{ "benchmark": "The Aggregate - Lichess Rating Prediction (10-game Pilot)", "board_id": "lichess-rating", "games": [ { "id": "game_01", "lichess_game_id": "n6H9m2oi", "url": "https://lichess.org/n6H9m2oi", "created_at": 1791744223231, "plies": 84, "selection_band": [ 400, 1300 ], "players": { "white": { "rating_at_game": 1363, "rapid_games_at_collection": 4696, "rapid_rd_at_collection": 46, "profile_retrieved_at": "2026-10-11T19:29:15.163331+00:00" }, "black": { "rating_at_game": 1159, "rapid_games_at_collection": 3260, "rapid_rd_at_collection": 45, "profile_retrieved_at": "2026-10-11T19:29:15.246000+00:00" } } }, { "id": "game_02", "lichess_game_id": "8foldxhD", "url": "https://lichess.org/8foldxhD", "created_at": 1791493705385, "plies": 50, "selection_band": [ 1900, 2200 ], "players": { "white": { "rating_at_game": 1575, "rapid_games_at_collection": 1928, "rapid_rd_at_collection": 45, "profile_retrieved_at": "2026-10-11T19:29:20.402616+00:00" }, "black": { "rating_at_game": 2410, "rapid_games_at_collection": 2936, "rapid_rd_at_collection": 45, "profile_retrieved_at": "2026-10-11T19:25:57.930514+00:00" } } }, { "id": "game_03", "lichess_game_id": "iSAnxtl2", "url": "https://lichess.org/iSAnxtl2", "created_at": 1791743013471, "plies": 91, "selection_band": [ 1600, 1900 ], "players": { "white": { "rating_at_game": 1782, "rapid_games_at_collection": 18588, "rapid_rd_at_collection": 45, "profile_retrieved_at": "2026-10-11T19:29:17.100336+00:00" }, "black": { "rating_at_game": 1901, "rapid_games_at_collection": 1366, "rapid_rd_at_collection": 50, "profile_retrieved_at": "2026-10-11T19:29:17.401277+00:00" } } }, { "id": "game_04", "lichess_game_id": "At0NmiOd", "url": "https://lichess.org/At0NmiOd", "created_at": 1791377582312, "plies": 52, "selection_band": [ 2200, 3200 ], "players": { "white": { "rating_at_game": 2508, "rapid_games_at_collection": 1324, "rapid_rd_at_collection": 47, "profile_retrieved_at": "2026-10-11T19:29:22.803205+00:00" }, "black": { "rating_at_game": 2718, "rapid_games_at_collection": 1136, "rapid_rd_at_collection": 59, "profile_retrieved_at": "2026-10-11T19:28:54.536331+00:00" } } }, { "id": "game_05", "lichess_game_id": "mCkfc84Y", "url": "https://lichess.org/mCkfc84Y", "created_at": 1791743797389, "plies": 58, "selection_band": [ 1300, 1600 ], "players": { "white": { "rating_at_game": 1120, "rapid_games_at_collection": 5115, "rapid_rd_at_collection": 45, "profile_retrieved_at": "2026-10-11T19:29:15.921476+00:00" }, "black": { "rating_at_game": 1867, "rapid_games_at_collection": 31199, "rapid_rd_at_collection": 45, "profile_retrieved_at": "2026-10-11T19:29:16.004197+00:00" } } }, { "id": "game_06", "lichess_game_id": "wLJhUOmU", "url": "https://lichess.org/wLJhUOmU", "created_at": 1791742657840, "plies": 62, "selection_band": [ 400, 1300 ], "players": { "white": { "rating_at_game": 1112, "rapid_games_at_collection": 3821, "rapid_rd_at_collection": 45, "profile_retrieved_at": "2026-10-11T19:25:53.582622+00:00" }, "black": { "rating_at_game": 1382, "rapid_games_at_collection": 5055, "rapid_rd_at_collection": 45, "profile_retrieved_at": "2026-10-11T19:29:15.750079+00:00" } } }, { "id": "game_07", "lichess_game_id": "qtL5x6uE", "url": "https://lichess.org/qtL5x6uE", "created_at": 1791743901913, "plies": 47, "selection_band": [ 1600, 1900 ], "players": { "white": { "rating_at_game": 1862, "rapid_games_at_collection": 7581, "rapid_rd_at_collection": 45, "profile_retrieved_at": "2026-10-11T19:29:19.502397+00:00" }, "black": { "rating_at_game": 1593, "rapid_games_at_collection": 13102, "rapid_rd_at_collection": 45, "profile_retrieved_at": "2026-10-11T19:29:19.801689+00:00" } } }, { "id": "game_08", "lichess_game_id": "5z2eLce5", "url": "https://lichess.org/5z2eLce5", "created_at": 1791745893736, "plies": 48, "selection_band": [ 1300, 1600 ], "players": { "white": { "rating_at_game": 1310, "rapid_games_at_collection": 4221, "rapid_rd_at_collection": 45, "profile_retrieved_at": "2026-10-11T19:29:16.090686+00:00" }, "black": { "rating_at_game": 1791, "rapid_games_at_collection": 4152, "rapid_rd_at_collection": 49, "profile_retrieved_at": "2026-10-11T19:29:16.199155+00:00" } } }, { "id": "game_09", "lichess_game_id": "tGYghpt5", "url": "https://lichess.org/tGYghpt5", "created_at": 1791181452386, "plies": 82, "selection_band": [ 1900, 2200 ], "players": { "white": { "rating_at_game": 2193, "rapid_games_at_collection": 4525, "rapid_rd_at_collection": 45, "profile_retrieved_at": "2026-10-11T19:29:21.302435+00:00" }, "black": { "rating_at_game": 2177, "rapid_games_at_collection": 8936, "rapid_rd_at_collection": 46, "profile_retrieved_at": "2026-10-11T19:25:54.928384+00:00" } } }, { "id": "game_10", "lichess_game_id": "ghN8Hbdl", "url": "https://lichess.org/ghN8Hbdl", "created_at": 1791398757874, "plies": 54, "selection_band": [ 2200, 3200 ], "players": { "white": { "rating_at_game": 2689, "rapid_games_at_collection": 9005, "rapid_rd_at_collection": 45, "profile_retrieved_at": "2026-10-11T19:28:55.121417+00:00" }, "black": { "rating_at_game": 2501, "rapid_games_at_collection": 3141, "rapid_rd_at_collection": 45, "profile_retrieved_at": "2026-10-11T19:29:22.503297+00:00" } } } ], "blind_games": [ { "id": "game_01", "pgn": "[White \"White\"]\n[Black \"Black\"]\n[Result \"0-1\"]\n[TimeControl \"600+0\"]\n\n1. e4 e5 2. d3 d5 3. Nc3 dxe4 4. Nxe4 f5 5. Nc3 Nc6 6. Nf3 Bd6 7. Bg5 Qd7 8.\nQe2 h6 9. Bd2 g5 10. O-O-O Nf6 11. h4 g4 12. Nh2 h5 13. g3 b6 14. Bg5 Nh7 15.\nNd5 Nxg5 16. Nf6+ Ke7 17. Nxd7 Kxd7 18. hxg5 Bb7 19. f3 Nd4 20. Qf2 gxf3 21.\nNxf3 Nxf3 22. Bg2 Rab8 23. Bxf3 Bxf3 24. Qxf3 Ke6 25. Rdf1 Rbf8 26. Qc6 a5 27.\nRe1 e4 28. Qc4+ Kd7 29. Qa4+ c6 30. dxe4 Bxg3 31. e5 Bxe1 32. Rxe1 h4 33. Qd4+\nKc7 34. Qd6+ Kb7 35. Qd7+ Ka6 36. Qd3+ b5 37. Qd6 Kb6 38. e6 h3 39. e7 Re8 40.\nQb8+ Rxb8 41. e8=Q Rbxe8 42. Rxe8 Rxe8 0-1" }, { "id": "game_02", "pgn": "[White \"White\"]\n[Black \"Black\"]\n[Result \"0-1\"]\n[TimeControl \"600+0\"]\n\n1. e4 e5 2. Nf3 Nc6 3. Bb5 a6 4. Ba4 Nf6 5. Nc3 b5 6. Bb3 Be7 7. O-O O-O 8. Nd5\nd6 9. c3 Nxe4 10. d4 exd4 11. cxd4 Nf6 12. Nxf6+ Bxf6 13. Be3 Na5 14. Bc2 Nc4\n15. Qd3 g6 16. Bh6 Re8 17. Qb3 Bb7 18. h4 Bxh4 19. Nxh4 Qxh4 20. Qh3 Qxh3 21.\ngxh3 Re2 22. Bd3 Rxb2 23. a3 Rb3 24. Bc2 Rxh3 25. Bg5 Rh1# 0-1" }, { "id": "game_03", "pgn": "[White \"White\"]\n[Black \"Black\"]\n[Result \"1-0\"]\n[TimeControl \"600+0\"]\n\n1. d4 f5 2. Bf4 Nf6 3. e3 e6 4. Nd2 b6 5. Ngf3 Bb7 6. c3 Be7 7. Be2 O-O 8. h4\nd6 9. Ng5 Qc8 10. O-O h6 11. Nh3 c5 12. dxc5 bxc5 13. Nc4 d5 14. Ne5 Bd6 15.\nNg6 Rd8 16. Bxd6 Kf7 17. Ne5+ Kg8 18. Be7 Re8 19. Bxf6 gxf6 20. Ng6 Qd7 21. Bh5\nQg7 22. Nhf4 e5 23. Nxd5 Nd7 24. Qb3 Bxd5 25. Qxd5+ Kh7 26. Rfd1 Nb6 27. Qxc5\nRac8 28. Qb4 Rc4 29. Qb5 Rg8 30. b3 Rc7 31. c4 f4 32. exf4 exf4 33. Qf5 Nc8 34.\nNe7+ Kh8 35. Nxg8 Qxg8 36. Qxf6+ Rg7 37. Rd8 Ne7 38. Rxg8+ Nxg8 39. Qd4 Kh7 40.\nQxf4 Ne7 41. Re1 Ng6 42. Qe4 a5 43. Bxg6+ Rxg6 44. h5 Kh8 45. Qxg6 a4 46. Re8#\n1-0" }, { "id": "game_04", "pgn": "[White \"White\"]\n[Black \"Black\"]\n[Result \"0-1\"]\n[TimeControl \"600+0\"]\n\n1. e4 c6 2. d4 d5 3. exd5 cxd5 4. Bd3 Nc6 5. c3 Qc7 6. Ne2 Bg4 7. O-O e6 8. Qe1\nNf6 9. Bg5 Bh5 10. f4 Bxe2 11. Qxe2 Be7 12. Nd2 h6 13. Bh4 O-O-O 14. b4 Kb8 15.\nNb3 Rc8 16. a4 Ne4 17. Be1 f5 18. Nc5 Bxc5 19. dxc5 g5 20. a5 a6 21. c4 gxf4\n22. b5 Nd4 23. Qb2 Qxc5 24. Bb4 Ne2+ 25. Kh1 Qg1+ 26. Rxg1 Nf2# 0-1" }, { "id": "game_05", "pgn": "[White \"White\"]\n[Black \"Black\"]\n[Result \"0-1\"]\n[TimeControl \"600+0\"]\n\n1. e4 c5 2. Bc4 e6 3. Nf3 a6 4. c3 b5 5. Bd3 c4 6. Be2 Bb7 7. d3 cxd3 8. Bxd3\nNf6 9. Nbd2 Be7 10. e5 Nd5 11. Be4 f5 12. Bd3 O-O 13. b3 Nc6 14. Bb2 Qc7 15. c4\nNf4 16. Be4 Nxg2+ 17. Kf1 fxe4 18. Nxe4 Nf4 19. Nfg5 h6 20. Nf3 Nb4 21. a3 Nbd3\n22. Bc3 Bxe4 23. Nd4 Bxh1 24. Kg1 Qb7 25. f3 Nh3+ 26. Kf1 Bxf3 27. Qxd3 Be4+\n28. Nf3 Bxd3+ 29. Kg2 Qxf3# 0-1" }, { "id": "game_06", "pgn": "[White \"White\"]\n[Black \"Black\"]\n[Result \"0-1\"]\n[TimeControl \"600+0\"]\n\n1. c4 d6 2. Nc3 Nf6 3. e3 Bd7 4. Bd3 e5 5. a3 Be7 6. Bb1 O-O 7. Qc2 g6 8. h4 h5\n9. b3 Bf5 10. e4 Bd7 11. Bb2 Nc6 12. Nf3 Bg4 13. Ng5 Nd4 14. Qd3 Nd7 15. f3 Be6\n16. f4 Bxg5 17. hxg5 Nc5 18. Qe3 Ncxb3 19. f5 Nxa1 20. Bxa1 Bxc4 21. fxg6 fxg6\n22. Nd5 Ne6 23. Qc3 Ba6 24. Nf6+ Kg7 25. Rf1 Bxf1 26. Kxf1 Nxg5 27. Ke1 Qxf6\n28. Qxc7+ Kh6 29. Bd3 Rac8 30. Qxb7 Rc1+ 31. Ke2 Qf2# 0-1" }, { "id": "game_07", "pgn": "[White \"White\"]\n[Black \"Black\"]\n[Result \"0-1\"]\n[TimeControl \"600+0\"]\n\n1. d4 d5 2. c4 c6 3. e3 Bf5 4. Nf3 Nf6 5. Nbd2 h6 6. Be2 Nbd7 7. O-O e6 8. cxd5\nexd5 9. h3 Bd6 10. b3 O-O 11. a4 Ne4 12. Nxe4 Bxe4 13. Ba3 Bxa3 14. Rxa3 Nf6\n15. Ne5 Nd7 16. Nxd7 Qxd7 17. b4 a6 18. Rb3 Qe6 19. b5 axb5 20. axb5 Qg6 21.\nBf3 Bc2 22. Bh5 Bxd1 23. Bxg6 Bxb3 24. Bd3 0-1" }, { "id": "game_08", "pgn": "[White \"White\"]\n[Black \"Black\"]\n[Result \"0-1\"]\n[TimeControl \"600+0\"]\n\n1. Nc3 e5 2. d3 d5 3. e3 Bb4 4. Bd2 Nf6 5. g3 O-O 6. Bg2 c6 7. Nf3 Bg4 8. O-O\nBxc3 9. Bxc3 e4 10. dxe4 Nxe4 11. Bb4 Re8 12. Qd4 Be6 13. Rad1 Nd7 14. Rfe1 c5\n15. Bxc5 Ndxc5 16. Ne5 Qb6 17. Bxe4 Nxe4 18. Qxb6 axb6 19. a3 f6 20. Nf3 Bg4\n21. Kg2 Ng5 22. Rd4 Bxf3+ 23. Kg1 Be4 24. h4 Nf3+ 0-1" }, { "id": "game_09", "pgn": "[White \"White\"]\n[Black \"Black\"]\n[Result \"0-1\"]\n[TimeControl \"600+0\"]\n\n1. d4 Nf6 2. c4 c5 3. d5 d6 4. Nc3 g6 5. e4 Bg7 6. g3 O-O 7. Bg2 Na6 8. a3 Bd7\n9. Nge2 Nc7 10. O-O b5 11. cxb5 Nxb5 12. Bd2 Rb8 13. Qc1 Ng4 14. h3 Ne5 15. Kh2\nNc4 16. Rb1 Nxd2 17. Qxd2 Qa5 18. Qc2 Rfc8 19. f4 Nd4 20. Qd3 Nxe2 21. Qxe2\nBxc3 22. bxc3 Qxa3 23. Ra1 Qxc3 24. Rfc1 Qd4 25. Rxa7 Rb2 26. Qf3 Qd2 27. Rg1\nRd8 28. e5 Bf5 29. g4 Bb1 30. f5 dxe5 31. fxg6 hxg6 32. Rxe7 Qf4+ 33. Qxf4 exf4\n34. Kh1 Ba2 35. Rc1 Bxd5 36. Bxd5 Rxd5 37. Re8+ Kg7 38. Rc8 Rdd2 39. R8xc5 Rh2+\n40. Kg1 Rbg2+ 41. Kf1 f3 0-1" }, { "id": "game_10", "pgn": "[White \"White\"]\n[Black \"Black\"]\n[Result \"0-1\"]\n[TimeControl \"600+0\"]\n\n1. c4 e6 2. Nf3 d5 3. g3 Nf6 4. Bg2 Nbd7 5. O-O Be7 6. b3 O-O 7. Bb2 c5 8. cxd5\nexd5 9. d4 b6 10. Nc3 Bb7 11. Rc1 Rc8 12. dxc5 Nxc5 13. Nd4 Qd7 14. Qd2 Rfe8\n15. Qf4 Ne6 16. Nxe6 fxe6 17. Rfd1 Bd6 18. Qh4 Rf8 19. e3 Qf7 20. Nb5 Be7 21.\nQf4 Qg6 22. Rxc8 Bxc8 23. Qc7 Bc5 24. Nxa7 Qc2 25. Rf1 Ba6 26. Be5 Bxf1 27.\nBxf1 Ne8 0-1" } ], "prompt": "Estimate each player's established Lichess RAPID rating immediately before this game, using only the moves and time control below. The target is Lichess's Glicko-2 rating, often informally called Elo; do not convert to FIDE or Chess.com ratings.\n\nThis is a standard rated 10-minute game with no increment. Both players have substantial Lichess rapid histories and non-provisional ratings. Their identities and ratings are hidden. You have no engine, browsing, or other tools. Assess the cumulative quality of each player's decisions, including opening understanding, tactics, positional choices, defence, and endgame play. A win or a single blunder does not determine a player's rating. Estimate the two players independently, allowing for uncertainty from seeing only one game.\n\nReturn only the requested JSON. For each color, provide a point estimate rounded to the nearest 50 rating points, an 80% uncertainty interval for the player's established rating, and a brief explanation citing concrete move numbers. The interval concerns the unknown rating, not the performance rating of this game. Keep each explanation to at most 80 words.\n\nGAME:\n{pgn}\n", "models": [ { "model": "google/gemini-3-flash-preview", "summary": { "games": 10, "player_estimates": 20, "actual_rating_range": [ 1112, 2718 ], "mae": 255.85, "rmse": 321.5524063041668, "signed_bias": -133.15, "pearson": 0.8762541146358507, "spearman": 0.8589782672582525, "within": { "100": { "count": 3, "fraction": 0.15 }, "200": { "count": 9, "fraction": 0.45 }, "300": { "count": 14, "fraction": 0.7 } }, "interval_80_coverage": 0.4, "interval_80_mean_width": 350.0, "game_mean_rating_mae": 216.05, "mae_game_bootstrap_95_interval": [ 178.95, 354.75374999999997 ], "baseline_1500_mae": 506.05, "baseline_leave_one_game_out_mean_mae": 471.5666666666666, "games_beating_loo_baseline": 8, "history_count_range": [ 1136, 31199 ], "max_profile_rating_deviation": 59 }, "predictions": [ { "game": "game_01", "color": "white", "actual": 1363, "predicted": 1350, "error": -13, "absolute_error": 13, "low_80": 1150, "high_80": 1550, "interval_covers": true, "loo_baseline": 1916.1666666666667, "rapid_games": 4696, "rating_deviation": 46, "explanation": "White plays a passive opening (2. d3) but finds a tactical fork (16. Nf6+) to win the queen. However, the conversion is erratic, losing the exchange with 31. e5 and failing to coordinate. The game concludes with 40. Qb8+??, a massive blunder that leads to the loss of all remaining material. This inability to convert a winning advantage and the final catastrophic oversight are typical of a 1350 Lichess rapid rating." }, { "game": "game_01", "color": "black", "actual": 1159, "predicted": 1400, "error": 241, "absolute_error": 241, "low_80": 1200, "high_80": 1600, "interval_covers": false, "loo_baseline": 1916.1666666666667, "rapid_games": 3260, "rating_deviation": 45, "explanation": "Black's aggressive expansion (4... f5, 9... g5) leads to a major tactical oversight with 15... Nxg5?, allowing a queen-winning fork. However, Black demonstrates resilience, creating counterplay with passed pawns and eventually winning the exchange (31... Bxe1). Black's persistence in a lost position and the final exploitation of White's blunders are characteristic of the 1400 level on Lichess, where games are often decided by late-game tactical errors rather than opening precision." }, { "game": "game_02", "color": "white", "actual": 1575, "predicted": 1600, "error": 25, "absolute_error": 25, "low_80": 1400, "high_80": 1800, "interval_covers": true, "loo_baseline": 1834.888888888889, "rapid_games": 1928, "rating_deviation": 45, "explanation": "White showed familiarity with Ruy Lopez structures, playing a coherent opening (1-10) and a thematic battery (15. Qd3). However, 18. h4 was a tactical oversight that lost a second pawn. The game concluded with 25. Bg5, which missed a simple mate in one. These errors suggest a player who understands basic strategy but lacks the tactical consistency required for higher ratings." }, { "game": "game_02", "color": "black", "actual": 2410, "predicted": 1800, "error": -610, "absolute_error": 610, "low_80": 1600, "high_80": 2000, "interval_covers": false, "loo_baseline": 1834.888888888889, "rapid_games": 2936, "rating_deviation": 45, "explanation": "Black displayed solid tactical awareness and positional understanding. After winning a pawn with 9... Nxe4, Black navigated the resulting pressure well. The maneuver 13... Na5 and 14... Nc4 demonstrated strong piece activity. Black's decision at 18... Bxh4 was precise, leading to a winning material advantage. The conversion was efficient, using the seventh rank effectively before finishing with 25... Rh1#." }, { "game": "game_03", "color": "white", "actual": 1782, "predicted": 1900, "error": 118, "absolute_error": 118, "low_80": 1750, "high_80": 2050, "interval_covers": true, "loo_baseline": 1851.6666666666667, "rapid_games": 18588, "rating_deviation": 45, "explanation": "White demonstrated strong tactical awareness and conversion skills. The maneuver starting with 9. Ng5 and culminating in the Ne5/Ng6 pressure showed a sophisticated understanding of the position. White capitalized on Black's inaccuracies with precise moves like 16. Bxd6 (creating the Ne7+ threat) and 23. Nxd5. The subsequent conversion was clean and methodical, utilizing the open d-file (26. Rfd1) and final mating net without allowing any counterplay, typical of a high 1800-rated player." }, { "game": "game_03", "color": "black", "actual": 1901, "predicted": 1700, "error": -201, "absolute_error": 201, "low_80": 1550, "high_80": 1850, "interval_covers": false, "loo_baseline": 1851.6666666666667, "rapid_games": 1366, "rating_deviation": 50, "explanation": "Black played a standard Dutch setup but faltered under tactical pressure. The move 16... Kf7 was forced by the threat of a fork on e7, but the resulting position after 19. Bxf6 left the kingside shattered. Black missed the tactical blow 23. Nxd5 and the follow-up 24. Qb3, leading to a quick collapse. While the opening phase was competent, the tactical oversights and lack of defensive resilience suggest a rating in the 1700 range." }, { "game": "game_04", "color": "white", "actual": 2508, "predicted": 1950, "error": -558, "absolute_error": 558, "low_80": 1800, "high_80": 2100, "interval_covers": false, "loo_baseline": 1765.9444444444443, "rapid_games": 1324, "rating_deviation": 47, "explanation": "White demonstrates solid opening knowledge in the Caro-Kann Exchange and initiates a logical queenside pawn storm (14. b4, 16. a4). However, the central break 21. c4 was mistimed, allowing Black's pieces to become too active. The failure to foresee the tactical consequences of the open a2-g8 diagonal and the subsequent smothered mate (24. Bb4?) indicates a lack of high-level tactical prophylaxis, typical of a strong club player who can buckle under pressure." }, { "game": "game_04", "color": "black", "actual": 2718, "predicted": 2050, "error": -668, "absolute_error": 668, "low_80": 1900, "high_80": 2200, "interval_covers": false, "loo_baseline": 1765.9444444444443, "rapid_games": 1136, "rating_deviation": 59, "explanation": "Black handles the opening with precision, employing a modern setup with 5... Qc7 and 9... Bh5. The choice to castle queenside led to a sharp battle where Black showed superior calculation. By stabilizing the queenside with 20... a6 and launching a decisive kingside counterattack (19... g5), Black maintained the initiative. The game concludes with a beautiful, correctly calculated smothered mate sequence (24... Ne2+ and 25... Qg1+), reflecting a high degree of tactical alertness and poise." }, { "game": "game_05", "color": "white", "actual": 1120, "predicted": 1250, "error": 130, "absolute_error": 130, "low_80": 1050, "high_80": 1450, "interval_covers": true, "loo_baseline": 1890.3333333333333, "rapid_games": 5115, "rating_deviation": 45, "explanation": "White's opening (2. Bc4 and 5. Bd3) displayed poor positional sense, obstructing development. Tactical vision was severely lacking, blundering a pawn on g2 with 16. Be4 and subsequently losing a piece. Failure to address the pressure on d3 (21... Nbd3) led to a total collapse. This performance reflects a player around 1250 who lacks defensive awareness and basic opening preparation." }, { "game": "game_05", "color": "black", "actual": 1867, "predicted": 1600, "error": -267, "absolute_error": 267, "low_80": 1400, "high_80": 1800, "interval_covers": false, "loo_baseline": 1890.3333333333333, "rapid_games": 31199, "rating_deviation": 45, "explanation": "Black capitalized well on White's inaccuracies, using 4... b5 and 5... c4 to seize space. Black demonstrated good tactical alertness by capturing the g2-pawn and utilizing the f4-outpost. The sequence starting from 20... Nb4 shows an ability to coordinate pieces for a decisive attack. Black's play was clean and efficient, typical of an established 1600 player who can punish significant mistakes without allowing counterplay." }, { "game": "game_06", "color": "white", "actual": 1112, "predicted": 1250, "error": 138, "absolute_error": 138, "low_80": 1050, "high_80": 1450, "interval_covers": true, "loo_baseline": 1917.7222222222222, "rapid_games": 3821, "rating_deviation": 45, "explanation": "White's opening (4. Bd3, 6. Bb1) is extremely suboptimal, prioritizing a crude battery over central development and control. Tactical awareness was poor; White allowed the powerful maneuvers 17... Nc5 and 18... Ncxb3, losing the exchange. The subsequent collapse, including losing a full rook with 25. Rf1 and the knight on g5, reflects a lack of board vision and piece safety common in the 1100-1300 range." }, { "game": "game_06", "color": "black", "actual": 1382, "predicted": 1700, "error": 318, "absolute_error": 318, "low_80": 1500, "high_80": 1900, "interval_covers": false, "loo_baseline": 1917.7222222222222, "rapid_games": 5055, "rating_deviation": 45, "explanation": "Black demonstrated solid fundamental play, calmly countering White's eccentric setup. Moves like 8... h5 and 9... Bf5 showed good understanding of flank attacks. Black effectively used centralized outposts (13... Nd4, 17... Nc5) to exert pressure and force tactical errors. The conversion of the material advantage was clean and accurate, culminating in a precise checkmate, characteristic of a high-intermediate player who punishes opening irregularities efficiently." }, { "game": "game_07", "color": "white", "actual": 1862, "predicted": 1750, "error": -112, "absolute_error": 112, "low_80": 1550, "high_80": 1950, "interval_covers": true, "loo_baseline": 1864.3333333333333, "rapid_games": 7581, "rating_deviation": 45, "explanation": "White demonstrates reasonable opening knowledge and positional awareness, utilizing a minority attack-style expansion with 17. b4. The trade of the dark-squared bishop (13. Ba3) shows a conceptual understanding beyond a basic level. However, the game ends due to a critical tactical oversight (21. Bf3), missing the 21...Bc2 fork. This combination of structural competence and sudden tactical vulnerability is characteristic of an established 1750 Lichess rapid player." }, { "game": "game_07", "color": "black", "actual": 1593, "predicted": 1850, "error": 257, "absolute_error": 257, "low_80": 1650, "high_80": 2050, "interval_covers": false, "loo_baseline": 1864.3333333333333, "rapid_games": 13102, "rating_deviation": 45, "explanation": "Black plays an accurate Slav, handling the development of the bishop and the knight maneuver 11...Ne4 proficiently. The tactical sequence starting with 20...Qg6, which provoked the game-ending error 21. Bf3, shows high alertness. Black’s ability to seize the initiative and immediately capitalize on a tactical blunder suggests a solid 1850-rated performance, displaying a level of consistency and tactical vision common at this rating on Lichess." }, { "game": "game_08", "color": "white", "actual": 1310, "predicted": 1300, "error": -10, "absolute_error": 10, "low_80": 1150, "high_80": 1450, "interval_covers": true, "loo_baseline": 1884.0, "rapid_games": 4221, "rating_deviation": 45, "explanation": "White's opening was passive and lacked central control. Significant tactical lapses occurred early, notably allowing the e4 fork/pin at move 9 and completely missing the c5 fork that trapped the bishop at move 14. Finally, 20. Nf3 was a decisive blunder that walked into a pin (20... Bg4) and a discovery (21... Ng5). These repeated one-move tactical oversights and lack of board awareness are characteristic of a lower-intermediate rating." }, { "game": "game_08", "color": "black", "actual": 1791, "predicted": 1600, "error": -191, "absolute_error": 191, "low_80": 1450, "high_80": 1750, "interval_covers": false, "loo_baseline": 1884.0, "rapid_games": 4152, "rating_deviation": 49, "explanation": "Black played a very solid game, occupying the center and developing naturally. They demonstrated good tactical vision by identifying and punishing White's mistakes, specifically the e4 push at move 9 and the winning c5 sequence. The final combination involving 20... Bg4 and 21... Ng5 showed an ability to calculate simple tactical sequences to force a resignation. Black's play was disciplined and efficient throughout the encounter." }, { "game": "game_09", "color": "white", "actual": 2193, "predicted": 1800, "error": -393, "absolute_error": 393, "low_80": 1650, "high_80": 1950, "interval_covers": false, "loo_baseline": 1813.5, "rapid_games": 4525, "rating_deviation": 45, "explanation": "White played a solid opening system but struggled to deal with Black's queenside pressure. The move 15. Kh2 was passive, allowing Black's knights to dominate the center. White's decision-making faltered under pressure, specifically failing to protect the b2/c3 structure at move 21. While White found some counterplay with 25. Rxa7, the inability to defend the kingside in the endgame (38-41) suggests a high-intermediate level rather than an expert one." }, { "game": "game_09", "color": "black", "actual": 2177, "predicted": 2050, "error": -127, "absolute_error": 127, "low_80": 1900, "high_80": 2200, "interval_covers": true, "loo_baseline": 1813.5, "rapid_games": 8936, "rating_deviation": 46, "explanation": "Black showed a high level of proficiency in Benoni-style structures, employing effective knight maneuvers (9... Nc7, 14... Ne5) and timely queenside expansion (10... b5). The tactical sequence starting with 21... Bxc3 displayed strong calculation, successfully converting a structural advantage into a material one. Black's technique in the heavy-piece endgame was precise, culminating in a clean mating net with 41... f3, indicating a player well-versed in both strategy and tactics." }, { "game": "game_10", "color": "white", "actual": 2689, "predicted": 2150, "error": -539, "absolute_error": 539, "low_80": 2000, "high_80": 2300, "interval_covers": false, "loo_baseline": 1767.9444444444443, "rapid_games": 9005, "rating_deviation": 45, "explanation": "White displays strong opening knowledge of the Reti/English, managing a balanced position through move 20. The decision 23. Qc7 aimed for activity but underestimated the tactical repercussions of 24. Nxa7. Losing the exchange and then failing to see the queen trap 27... Ne8 indicates a lapse in tactical vigilance, contrasting with the sophisticated positional play shown earlier in the game." }, { "game": "game_10", "color": "black", "actual": 2501, "predicted": 2300, "error": -201, "absolute_error": 201, "low_80": 2150, "high_80": 2450, "interval_covers": false, "loo_baseline": 1767.9444444444443, "rapid_games": 3141, "rating_deviation": 45, "explanation": "Black demonstrated excellent theoretical preparation and positional patience. Piece coordination was superior, particularly utilizing the d-file and kingside pressure. The tactical sequence starting with 24... Qc2 shows high-level calculation, followed by the decisive 27... Ne8 to trap the White queen. This level of precision, combined with a clean conversion of a tactical advantage, is characteristic of a player who exploits minor inaccuracies ruthlessly." } ], "cost_usd": 0.102385, "reasoning_effort": "medium", "max_tokens": 8192, "run_start": "2026-10-11", "run_end": "2026-10-11", "display_name": "Gemini 3 Flash (Preview)" }, { "model": "google/gemini-3.1-pro-preview", "summary": { "games": 10, "player_estimates": 20, "actual_rating_range": [ 1112, 2718 ], "mae": 263.95, "rmse": 351.6688072604677, "signed_bias": -218.15, "pearson": 0.872304918722874, "spearman": 0.8790976040984101, "within": { "100": { "count": 8, "fraction": 0.4 }, "200": { "count": 10, "fraction": 0.5 }, "300": { "count": 12, "fraction": 0.6 } }, "interval_80_coverage": 0.5, "interval_80_mean_width": 342.5, "game_mean_rating_mae": 233.75, "mae_game_bootstrap_95_interval": [ 153.2975, 386.00249999999994 ], "baseline_1500_mae": 506.05, "baseline_leave_one_game_out_mean_mae": 471.5666666666666, "games_beating_loo_baseline": 7, "history_count_range": [ 1136, 31199 ], "max_profile_rating_deviation": 59 }, "predictions": [ { "game": "game_01", "color": "white", "actual": 1363, "predicted": 1250, "error": -113, "absolute_error": 113, "low_80": 1100, "high_80": 1400, "interval_covers": true, "loo_baseline": 1916.1666666666667, "rapid_games": 4696, "rating_deviation": 46, "explanation": "White played a passive and questionable opening with 2.d3 and 3.Nc3. They showed some tactical awareness by finding the royal fork 16.Nf6+, but their game completely fell apart later, blundering a rook with 31.e5?? and finally giving away their queen with the absurd 40.Qb8+??, suggesting a low rating." }, { "game": "game_01", "color": "black", "actual": 1159, "predicted": 1200, "error": 41, "absolute_error": 41, "low_80": 1050, "high_80": 1350, "interval_covers": true, "loo_baseline": 1916.1666666666667, "rapid_games": 3260, "rating_deviation": 45, "explanation": "Black responded aggressively in the opening but showed poor board vision by walking into a royal fork on move 16, losing their queen for a knight. They played on and capitalized on White's later blunders (31...Bxe1 and 40...Rxb8), which indicates some resilience but overall erratic play typical of this level." }, { "game": "game_02", "color": "white", "actual": 1575, "predicted": 1300, "error": -275, "absolute_error": 275, "low_80": 1100, "high_80": 1500, "interval_covers": false, "loo_baseline": 1834.888888888889, "rapid_games": 1928, "rating_deviation": 45, "explanation": "White played a sub-optimal opening setup with 5. Nc3 and blundered a central pawn with 9. c3. Further blunders like 18. h4 hanging another pawn and completely missing a one-move checkmate with 24. Bc2 suggest a beginner to intermediate level of tactical awareness." }, { "game": "game_02", "color": "black", "actual": 2410, "predicted": 1650, "error": -760, "absolute_error": 760, "low_80": 1450, "high_80": 1850, "interval_covers": false, "loo_baseline": 1834.888888888889, "rapid_games": 2936, "rating_deviation": 45, "explanation": "Black played a very solid and punishing game. They correctly grabbed the hanging pawn on e4, found the strong maneuver 13... Na5 and 14... Nc4, calmly defended against White's superficial attack with 15... g6, and efficiently converted their advantage into a straightforward checkmate." }, { "game": "game_03", "color": "white", "actual": 1782, "predicted": 1800, "error": 18, "absolute_error": 18, "low_80": 1600, "high_80": 2000, "interval_covers": true, "loo_baseline": 1851.6666666666667, "rapid_games": 18588, "rating_deviation": 45, "explanation": "White played a solid London System and demonstrated good tactical vision, maneuvering well with Nc4 and Ne5. White capitalized on Black's blunders efficiently, finding strong continuations like 23. Nxd5 and the decisive 34. Ne7+ and 37. Rd8 to convert the advantage smoothly." }, { "game": "game_03", "color": "black", "actual": 1901, "predicted": 1500, "error": -401, "absolute_error": 401, "low_80": 1300, "high_80": 1700, "interval_covers": false, "loo_baseline": 1851.6666666666667, "rapid_games": 1366, "rating_deviation": 50, "explanation": "Black played a reasonable Dutch setup initially but completely collapsed tactically in the middlegame. The bizarre 16... Kf7? blundered a piece inexplicably (missing 16... Qxd6). Black continued to play on but was dismantled by White's simple threats and pins, showing a lack of basic board awareness." }, { "game": "game_04", "color": "white", "actual": 2508, "predicted": 1950, "error": -558, "absolute_error": 558, "low_80": 1800, "high_80": 2100, "interval_covers": false, "loo_baseline": 1765.9444444444443, "rapid_games": 1324, "rating_deviation": 47, "explanation": "White played the opening and early middlegame reasonably well, launching a logical queenside pawn storm with 14. b4 and 16. a4. However, White overextended and completely missed Black's tactical threats in the center, blundering into a smothered mate sequence with 24. Bb4." }, { "game": "game_04", "color": "black", "actual": 2718, "predicted": 2150, "error": -568, "absolute_error": 568, "low_80": 1950, "high_80": 2300, "interval_covers": false, "loo_baseline": 1765.9444444444443, "rapid_games": 1136, "rating_deviation": 59, "explanation": "Black demonstrated strong positional understanding with 13... O-O-O and 16... Ne4, launching a well-timed kingside counterattack. The setup and execution of the smothered mate starting with 22... Nd4 and culminating in 26... Nf2# was brilliant and indicates high tactical awareness." }, { "game": "game_05", "color": "white", "actual": 1120, "predicted": 1250, "error": 130, "absolute_error": 130, "low_80": 1100, "high_80": 1400, "interval_covers": true, "loo_baseline": 1890.3333333333333, "rapid_games": 5115, "rating_deviation": 45, "explanation": "White played an awkward opening with 2. Bc4 and 5. Bd3, followed by a series of tactical blunders including 11. Be4 and 16. Be4 which hung pieces and allowed devastating attacks. White showed limited awareness of king safety and piece coordination." }, { "game": "game_05", "color": "black", "actual": 1867, "predicted": 1550, "error": -317, "absolute_error": 317, "low_80": 1350, "high_80": 1750, "interval_covers": false, "loo_baseline": 1890.3333333333333, "rapid_games": 31199, "rating_deviation": 45, "explanation": "Black demonstrated a good grasp of the Sicilian structure, expanding purposefully on the queenside. Black capitalized well on White's mistakes with strong tactical shots like 15... Nf4! and smoothly executed the resulting attack and mating sequence." }, { "game": "game_06", "color": "white", "actual": 1112, "predicted": 1200, "error": 88, "absolute_error": 88, "low_80": 1000, "high_80": 1400, "interval_covers": true, "loo_baseline": 1917.7222222222222, "rapid_games": 3821, "rating_deviation": 45, "explanation": "White demonstrates a lack of basic opening principles with the bizarre 4. Bd3 and 6. Bb1 maneuver to set up a primitive battery. White neglects development, pushes pawns aggressively with 8. h4, and misses simple tactics, dropping massive material to knight forks (18. Qe3 Ncxb3)." }, { "game": "game_06", "color": "black", "actual": 1382, "predicted": 1450, "error": 68, "absolute_error": 68, "low_80": 1250, "high_80": 1650, "interval_covers": true, "loo_baseline": 1917.7222222222222, "rapid_games": 5055, "rating_deviation": 45, "explanation": "Black plays logically, easily parrying White's transparent threats with 7...g6 and 8...h5. Black exploits White's lack of development excellently using outposts (13...Nd4, 17...Nc5) and converts the material advantage efficiently, culminating in a sharp mating net with 29...Rac8 and 30...Rc1+." }, { "game": "game_07", "color": "white", "actual": 1862, "predicted": 1400, "error": -462, "absolute_error": 462, "low_80": 1250, "high_80": 1550, "interval_covers": false, "loo_baseline": 1864.3333333333333, "rapid_games": 7581, "rating_deviation": 45, "explanation": "White played a very passive opening setup against the Slav (3. e3, 5. Nbd2, 6. Be2) and lacked a clear middlegame plan. The decisive blunder 21. Bf3 showed a lack of tactical awareness, missing the fork 21...Bc2 after defending against the mate threat on g2." }, { "game": "game_07", "color": "black", "actual": 1593, "predicted": 1550, "error": -43, "absolute_error": 43, "low_80": 1400, "high_80": 1700, "interval_covers": true, "loo_baseline": 1864.3333333333333, "rapid_games": 13102, "rating_deviation": 45, "explanation": "Black developed sensibly and actively, smoothly centralizing pieces and controlling the center. The multi-purpose move 20...Qg6 showed good situational awareness, creating a mate threat while setting up the winning tactical fork with 21...Bc2. Black capitalized efficiently on White's mistakes." }, { "game": "game_08", "color": "white", "actual": 1310, "predicted": 1350, "error": 40, "absolute_error": 40, "low_80": 1200, "high_80": 1500, "interval_covers": true, "loo_baseline": 1884.0, "rapid_games": 4221, "rating_deviation": 45, "explanation": "White played a very passive and timid opening (2. d3, 3. e3). Tactically, White struggled throughout, blundering a piece to a simple pawn fork (14... c5), failing to handle the pin on the f3 knight (20... Bg4, 21... Ng5), and finally walking into a family fork with 24. h4." }, { "game": "game_08", "color": "black", "actual": 1791, "predicted": 1750, "error": -41, "absolute_error": 41, "low_80": 1550, "high_80": 1950, "interval_covers": true, "loo_baseline": 1884.0, "rapid_games": 4152, "rating_deviation": 49, "explanation": "Black demonstrated strong fundamentals, seizing space and developing naturally against White's passive setup. Black showed good tactical awareness, spotting the 14... c5 pawn fork and later clinically exploiting the pinned knight on f3 with 20... Bg4 and 21... Ng5. The play was crisp and decisive." }, { "game": "game_09", "color": "white", "actual": 2193, "predicted": 1950, "error": -243, "absolute_error": 243, "low_80": 1800, "high_80": 2100, "interval_covers": false, "loo_baseline": 1813.5, "rapid_games": 4525, "rating_deviation": 45, "explanation": "White played the opening passively with 8. a3 and 13. Qc1, mishandling the Benoni structure. They allowed Black's knights to completely dominate the center (15...Nc4, 19...Nd4) and lost a pawn with 21. Qxe2. White showed basic resilience but lacked strategic depth." }, { "game": "game_09", "color": "black", "actual": 2177, "predicted": 2250, "error": 73, "absolute_error": 73, "low_80": 2100, "high_80": 2400, "interval_covers": true, "loo_baseline": 1813.5, "rapid_games": 8936, "rating_deviation": 46, "explanation": "Black demonstrated an excellent understanding of the Benoni structure with the textbook maneuvers 7...Na6, 9...Nc7, and the 10...b5 break. The knight tour 13...Ng4 to c4 was devastating. Black played precisely to convert the advantage, capping it off with the clinical 32...Qf4+ and strong endgame technique." }, { "game": "game_10", "color": "white", "actual": 2689, "predicted": 2000, "error": -689, "absolute_error": 689, "low_80": 1850, "high_80": 2150, "interval_covers": false, "loo_baseline": 1767.9444444444443, "rapid_games": 9005, "rating_deviation": 45, "explanation": "White demonstrated a solid understanding of the Reti opening and played a logical middlegame, but their greedy 23. Qc7 allowed a devastating tactical sequence starting with 24... Qc2." }, { "game": "game_10", "color": "black", "actual": 2501, "predicted": 2150, "error": -351, "absolute_error": 351, "low_80": 2000, "high_80": 2300, "interval_covers": false, "loo_baseline": 1767.9444444444443, "rapid_games": 3141, "rating_deviation": 45, "explanation": "Black played a very smooth and high-quality game, handling the opening comfortably and remaining alert to launch a powerful counterattack with 24... Qc2 and the beautiful finishing blow 27... Ne8." } ], "cost_usd": 0.139576, "reasoning_effort": "medium", "max_tokens": 8192, "run_start": "2026-10-11", "run_end": "2026-10-11", "display_name": "Gemini 3.1 Pro (Preview)" }, { "model": "google/gemini-3.8-flash", "summary": { "games": 10, "player_estimates": 20, "actual_rating_range": [ 1112, 2718 ], "mae": 280.85, "rmse": 367.06395900442203, "signed_bias": -235.65, "pearson": 0.8961349192924846, "spearman": 0.8902352321346022, "within": { "100": { "count": 6, "fraction": 0.3 }, "200": { "count": 10, "fraction": 0.5 }, "300": { "count": 12, "fraction": 0.6 } }, "interval_80_coverage": 0.45, "interval_80_mean_width": 300.0, "game_mean_rating_mae": 256.25, "mae_game_bootstrap_95_interval": [ 174.19875, 410.95749999999987 ], "baseline_1500_mae": 506.05, "baseline_leave_one_game_out_mean_mae": 471.5666666666666, "games_beating_loo_baseline": 7, "history_count_range": [ 1136, 31199 ], "max_profile_rating_deviation": 59 }, "predictions": [ { "game": "game_01", "color": "white", "actual": 1363, "predicted": 1200, "error": -163, "absolute_error": 163, "low_80": 1050, "high_80": 1350, "interval_covers": false, "loo_baseline": 1916.1666666666667, "rapid_games": 4696, "rating_deviation": 46, "explanation": "White spotted the nice tactic 16. Nf6+ to win Black's queen, but overall play demonstrates significant tactical instability typical of beginner-intermediate players. White quickly gave back material with 21. Nxf3?? (hanging a piece to 21...Nxf3), blundered an exchange with 31. e5?? allowing 31...Bxe1, and finally threw the game outright with 40. Qb8+?? hanging the queen in one move." }, { "game": "game_01", "color": "black", "actual": 1159, "predicted": 1250, "error": 91, "absolute_error": 91, "low_80": 1100, "high_80": 1400, "interval_covers": true, "loo_baseline": 1916.1666666666667, "rapid_games": 3260, "rating_deviation": 45, "explanation": "Black pursued a crude, aggressive kingside pawn storm (4...f5, 9...g5, 11...g4) while leaving the king uncastled. Black blundered queen for minor piece due to tactical blindness to 16. Nf6+ after 15...Nxg5. However, Black kept generating practical counterplay with the h-pawn, spotted White's tactical gifts on f3 and e1, and ultimately converted after White collapsed." }, { "game": "game_02", "color": "white", "actual": 1575, "predicted": 1450, "error": -125, "absolute_error": 125, "low_80": 1300, "high_80": 1600, "interval_covers": true, "loo_baseline": 1834.888888888889, "rapid_games": 1928, "rating_deviation": 45, "explanation": "White navigates a standard Ruy Lopez reasonably well early on but quickly loses control. Sacrificing or hanging the e4 pawn on move 9 yielded little compensation. White then dropped another pawn with 18. h4? and traded into a hopeless endgame. Finally, White suffered total tactical blindness by missing the imminent mate threat on h1 after 24...Rxh3, casually playing 25. Bg5?? and allowing 25...Rh1#." }, { "game": "game_02", "color": "black", "actual": 2410, "predicted": 1650, "error": -760, "absolute_error": 760, "low_80": 1500, "high_80": 1800, "interval_covers": false, "loo_baseline": 1834.888888888889, "rapid_games": 2936, "rating_deviation": 45, "explanation": "Black demonstrated clean, disciplined chess throughout. Black correctly snapped up the free central pawn with 9...Nxe4, developed actively with 13...Na5 and 14...Nc4, and defended against White's crude mate threats with 15...g6. Black immediately punished White's blunder on move 18, traded queens into a completely winning position, effectively dominated the second rank with 21...Re2, and efficiently executed the mating net." }, { "game": "game_03", "color": "white", "actual": 1782, "predicted": 1800, "error": 18, "absolute_error": 18, "low_80": 1650, "high_80": 1950, "interval_covers": true, "loo_baseline": 1851.6666666666667, "rapid_games": 18588, "rating_deviation": 45, "explanation": "White demonstrated strong tactical alertness and positional control throughout. Moves like 16. Bxd6! leveraged the threat of a knight fork on e7, and 24. Qb3! exerted decisive multi-directional pressure against Black's loose pieces. The conversion was very clean, seizing open files (26. Rfd1), creating mating nets (33. Qf5 followed by 34. Ne7+), and executing a textbook mate with 46. Re8# without allowing counterplay." }, { "game": "game_03", "color": "black", "actual": 1901, "predicted": 1600, "error": -301, "absolute_error": 301, "low_80": 1450, "high_80": 1750, "interval_covers": false, "loo_baseline": 1851.6666666666667, "rapid_games": 1366, "rating_deviation": 50, "explanation": "Black navigated the opening with a coherent Dutch setup but struggled as soon as tactical complications arose. 14... Bd6 ran into 15. Ng6 and the tactic 16. Bxd6! (exploiting the e7 fork). Later, 22... e5 walked directly into 23. Nxd5, and 33... Nc8 allowed the devastating 34. Ne7+ discovery. The passive defense and repeated oversights to knight forks and pins are typical of the 1500–1700 range." }, { "game": "game_04", "color": "white", "actual": 2508, "predicted": 1850, "error": -658, "absolute_error": 658, "low_80": 1700, "high_80": 2000, "interval_covers": false, "loo_baseline": 1765.9444444444443, "rapid_games": 1324, "rating_deviation": 47, "explanation": "White demonstrates solid opening preparation in the Caro-Kann Exchange (5. c3, 6. Ne2) and formulates a standard queenside pawn storm plan (14. b4, 16. a4). Positional handling is reasonable through move 20. However, White neglects king safety on the g1-a7 diagonal and collapses tactically under pressure, blundering into an elementary smothered mate combination with 24. Bb4 instead of defending against Black's threats." }, { "game": "game_04", "color": "black", "actual": 2718, "predicted": 1950, "error": -768, "absolute_error": 768, "low_80": 1800, "high_80": 2100, "interval_covers": false, "loo_baseline": 1765.9444444444443, "rapid_games": 1136, "rating_deviation": 59, "explanation": "Black demonstrates confident opening mastery with 5...Qc7 and 6...Bg4, followed by sharp opposite-side castling plans. Moves like 16...Ne4 and 19...g5 seize dynamic counterplay on both flanks. When White miscalculates with 24. Bb4, Black instantly spots the classic smothered mate combination starting with 24...Ne2+ and finishes ruthlessly with 25...Qg1+ and 26...Nf2#." }, { "game": "game_05", "color": "white", "actual": 1120, "predicted": 1200, "error": 80, "absolute_error": 80, "low_80": 1050, "high_80": 1350, "interval_covers": true, "loo_baseline": 1890.3333333333333, "rapid_games": 5115, "rating_deviation": 45, "explanation": "White displays weak opening understanding, wasting multiple tempi moving the light-squared bishop (moves 2, 5, 6, 8, 11, 12, and 16). Move 13. b3 leaves c3 undefended, and 16. Be4 is a catastrophic tactical blunder that loses pawns and opens lines against White's king. After 17. Kf1, White's position collapses rapidly due to complete lack of coordination and neglect of king safety." }, { "game": "game_05", "color": "black", "actual": 1867, "predicted": 1550, "error": -317, "absolute_error": 317, "low_80": 1400, "high_80": 1700, "interval_covers": false, "loo_baseline": 1890.3333333333333, "rapid_games": 31199, "rating_deviation": 45, "explanation": "Black handles the opening purposefully, seizing queenside space with 4...b5 and 5...c4. When White blunders with 16. Be4, Black spots the tactical sequence 16...Nxg2+ and 17...fxe4 cleanly. From there, Black demonstrates good board vision and tactical conversion, invading effectively with 20...Nb4, 21...Nbd3, and setting up the unstoppable mating attack with 24...Qb7 and 25...Nh3+." }, { "game": "game_06", "color": "white", "actual": 1112, "predicted": 1250, "error": 138, "absolute_error": 138, "low_80": 1100, "high_80": 1400, "interval_covers": true, "loo_baseline": 1917.7222222222222, "rapid_games": 3821, "rating_deviation": 45, "explanation": "White's opening with 4. Bd3 and 6. Bb1 was bizarre and passive, followed by a crude kingside attack (7. Qc2, 8. h4). White consistently missed tactics, leaving the a1-rook undefended to 18...Ncxb3 and 19...Nxa1, and inexplicably handing over the other rook with 25. Rf1? into Black's bishop. King safety was completely neglected, allowing a swift collapse." }, { "game": "game_06", "color": "black", "actual": 1382, "predicted": 1450, "error": 68, "absolute_error": 68, "low_80": 1300, "high_80": 1600, "interval_covers": true, "loo_baseline": 1917.7222222222222, "rapid_games": 5055, "rating_deviation": 45, "explanation": "Black handled the defense cleanly against White's crude attack with 7...g6 and 8...h5. Black promptly seized central initiative, finding strong tactical strikes like 13...Nd4, 17...Nc5, and 18...Ncxb3 winning material. When presented with further gifts (25...Bxf1), Black converted effortlessly and finished the game with an accurate and decisive mating sequence." }, { "game": "game_07", "color": "white", "actual": 1862, "predicted": 1550, "error": -312, "absolute_error": 312, "low_80": 1400, "high_80": 1700, "interval_covers": false, "loo_baseline": 1864.3333333333333, "rapid_games": 7581, "rating_deviation": 45, "explanation": "White handled the opening solidly and logically, deploying harmonious queenside play with moves like 10. b3 and 17. b4. However, after parrying the mate threat on g2 with 21. Bf3, White committed a critical tactical oversight by missing the double attack 21...Bc2, which cleanly lost the rook on b3." }, { "game": "game_07", "color": "black", "actual": 1593, "predicted": 1650, "error": 57, "absolute_error": 57, "low_80": 1500, "high_80": 1800, "interval_covers": true, "loo_baseline": 1864.3333333333333, "rapid_games": 13102, "rating_deviation": 45, "explanation": "Black demonstrated sound opening knowledge in the Slav setup and followed classical strategic principles with patient piece development. The transition to attack with 20...Qg6 set up an effective threat, and Black immediately spotted and executed the decisive winning fork with 21...Bc2, wrapping up the win cleanly." }, { "game": "game_08", "color": "white", "actual": 1310, "predicted": 1250, "error": -60, "absolute_error": 60, "low_80": 1100, "high_80": 1400, "interval_covers": true, "loo_baseline": 1884.0, "rapid_games": 4221, "rating_deviation": 45, "explanation": "White chose an overly passive opening setup (1. Nc3, 2. d3, 3. e3) and struggled with pins and tactical awareness. Allowing 9...e4 caused immediate tension, and White missed Black's straightforward fork 14...c5, losing the bishop on b4. Later, White walked into another pinned knight with 21. Kg2 and crumbled to Black's simple knight coordination (21...Ng5 and 24...Nf3+). These oversights reflect typical lower-intermediate rapid play." }, { "game": "game_08", "color": "black", "actual": 1791, "predicted": 1500, "error": -291, "absolute_error": 291, "low_80": 1350, "high_80": 1650, "interval_covers": false, "loo_baseline": 1884.0, "rapid_games": 4152, "rating_deviation": 49, "explanation": "Black played a very clean and direct game. They staked out central control early and repeatedly exploited tactical motifs. Black correctly struck with 9...e4 to open the pin on f3, found the winning pawn fork 14...c5 to net a piece, and traded queens favorably on b6. The finish was precise, exploiting the pinned f3 knight via 20...Bg4 and 21...Ng5, concluding with the decisive 24...Nf3+ fork." }, { "game": "game_09", "color": "white", "actual": 2193, "predicted": 1900, "error": -293, "absolute_error": 293, "low_80": 1750, "high_80": 2050, "interval_covers": false, "loo_baseline": 1813.5, "rapid_games": 4525, "rating_deviation": 45, "explanation": "White chose a sensible fianchetto setup against Black's Benoni structure but drifted passively with moves like 8. a3 and 13. Qc1. White allowed Black to dominate the queenside, ultimately missing the tactical sequence 21. Qxe2 Bxc3 22. bxc3 Qxa3 which dropped a key pawn. White tried to create counterplay with e5 and f5, but Black's pieces dominated, leading to an indefensible mating net after 41. Kf1 f3." }, { "game": "game_09", "color": "black", "actual": 2177, "predicted": 2050, "error": -127, "absolute_error": 127, "low_80": 1900, "high_80": 2200, "interval_covers": true, "loo_baseline": 1813.5, "rapid_games": 8936, "rating_deviation": 46, "explanation": "Black demonstrated high thematic understanding of the Benoni/King's Indian structure, executing the queenside expansion with ...b5 and purposeful knight maneuvers (...Ng4-e5-c4). Black accurately calculated the tactical pawn grab 21...Bxc3 22. bxc3 Qxa3, then infiltrated effectively with 25...Rb2 and 26...Qd2. The technical conversion was precise, cleanly trading into a winning rook endgame and setting up the unstoppable mating net with 41...f3." }, { "game": "game_10", "color": "white", "actual": 2689, "predicted": 2050, "error": -639, "absolute_error": 639, "low_80": 1900, "high_80": 2200, "interval_covers": false, "loo_baseline": 1767.9444444444443, "rapid_games": 9005, "rating_deviation": 45, "explanation": "White displays fluent Catalan/English opening understanding and sound piece coordination through move 22. However, White gets overly ambitious on the queenside with 23. Qc7 and 24. Nxa7?, missing the decisive double attack 24...Qc2! which targets both d1 and b2. After failing to calculate Black's tactical counterplay, White quickly collapses against Black's mating threats." }, { "game": "game_10", "color": "black", "actual": 2501, "predicted": 2150, "error": -351, "absolute_error": 351, "low_80": 2000, "high_80": 2300, "interval_covers": false, "loo_baseline": 1767.9444444444443, "rapid_games": 3141, "rating_deviation": 45, "explanation": "Black handles the opening and middlegame with high positional clarity, smoothly mobilizing pieces and redirecting them toward White's kingside (17...Bd6, 18...Rf8, 19...Qf7). When White overextends, Black spots the sharp tactical shot 24...Qc2! and follows up with clinical accuracy (25...Ba6, 27...Ne8!), wrapping up the game forcefully without giving White any counterchances." } ], "cost_usd": 0.04474575, "reasoning_effort": "medium", "max_tokens": 8192, "run_start": "2026-10-11", "run_end": "2026-10-11", "display_name": "Gemini 3.8 Flash" }, { "model": "openai/gpt-6.1-sol", "summary": { "games": 10, "player_estimates": 20, "actual_rating_range": [ 1112, 2718 ], "mae": 298.35, "rmse": 392.21926265801886, "signed_bias": -228.15, "pearson": 0.8314309410613463, "spearman": 0.8166482929480172, "within": { "100": { "count": 6, "fraction": 0.3 }, "200": { "count": 9, "fraction": 0.45 }, "300": { "count": 13, "fraction": 0.65 } }, "interval_80_coverage": 0.7, "interval_80_mean_width": 680.0, "game_mean_rating_mae": 261.55, "mae_game_bootstrap_95_interval": [ 169.43875, 450.91374999999977 ], "baseline_1500_mae": 506.05, "baseline_leave_one_game_out_mean_mae": 471.5666666666666, "games_beating_loo_baseline": 7, "history_count_range": [ 1136, 31199 ], "max_profile_rating_deviation": 59 }, "predictions": [ { "game": "game_01", "color": "white", "actual": 1363, "predicted": 1300, "error": -63, "absolute_error": 63, "low_80": 950, "high_80": 1650, "interval_covers": true, "loo_baseline": 1916.1666666666667, "rapid_games": 4696, "rating_deviation": 46, "explanation": "White found the effective 15.Nd5–16.Nf6+–17.Nxd7 queen-winning tactic and coordinated the central passed pawn with checks on moves 33–39. However, the opening involved repeated piece moves, 31.e5 conceded the exchange without clear necessity, and 40.Qb8+ overlooked the second rook's ability to defeat the promotion combination. This mixture of tactical awareness and incomplete calculation suggests an intermediate club-level Lichess rapid rating." }, { "game": "game_01", "color": "black", "actual": 1159, "predicted": 1300, "error": 141, "absolute_error": 141, "low_80": 950, "high_80": 1700, "interval_covers": true, "loo_baseline": 1916.1666666666667, "rapid_games": 3260, "rating_deviation": 45, "explanation": "Black's aggressive pawn expansion left the king exposed, and 15...Nxg5 missed the queen-winning knight fork. Nevertheless, 19...Nd4 and the subsequent exchanges showed useful defensive activity. The sequence 27...e4 and 30...Bxg3 created an exchange-winning opportunity, while 32...h4 and 38...h3 supplied counterplay. Black also correctly handled 39...Re8 and the final liquidation. The substantial opening oversight limits confidence despite resourceful later play." }, { "game": "game_02", "color": "white", "actual": 1575, "predicted": 1250, "error": -325, "absolute_error": 325, "low_80": 950, "high_80": 1600, "interval_covers": true, "loo_baseline": 1834.888888888889, "rapid_games": 1928, "rating_deviation": 45, "explanation": "White develops sensibly through 7.O-O and shows attacking intent with 15.Qd3 and 16.Bh6. However, 9.c3 permits the loss of e4, while 18.h4 sheds another pawn without sufficient compensation. The queen exchange with 20.Qh3 leaves a vulnerable position, and 22.Bd3–24.Bc2 fails to contain the rook invasion. Finally, 25.Bg5 overlooks mate in one. These cumulative defensive shortcomings outweigh the reasonable opening development." }, { "game": "game_02", "color": "black", "actual": 2410, "predicted": 1650, "error": -760, "absolute_error": 760, "low_80": 1250, "high_80": 2000, "interval_covers": false, "loo_baseline": 1834.888888888889, "rapid_games": 2936, "rating_deviation": 45, "explanation": "Black combines sound development with concrete play: 9...Nxe4 wins a pawn, 13...Na5–14...Nc4 establishes an active knight, and 15...g6 meets White's attacking threat. The sequence 18...Bxh4–20...Qxh3 gains another pawn and simplifies favorably. After 21...Re2, Black exploits the second rank and coordinates the rook with the b7 bishop for 25...Rh1#. The execution is convincing, although White's limited resistance leaves substantial uncertainty." }, { "game": "game_03", "color": "white", "actual": 1782, "predicted": 1800, "error": 18, "absolute_error": 18, "low_80": 1450, "high_80": 2100, "interval_covers": true, "loo_baseline": 1851.6666666666667, "rapid_games": 18588, "rating_deviation": 45, "explanation": "White showed sustained tactical awareness: 15.Ng6 and 16.Bxd6 exploit the fork on e7, while 18.Be7–19.Bxf6 secures extra material. Active queen play and 34.Ne7+ convert pressure into further gains; 37.Rd8 and 44.h5 finish efficiently. The time-consuming 8.h4–11.Nh3 maneuver and relatively straightforward conversion temper the estimate, but the overall coordination suggests an experienced intermediate player." }, { "game": "game_03", "color": "black", "actual": 1901, "predicted": 1500, "error": -401, "absolute_error": 401, "low_80": 1200, "high_80": 1850, "interval_covers": false, "loo_baseline": 1851.6666666666667, "rapid_games": 1366, "rating_deviation": 50, "explanation": "Black developed coherently and sought counterplay with 27...Rac8, 28...Rc4 and 31...f4. However, 14...Bd6–15...Rd8 allows the tactical loss of a minor piece; 16...Kf7 recognizes the e7 fork without adequately resolving the position. Later, 33...Nc8 overlooks 34.Ne7+, losing an exchange. These repeated tactical vulnerabilities outweigh the reasonable opening structure and active resistance, though one game leaves substantial uncertainty." }, { "game": "game_04", "color": "white", "actual": 2508, "predicted": 1750, "error": -758, "absolute_error": 758, "low_80": 1400, "high_80": 2050, "interval_covers": false, "loo_baseline": 1765.9444444444443, "rapid_games": 1324, "rating_deviation": 47, "explanation": "White showed reasonable opening understanding and a coherent queenside plan with 14.b4, 16.a4 and 18.Nc5. However, repeated bishop moves cost time, while 21.c4 and 22.b5 prioritized expansion over Black's central threats and kingside attack. The forcing 24.Bb4 overlooked the mating sequence. This suggests an intermediate player with useful strategic ideas but inconsistent tactical vigilance." }, { "game": "game_04", "color": "black", "actual": 2718, "predicted": 2000, "error": -718, "absolute_error": 718, "low_80": 1650, "high_80": 2300, "interval_covers": false, "loo_baseline": 1765.9444444444443, "rapid_games": 1136, "rating_deviation": 59, "explanation": "Black coordinated opposite-side castling and central control effectively: 13...O-O-O, 14...Kb8, 15...Rc8 and 16...Ne4 formed a coherent setup. The subsequent ...f5 and ...g5 supported an energetic attack. Particularly strong were 22...Nd4 and the forcing finish 24...Ne2+, 25...Qg1+ and 26...Nf2#. The game shows strong tactical and positional coordination, although one short attacking game leaves substantial rating uncertainty." }, { "game": "game_05", "color": "white", "actual": 1120, "predicted": 1200, "error": 80, "absolute_error": 80, "low_80": 900, "high_80": 1550, "interval_covers": true, "loo_baseline": 1890.3333333333333, "rapid_games": 5115, "rating_deviation": 45, "explanation": "White shows basic development and central ambitions with 9.Nbd2, 10.e5 and 14.Bb2, but spends many opening tempi relocating the bishop and neglects king safety. The sequence 15.c4–17.Kf1 allows Black to win a bishop, while 23.Nd4 exposes the rook to capture. Finally, 27.Qxd3 overlooks the discovered-check sequence winning the queen. These cumulative tactical and coordination problems suggest a lower-intermediate established rating." }, { "game": "game_05", "color": "black", "actual": 1867, "predicted": 1650, "error": -217, "absolute_error": 217, "low_80": 1300, "high_80": 2000, "interval_covers": true, "loo_baseline": 1890.3333333333333, "rapid_games": 31199, "rating_deviation": 45, "explanation": "Black exploits White's opening tempi with 4...b5 and 5...c4, then develops coherently. The sequence 15...Nf4, 16...Nxg2+ and 17...fxe4 demonstrates tactical awareness. Knight penetration with 20...Nb4 and 21...Nbd3, followed by 22...Bxe4, creates further material gains. The finish with 25...Nh3+ and 27...Be4+ shows good coordination and discovered-check recognition. However, White provides little resistance, limiting how strongly this game establishes Black's rating." }, { "game": "game_06", "color": "white", "actual": 1112, "predicted": 1200, "error": 88, "absolute_error": 88, "low_80": 850, "high_80": 1550, "interval_covers": true, "loo_baseline": 1917.7222222222222, "rapid_games": 3821, "rating_deviation": 45, "explanation": "The slow 4.Bd3–6.Bb1 setup and repeated queen moves suggest limited opening efficiency. White pursued a coherent kingside attack with 13.Ng5 and 19.f5, but underestimated material losses after 18...Ncxb3. The decisive 25.Rf1 allowed ...Bxf1, followed by loss of the knight on f6. Finally, 30.Qxb7 overlooked the mating attack. Tactical and defensive awareness appear substantially weaker than attacking ambition." }, { "game": "game_06", "color": "black", "actual": 1382, "predicted": 1550, "error": 168, "absolute_error": 168, "low_80": 1200, "high_80": 1900, "interval_covers": true, "loo_baseline": 1917.7222222222222, "rapid_games": 5055, "rating_deviation": 45, "explanation": "Repeated bishop moves between moves 3 and 15 were inefficient, but Black showed useful central understanding with 13...Nd4 and tactical awareness through 17...Nc5 and 18...Ncxb3. The sequence 25...Bxf1, 26...Nxg5 and 27...Qxf6 exploited both the exposed rook and queen diagonal. Black then converted actively with 29...Rac8 and 30...Rc1+, finding mate. Strong exploitation of accessible opportunities, though this game provides limited evidence of deeper positional skill." }, { "game": "game_07", "color": "white", "actual": 1862, "predicted": 1650, "error": -212, "absolute_error": 212, "low_80": 1250, "high_80": 2000, "interval_covers": true, "loo_baseline": 1864.3333333333333, "rapid_games": 7581, "rating_deviation": 45, "explanation": "White developed sensibly, centralized with 15.Ne5, and pursued a coherent queenside expansion with 17.b4–20.axb5. However, the rook placement on b3 became tactically vulnerable. After 21...Bc2, 22.Bh5 counterattacked the queen but overlooked that 22...Bxd1 23.Bxg6 Bxb3 loses a full rook. This suggests reasonable strategic understanding with imperfect tactical calculation; the short game provides no endgame evidence." }, { "game": "game_07", "color": "black", "actual": 1593, "predicted": 1800, "error": 207, "absolute_error": 207, "low_80": 1400, "high_80": 2150, "interval_covers": true, "loo_baseline": 1864.3333333333333, "rapid_games": 13102, "rating_deviation": 45, "explanation": "Black showed coherent development and central play with 11...Ne4, then redirected the queen effectively through 18...Qe6 and 20...Qg6. The defended fork 21...Bc2 exploited White's exposed queen and rook, and 22...Bxd1 followed by 23...Bxb3 correctly answered White's counterattack. Some opening choices were unambitious, and one successful tactical sequence cannot establish a precise rating, but the overall coordination suggests an experienced club-level player." }, { "game": "game_08", "color": "white", "actual": 1310, "predicted": 1250, "error": -60, "absolute_error": 60, "low_80": 950, "high_80": 1550, "interval_covers": true, "loo_baseline": 1884.0, "rapid_games": 4221, "rating_deviation": 45, "explanation": "White developed and coordinated rooks sensibly with 13.Rad1 and 14.Rfe1, but repeatedly overlooked tactical relationships. 15.Bxc5 allowed a bishop to be lost for a pawn, with 16.Ne5 and 17.Bxe4 providing insufficient compensation. After 21...Ng5, 22.Rd4 allowed another minor piece to fall because the knight protected f3. Finally, 24.h4 permitted a decisive knight fork. These cumulative errors suggest modest tactical awareness rather than merely one unfortunate blunder." }, { "game": "game_08", "color": "black", "actual": 1791, "predicted": 1650, "error": -141, "absolute_error": 141, "low_80": 1300, "high_80": 2000, "interval_covers": true, "loo_baseline": 1884.0, "rapid_games": 4152, "rating_deviation": 49, "explanation": "Black showed coherent central play with 9...e4 and 14...c5, correctly exploiting 15.Bxc5 with 15...Ndxc5. The queen exchange on moves 16–18 simplified a material advantage. Subsequently, 19...f6, 20...Bg4 and 21...Ng5 built pressure, while 22...Bxf3+ and 23...Be4 established the decisive 24...Nf3+ fork. This was consistently competent play, although White's limited resistance and the absence of an endgame leave substantial rating uncertainty." }, { "game": "game_09", "color": "white", "actual": 2193, "predicted": 1900, "error": -293, "absolute_error": 293, "low_80": 1600, "high_80": 2200, "interval_covers": true, "loo_baseline": 1813.5, "rapid_games": 4525, "rating_deviation": 45, "explanation": "White showed coherent fianchetto development and sought active compensation with 23.Ra1, 25.Rxa7 and 28.e5–30.f5. However, 19.f4 allowed Black's simplifying sequence through 23...Qxc3, leaving lasting structural and material disadvantages. The rook activity after 35.Rc1 was purposeful, but 38.Rc8 and 39.R8xc5 underestimated the kingside mating attack. This suggests an experienced player with uneven positional judgment and defensive calculation." }, { "game": "game_09", "color": "black", "actual": 2177, "predicted": 2100, "error": -77, "absolute_error": 77, "low_80": 1800, "high_80": 2400, "interval_covers": true, "loo_baseline": 1813.5, "rapid_games": 8936, "rating_deviation": 46, "explanation": "Black demonstrated coherent Benoni counterplay with 7...Na6, 9...Nc7 and 10...b5, followed by effective knight rerouting through 13...Ng4 and 15...Nc4. The sequence 19...Nd4–23...Qxc3 exploited White's structure, while 25...Rb2 maintained activity. Resourceful defence with 29...Bb1 and 32...Qf4+ led into a favorable endgame. The coordinated finish, 38...Rdd2 through 41...f3, supports a strong club-level estimate." }, { "game": "game_10", "color": "white", "actual": 2689, "predicted": 1900, "error": -789, "absolute_error": 789, "low_80": 1550, "high_80": 2250, "interval_covers": false, "loo_baseline": 1767.9444444444443, "rapid_games": 9005, "rating_deviation": 45, "explanation": "White showed coherent opening understanding with 9.d4, 11.Rc1 and 13.Nd4, followed by active queenside play. However, the queen manoeuvres on moves 15–21 achieved little, and 24.Nxa7 underestimated Black's counterplay: 24...Qc2 and 25...Ba6 led to an exchange loss. This suggests a competent intermediate player, but one tactical sequence provides limited evidence about an established rating." }, { "game": "game_10", "color": "black", "actual": 2501, "predicted": 2050, "error": -451, "absolute_error": 451, "low_80": 1700, "high_80": 2400, "interval_covers": false, "loo_baseline": 1767.9444444444443, "rapid_games": 3141, "rating_deviation": 45, "explanation": "Black developed harmoniously, accepted the structural change with 16...fxe6, and coordinated queen and rook through 18...Rf8 and 21...Qg6. The sequence 23...Bc5, 24...Qc2 and 25...Ba6 exploited White's loose coordination to win an exchange. Finally, 27...Ne8 saved the attacked knight while attacking White's queen. This was convincing tactical and positional play, though no substantial endgame or difficult defence was tested." } ], "cost_usd": 0.145124, "reasoning_effort": "medium", "max_tokens": 8192, "run_start": "2026-10-11", "run_end": "2026-10-11", "display_name": "GPT-6.1 Sol" }, { "model": "google/gemini-3.7-flash", "summary": { "games": 10, "player_estimates": 20, "actual_rating_range": [ 1112, 2718 ], "mae": 300.55, "rmse": 376.6376906258852, "signed_bias": -208.15, "pearson": 0.8878377070397115, "spearman": 0.8578997069620994, "within": { "100": { "count": 5, "fraction": 0.25 }, "200": { "count": 7, "fraction": 0.35 }, "300": { "count": 13, "fraction": 0.65 } }, "interval_80_coverage": 0.3, "interval_80_mean_width": 300.0, "game_mean_rating_mae": 252.85, "mae_game_bootstrap_95_interval": [ 197.89875, 424.50499999999994 ], "baseline_1500_mae": 506.05, "baseline_leave_one_game_out_mean_mae": 471.5666666666666, "games_beating_loo_baseline": 7, "history_count_range": [ 1136, 31199 ], "max_profile_rating_deviation": 59 }, "predictions": [ { "game": "game_01", "color": "white", "actual": 1363, "predicted": 1300, "error": -63, "absolute_error": 63, "low_80": 1150, "high_80": 1450, "interval_covers": true, "loo_baseline": 1916.1666666666667, "rapid_games": 4696, "rating_deviation": 46, "explanation": "White spotted a sharp tactical royal fork with 16. Nf6+ to win Black's queen and converted the advantage reasonably well into a dominant passed e-pawn. However, the opening was passive and White suffered an extreme tactical hallucination with 40. Qb8+??, simply hanging the queen for nothing and instantly losing. This mix of finding basic tactics alongside severe one-move blunders is typical around 1300 Lichess rapid." }, { "game": "game_01", "color": "black", "actual": 1159, "predicted": 1250, "error": 91, "absolute_error": 91, "low_80": 1100, "high_80": 1400, "interval_covers": true, "loo_baseline": 1916.1666666666667, "rapid_games": 3260, "rating_deviation": 45, "explanation": "Black overextended their kingside pawns early (4...f5, 9...g5, 12...h5) while neglecting king safety, leading directly to the decisive blunder 15...Nxg5?? allowing 16. Nf6+. After losing the queen, Black kept fighting with active piece play (19...Nd4) and advanced the h-pawn, ultimately winning entirely due to White's gift on move 40. The overall lack of prophylactic awareness and structural safety points to the mid-1200s." }, { "game": "game_02", "color": "white", "actual": 1575, "predicted": 1500, "error": -75, "absolute_error": 75, "low_80": 1350, "high_80": 1650, "interval_covers": true, "loo_baseline": 1834.888888888889, "rapid_games": 1928, "rating_deviation": 45, "explanation": "White navigates a standard Ruy Lopez opening reasonably well until needlessly losing the e4 pawn (9. c3 Nxe4). After decent central play, White blunders tactically with 18. h4? and allows Black's pieces to dominate. White then misses the back-rank/h-file mating threat, walking into 24...Rxh3 and playing 25. Bg5?? allowing mate in one (25...Rh1#). This combination of basic opening familiarity and major tactical blindness is characteristic of around 1500 Lichess rapid." }, { "game": "game_02", "color": "black", "actual": 2410, "predicted": 1650, "error": -760, "absolute_error": 760, "low_80": 1500, "high_80": 1800, "interval_covers": false, "loo_baseline": 1834.888888888889, "rapid_games": 2936, "rating_deviation": 45, "explanation": "Black plays cleanly and efficiently throughout. They accurately punish White's pawn loose play with 9...Nxe4, comfortably handle the position with 14...Nc4 and 15...g6, and methodically convert tactical opportunities (18...Bxh4, 21...Re2, 24...Rxh3). Black avoids mistakes and cleanly finds 25...Rh1# to end the game. Such disciplined punishment of basic mistakes without counter-blunders indicates a solid intermediate player around 1650 Lichess." }, { "game": "game_03", "color": "white", "actual": 1782, "predicted": 1700, "error": -82, "absolute_error": 82, "low_80": 1550, "high_80": 1850, "interval_covers": true, "loo_baseline": 1851.6666666666667, "rapid_games": 18588, "rating_deviation": 45, "explanation": "White demonstrates solid tactical alertness and smooth technical conversion. After an aggressive flank plan (8. h4), White systematically exploited Black's weaknesses around e5 and the d-file (14. Ne5, 23. Nxd5, 24. Qb3). White never lost control, finding direct winning tactics like 34. Ne7+ and 37. Rd8 to finish ruthlessly." }, { "game": "game_03", "color": "black", "actual": 1901, "predicted": 1500, "error": -401, "absolute_error": 401, "low_80": 1350, "high_80": 1650, "interval_covers": false, "loo_baseline": 1851.6666666666667, "rapid_games": 1366, "rating_deviation": 50, "explanation": "Black opened reasonably in the Dutch but struggled under pressure. Moves like 13...d5 created severe dark-square concessions, and 15...Rd8 overlooked the tactic 16. Bxd6. Further oversights like 23...Nd7 (allowing 24. Qb3) and 36...Rg7 (allowing the pin 37. Rd8) show deficiencies in king safety and tactical defense." }, { "game": "game_04", "color": "white", "actual": 2508, "predicted": 1850, "error": -658, "absolute_error": 658, "low_80": 1700, "high_80": 2000, "interval_covers": false, "loo_baseline": 1765.9444444444443, "rapid_games": 1324, "rating_deviation": 47, "explanation": "White demonstrates solid opening knowledge in the Exchange Caro-Kann and formulates a coherent opposite-side castling attack with b4-a4-b5. However, White neglects king safety on the open f/g files and tactical vulnerabilities on the a7-g1 diagonal. The critical blunder 24. Bb4?? completely misses the classic smothered mate combination starting with the double check 24...Ne2+." }, { "game": "game_04", "color": "black", "actual": 2718, "predicted": 1950, "error": -768, "absolute_error": 768, "low_80": 1800, "high_80": 2100, "interval_covers": false, "loo_baseline": 1765.9444444444443, "rapid_games": 1136, "rating_deviation": 59, "explanation": "Black plays cleanly throughout, utilizing the standard plan with ...Qc7, ...O-O-O, and central clamping with ...Ne4 and ...f5. Black accurately counters White's flank attack by undermining the kingside with ...g5 and rapidly activating the knight and queen. Black spotlessly calculates and executes the classic smothered mate finish with 24...Ne2+ and 25...Qg1+." }, { "game": "game_05", "color": "white", "actual": 1120, "predicted": 1400, "error": 280, "absolute_error": 280, "low_80": 1250, "high_80": 1550, "interval_covers": false, "loo_baseline": 1890.3333333333333, "rapid_games": 5115, "rating_deviation": 45, "explanation": "White struggles with opening coordination, moving the light-squared bishop four times by move 12 (Bc4-d3-e2-d3) without achieving central control. Move 13. b3 weakens c3, and 16. Be4 critically blunders the g2 pawn with check, losing a piece to 16...Nxg2+ and 17...fxe4. White's king position collapses immediately, offering minimal resistance to Black's finishing attack." }, { "game": "game_05", "color": "black", "actual": 1867, "predicted": 1600, "error": -267, "absolute_error": 267, "low_80": 1450, "high_80": 1750, "interval_covers": false, "loo_baseline": 1890.3333333333333, "rapid_games": 31199, "rating_deviation": 45, "explanation": "Black handles the opening cleanly with queenside space expansion (...b5, ...c4, ...Bb7). Black spots decisive tactical opportunities, infiltrating with 15...Nf4 and punishing White's blunder with 16...Nxg2+ and 17...fxe4. The conversion is swift and accurate, utilizing knight outposts on d3 and executing a forceful mating sequence (24...Qb7, 25...Nh3+, 27...Be4+)." }, { "game": "game_06", "color": "white", "actual": 1112, "predicted": 1350, "error": 238, "absolute_error": 238, "low_80": 1200, "high_80": 1500, "interval_covers": false, "loo_baseline": 1917.7222222222222, "rapid_games": 3821, "rating_deviation": 45, "explanation": "White adopts an artificial setup with 4. Bd3, 6. Bb1, and 7. Qc2, neglecting central development and king safety. Tactics were severely miscalculated: 18. Qe3 walked into 18...Ncxb3, losing material, and 25. Rf1 surrendered an exchange under the pin. White continued grabbing pawns with 28. Qxc7+ while allowing a swift mating attack on the uncastled king." }, { "game": "game_06", "color": "black", "actual": 1382, "predicted": 1550, "error": 168, "absolute_error": 168, "low_80": 1400, "high_80": 1700, "interval_covers": false, "loo_baseline": 1917.7222222222222, "rapid_games": 5055, "rating_deviation": 45, "explanation": "Black handled White's unusual opening calmly, neutralizing the kingside battery with 7...g6 and 8...h5. Black systematically seized control of key outposts with 13...Nd4 and 17...Nc5, accurately winning material via 18...Ncxb3 and 19...Nxa1. The conversion was smooth and decisive, pinning the knight on f6 and concluding with an efficient mating sequence (29...Rac8 to 31...Qf2#)." }, { "game": "game_07", "color": "white", "actual": 1862, "predicted": 1650, "error": -212, "absolute_error": 212, "low_80": 1500, "high_80": 1800, "interval_covers": false, "loo_baseline": 1864.3333333333333, "rapid_games": 7581, "rating_deviation": 45, "explanation": "White demonstrates solid opening understanding and natural development with sensible queenside expansion (10. b3, 17. b4, 18. Rb3). However, White misses Black's tactical threat after 20...Qg6, playing 21. Bf3?? which immediately allows the devastating fork 21...Bc2, losing a full rook on b3. This level of positional play combined with a single critical one-move oversight is characteristic of the mid-1600s range." }, { "game": "game_07", "color": "black", "actual": 1593, "predicted": 1700, "error": 107, "absolute_error": 107, "low_80": 1550, "high_80": 1850, "interval_covers": true, "loo_baseline": 1864.3333333333333, "rapid_games": 13102, "rating_deviation": 45, "explanation": "Black handles the Slav setup cleanly with harmonious piece development and comfortable piece trades (12...Bxe4, 13...Bxa3). Black creates active counterplay with 18...Qe6 and 20...Qg6, creating a mating threat, and immediately capitalizes on White's tactical blind spot with the winning fork 21...Bc2!, converting cleanly to force resignation on move 24." }, { "game": "game_08", "color": "white", "actual": 1310, "predicted": 1350, "error": 40, "absolute_error": 40, "low_80": 1200, "high_80": 1500, "interval_covers": true, "loo_baseline": 1884.0, "rapid_games": 4221, "rating_deviation": 45, "explanation": "White adopts a passive, disjointed setup (1. Nc3, 2. d3, 3. e3) and concedes central control. White fails to anticipate basic tactics, walking into the pin with 8. O-O, entirely missing the pawn fork 14... c5 which dropped a minor piece, and then allowing a decisive pin/fork motif with 20... Bg4 and 21... Ng5, culminating in 24... Nf3+." }, { "game": "game_08", "color": "black", "actual": 1791, "predicted": 1550, "error": -241, "absolute_error": 241, "low_80": 1400, "high_80": 1700, "interval_covers": false, "loo_baseline": 1884.0, "rapid_games": 4152, "rating_deviation": 49, "explanation": "Black demonstrates good fundamental understanding by claiming the center and developing smoothly. Black spots tactical flaws immediately, exploiting the pin on the f3-knight with 9... e4 and winning a clean piece with the 14... c5 fork. Black converts the advantage efficiently by relentlessly targeting the pinned f3-knight with 20... Bg4 and 21... Ng5." }, { "game": "game_09", "color": "white", "actual": 2193, "predicted": 1850, "error": -343, "absolute_error": 343, "low_80": 1700, "high_80": 2000, "interval_covers": false, "loo_baseline": 1813.5, "rapid_games": 4525, "rating_deviation": 45, "explanation": "White chose a playable King's Indian/Benoni setup (6. g3, 7. Bg2), but failed to counter Black's standard queenside expansion. White allowed Black's knights to dominate (14...Ne5, 15...Nc4) and conceded a clean pawn with 21. Qxe2 Bxc3 22. bxc3 Qxa3. The counterplay attempt with 25. Rxa7 was natural, but White could not contain Black's major piece activity along the second rank." }, { "game": "game_09", "color": "black", "actual": 2177, "predicted": 1950, "error": -227, "absolute_error": 227, "low_80": 1800, "high_80": 2100, "interval_covers": false, "loo_baseline": 1813.5, "rapid_games": 8936, "rating_deviation": 46, "explanation": "Black demonstrated strong opening and strategic understanding with ...Na6-c7 and the timely 10...b5 break. Tactically precise manoeuvres (14...Ne5, 15...Nc4) won the bishop pair and exploited White's queenside weaknesses with 21...Bxc3 and 22...Qxa3. Black maintained relentless pressure on the second rank (25...Rb2, 26...Qd2) and smoothly converted the rook endgame into an unstoppable mating net." }, { "game": "game_10", "color": "white", "actual": 2689, "predicted": 2050, "error": -639, "absolute_error": 639, "low_80": 1900, "high_80": 2200, "interval_covers": false, "loo_baseline": 1767.9444444444443, "rapid_games": 9005, "rating_deviation": 45, "explanation": "White demonstrates solid opening preparation in the English/Catalan setup and handles the early middlegame smoothly (1. c4 through 18. Qh4). However, White falters tactically with 24. Nxa7?, overlooking the fork 24...Qc2, and quickly collapses under Black's pressure without putting up stubborn defense." }, { "game": "game_10", "color": "black", "actual": 2501, "predicted": 2150, "error": -351, "absolute_error": 351, "low_80": 2000, "high_80": 2300, "interval_covers": false, "loo_baseline": 1767.9444444444443, "rapid_games": 3141, "rating_deviation": 45, "explanation": "Black plays high-quality positional chess, handling the opening and pawn structure seamlessly (8...exd5, 12...Nxc5, 16...fxe6). Once White errs, Black calculates and executes decisively with 24...Qc2, 25...Ba6, and the killer blow 27...Ne8!, showing strong tactical vigilance and conversion." } ], "cost_usd": 0.0394545, "reasoning_effort": "medium", "max_tokens": 8192, "run_start": "2026-10-11", "run_end": "2026-10-11", "display_name": "Gemini 3.7 Flash" }, { "model": "openai/gpt-6-astra", "summary": { "games": 10, "player_estimates": 20, "actual_rating_range": [ 1112, 2718 ], "mae": 302.35, "rmse": 393.4157470158001, "signed_bias": -198.15, "pearson": 0.8242017235788925, "spearman": 0.8249115337448945, "within": { "100": { "count": 5, "fraction": 0.25 }, "200": { "count": 8, "fraction": 0.4 }, "300": { "count": 14, "fraction": 0.7 } }, "interval_80_coverage": 0.65, "interval_80_mean_width": 622.5, "game_mean_rating_mae": 256.55, "mae_game_bootstrap_95_interval": [ 176.94625, 450.9637499999998 ], "baseline_1500_mae": 506.05, "baseline_leave_one_game_out_mean_mae": 471.5666666666666, "games_beating_loo_baseline": 7, "history_count_range": [ 1136, 31199 ], "max_profile_rating_deviation": 59 }, "predictions": [ { "game": "game_01", "color": "white", "actual": 1363, "predicted": 1350, "error": -13, "absolute_error": 13, "low_80": 1000, "high_80": 1700, "interval_covers": true, "loo_baseline": 1916.1666666666667, "rapid_games": 4696, "rating_deviation": 46, "explanation": "White showed purposeful development with 10.O-O-O and 11.h4, and spotted the queen-winning sequence 15.Nd5–17.Nxd7. The subsequent exchanges were sensible, but conversion was inconsistent: 31.e5 allowed ...Bxe1, while 40.Qb8+ overlooked the decisive recapture sequence involving both black rooks. This suggests some tactical competence but unreliable calculation; one game leaves substantial uncertainty." }, { "game": "game_01", "color": "black", "actual": 1159, "predicted": 1400, "error": 241, "absolute_error": 241, "low_80": 1050, "high_80": 1750, "interval_covers": true, "loo_baseline": 1916.1666666666667, "rapid_games": 3260, "rating_deviation": 45, "explanation": "Black's ambitious kingside expansion showed a coherent plan, but 15...Nxg5 missed a straightforward queen-winning fork. Despite that major oversight, Black continued resourcefully with 19...Nd4, coordinated the rooks through 25...Rbf8, exploited the bishop diagonal with 30...Bxg3 and 31...Bxe1, and created counterplay with the h-pawn. The accurate response to 40.Qb8+ contrasts with the earlier tactical lapse." }, { "game": "game_02", "color": "white", "actual": 1575, "predicted": 1350, "error": -225, "absolute_error": 225, "low_80": 1050, "high_80": 1650, "interval_covers": true, "loo_baseline": 1834.888888888889, "rapid_games": 1928, "rating_deviation": 45, "explanation": "White showed reasonable development and attacking ideas with 15.Qd3 and 16.Bh6, but 9.c3 allowed the undefended e4 pawn to fall. The attack lacked sufficient calculation: 18.h4 lost another pawn, and the queen exchange left serious weaknesses. The slow 23.a3 and especially 25.Bg5, allowing mate, suggest limited defensive awareness. There is no meaningful endgame evidence." }, { "game": "game_02", "color": "black", "actual": 2410, "predicted": 1650, "error": -760, "absolute_error": 760, "low_80": 1300, "high_80": 2000, "interval_covers": false, "loo_baseline": 1834.888888888889, "rapid_games": 2936, "rating_deviation": 45, "explanation": "Black developed coherently, exploited 9...Nxe4, and found the useful regrouping 13...Na5–14...Nc4. The sequence 18...Bxh4 through the queen exchange converted White's attack into a material advantage. Active rook play with 21...Re2 and 23...Rb3 culminated in the bishop-supported 25...Rh1#. This was convincing practical play, although White's errors meant Black faced little demanding defence or endgame technique." }, { "game": "game_03", "color": "white", "actual": 1782, "predicted": 1850, "error": 68, "absolute_error": 68, "low_80": 1550, "high_80": 2150, "interval_covers": true, "loo_baseline": 1851.6666666666667, "rapid_games": 18588, "rating_deviation": 45, "explanation": "White showed tactical awareness with 16.Bxd6, when ...Rxd6 permits a royal fork, and purposeful piece coordination through 18.Be7 and 22.Nhf4. The sequence 33.Qf5–38.Rxg8+ exploited the exposed king convincingly. Opening play was less polished: 8.h4 followed by castling and repeated knight moves spent considerable time. Strong attacking execution suggests an experienced intermediate player, though this favorable position provides limited evidence about defence or endgames." }, { "game": "game_03", "color": "black", "actual": 1901, "predicted": 1650, "error": -251, "absolute_error": 251, "low_80": 1350, "high_80": 1950, "interval_covers": true, "loo_baseline": 1851.6666666666667, "rapid_games": 1366, "rating_deviation": 50, "explanation": "Black developed coherently and found 16...Kf7 rather than immediately falling for the fork after ...Rxd6. The central advances 11...c5 and 13...d5, followed by active rook play on moves 27–29, showed reasonable understanding. However, 22...e5 conceded central material, and 33...Nc8 failed to address White's discovered-check threats. The ensuing defence allowed decisive tactical gains. These mixed decisions suggest intermediate strength rather than a rating determined solely by the loss." }, { "game": "game_04", "color": "white", "actual": 2508, "predicted": 1800, "error": -708, "absolute_error": 708, "low_80": 1500, "high_80": 2100, "interval_covers": false, "loo_baseline": 1765.9444444444443, "rapid_games": 1324, "rating_deviation": 47, "explanation": "White shows coherent development and attacking intent: 14.b4, 15.Nb3 and 18.Nc5 coordinate queenside expansion, followed by 21.c4 and 22.b5. However, 19.dxc5 leaves Black's e4 knight entrenched, and White underestimates the central counterattack. Most notably, 24.Bb4 attacks the queen while overlooking the forcing double-check and smothered-mate sequence. This suggests intermediate strength, with substantial uncertainty from one short game." }, { "game": "game_04", "color": "black", "actual": 2718, "predicted": 2000, "error": -718, "absolute_error": 718, "low_80": 1700, "high_80": 2300, "interval_covers": false, "loo_baseline": 1765.9444444444443, "rapid_games": 1136, "rating_deviation": 59, "explanation": "Black combines sensible king safety and rook placement with a coherent central strategy: 16...Ne4, 17...f5 and 19...g5. The active 22...Nd4 gains time against the queen, and 23...Qxc5 creates the tactical opportunity. The precise 24...Ne2+ double check, 25...Qg1+ and 26...Nf2# demonstrate strong tactical awareness. The game provides little evidence about prolonged defence or endgame technique, so the established-rating estimate remains broad." }, { "game": "game_05", "color": "white", "actual": 1120, "predicted": 1300, "error": 180, "absolute_error": 180, "low_80": 1000, "high_80": 1600, "interval_covers": true, "loo_baseline": 1890.3333333333333, "rapid_games": 5115, "rating_deviation": 45, "explanation": "White developed pieces and sought central activity, but 2.Bc4 and repeated bishop moves conceded substantial time. The sequence 15.c4 Nf4 16.Be4 overlooked concrete threats, allowing ...Nxg2+ and ...fxe4. Later, 21.a3 permitted ...Nbd3, and 27.Qxd3 missed the discovered-check skewer. These suggest limited tactical checking, though errors after the position was already lost deserve less weight." }, { "game": "game_05", "color": "black", "actual": 1867, "predicted": 1650, "error": -217, "absolute_error": 217, "low_80": 1350, "high_80": 1950, "interval_covers": true, "loo_baseline": 1890.3333333333333, "rapid_games": 31199, "rating_deviation": 45, "explanation": "Black exploited the exposed bishop with 3...a6, 4...b5 and 5...c4, then completed development and castled. The tactical sequence 15...Nf4 through 18...Nf4 won material efficiently. Strong coordination continued with 20...Nb4, 21...Nbd3 and 22...Bxe4. The finishing 27...Be4+ and 29...Qxf3# showed accurate tactical vision, although White's cooperation limits evidence of higher-level positional or defensive skill." }, { "game": "game_06", "color": "white", "actual": 1112, "predicted": 1250, "error": 138, "absolute_error": 138, "low_80": 900, "high_80": 1600, "interval_covers": true, "loo_baseline": 1917.7222222222222, "rapid_games": 3821, "rating_deviation": 45, "explanation": "White showed coherent attacking ideas with 13.Ng5, 16.f4 and 22.Nd5–24.Nf6+, but neglected development and king safety. The slow 4.Bd3–6.Bb1 setup and subsequent pawn moves allowed central domination. Continuing the attack after 18...Ncxb3 conceded substantial material, while 25.Rf1 and 27.Ke1 overlooked further losses. The attacking coordination suggests some experience, but threat recognition was inconsistent." }, { "game": "game_06", "color": "black", "actual": 1382, "predicted": 1600, "error": 218, "absolute_error": 218, "low_80": 1250, "high_80": 1950, "interval_covers": true, "loo_baseline": 1917.7222222222222, "rapid_games": 5055, "rating_deviation": 45, "explanation": "Despite losing tempi with 9...Bf5–10...Bd7, Black developed an effective central initiative through 13...Nd4 and 17...Nc5. The captures on b3 and a1 exploited White's neglected queenside. Black handled the attack resourcefully with 22...Ne6 and 23...Ba6, then found 25...Bxf1 and 27...Qxf6. The forcing finish with 29...Rac8 and 30...Rc1+ was convincing, though many opportunities arose from straightforward opponent errors." }, { "game": "game_07", "color": "white", "actual": 1862, "predicted": 1550, "error": -312, "absolute_error": 312, "low_80": 1250, "high_80": 1850, "interval_covers": false, "loo_baseline": 1864.3333333333333, "rapid_games": 7581, "rating_deviation": 45, "explanation": "White develops coherently and shows a recognizable queenside plan with 13.Ba3 and 17–20.b4–b5. However, the rook on b3 and queen on d1 become vulnerable to a bishop fork, overlooked during that expansion. After 20...Qg6, 21.Bf3 addresses the mating threat but permits 21...Bc2. The resource 22.Bh5 shows tactical awareness, although 23...Bxb3 still loses a rook. No endgame evidence is available." }, { "game": "game_07", "color": "black", "actual": 1593, "predicted": 1750, "error": 157, "absolute_error": 157, "low_80": 1400, "high_80": 2050, "interval_covers": true, "loo_baseline": 1864.3333333333333, "rapid_games": 13102, "rating_deviation": 45, "explanation": "Black combines sound development with purposeful central play through 11...Ne4 and 15...Nd7. The sequence 18...Qe6 and 20...Qg6 creates coordinated threats against White’s king and loose pieces. Black spots 21...Bc2 and handles 22.Bh5 correctly with 22...Bxd1, followed by 23...Bxb3. This is convincing tactical play, but a short game with limited defensive or endgame demands leaves substantial rating uncertainty." }, { "game": "game_08", "color": "white", "actual": 1310, "predicted": 1350, "error": 40, "absolute_error": 40, "low_80": 1050, "high_80": 1650, "interval_covers": true, "loo_baseline": 1884.0, "rapid_games": 4221, "rating_deviation": 45, "explanation": "White developed and coordinated the rooks sensibly through move 14, but tactical awareness was inconsistent. 14.Rfe1 allowed the pawn fork 14...c5, costing a bishop for a pawn. Later, 22.Rd4 overlooked that ...Bxf3+ could not be answered by Kxf3, and 24.h4 allowed the decisive knight fork. These repeated oversights suggest a developing intermediate player rather than merely one isolated blunder." }, { "game": "game_08", "color": "black", "actual": 1791, "predicted": 1700, "error": -91, "absolute_error": 91, "low_80": 1400, "high_80": 2000, "interval_covers": true, "loo_baseline": 1884.0, "rapid_games": 4152, "rating_deviation": 49, "explanation": "Black established a strong center, used 9...e4 to gain activity, and developed coherently with ...Re8, ...Be6 and ...Nd7. The fork 14...c5 won material, while 16...Qb6 facilitated a favorable queen exchange. The sequence 20...Bg4, 21...Ng5 and 23...Be4 showed coordinated tactical play, culminating in 24...Nf3+. The estimate remains broad because White offered relatively straightforward opportunities and Black faced little sustained defensive pressure." }, { "game": "game_09", "color": "white", "actual": 2193, "predicted": 1900, "error": -293, "absolute_error": 293, "low_80": 1650, "high_80": 2150, "interval_covers": false, "loo_baseline": 1813.5, "rapid_games": 4525, "rating_deviation": 45, "explanation": "White developed coherently and sought practical counterplay with 25.Rxa7 and 28.e5–30.f5. However, the queen manoeuvres on moves 13–20 were inefficient, and 21.Qxe2 allowed liquidation that exposed the queenside pawns. Defensive play became less convincing after the queen exchange; 38.Rc8 and 39.R8xc5 underestimated Black's mating threats. This suggests an experienced intermediate player, with substantial uncertainty from one game." }, { "game": "game_09", "color": "black", "actual": 2177, "predicted": 2100, "error": -77, "absolute_error": 77, "low_80": 1850, "high_80": 2350, "interval_covers": true, "loo_baseline": 1813.5, "rapid_games": 8936, "rating_deviation": 46, "explanation": "Black showed coherent Benoni counterplay with 7...Na6–10...b5 and 13...Ng4–15...Nc4. The sequence 19...Nd4–23...Qxc3 converted positional pressure into material, while 25...Rb2 maintained activity. The bishop retreat 29...Bb1 and simplifying 32...Qf4+ handled White's initiative effectively. Finally, 38...Rdd2 and 41...f3 demonstrated strong attacking technique. Consistently purposeful play suggests a strong club-level Lichess rapid player." }, { "game": "game_10", "color": "white", "actual": 2689, "predicted": 1850, "error": -839, "absolute_error": 839, "low_80": 1500, "high_80": 2150, "interval_covers": false, "loo_baseline": 1767.9444444444443, "rapid_games": 9005, "rating_deviation": 45, "explanation": "White showed coherent opening understanding through 14.Qd2 and purposeful activity with 20.Nb5 and 23.Qc7. However, 24.Nxa7 underestimated Black’s counterplay: 24...Qc2 attacked rook and bishop, while 25...Ba6 subsequently won the exchange. The earlier coordination suggests an experienced player, but the pawn grab and inadequate threat assessment limit the estimate. There is no endgame evidence." }, { "game": "game_10", "color": "black", "actual": 2501, "predicted": 2000, "error": -501, "absolute_error": 501, "low_80": 1650, "high_80": 2300, "interval_covers": false, "loo_baseline": 1767.9444444444443, "rapid_games": 3141, "rating_deviation": 45, "explanation": "Black developed harmoniously, then adapted well with 16...fxe6 and 18...Rf8, obtaining central control and attacking prospects. The sequence 23...Bc5, 24...Qc2 and 25...Ba6 exploited White’s loose pieces efficiently. Particularly telling was 27...Ne8: it attacked the queen while clearing the f-file to support ...Qxf2+. This was convincing tactical and positional play, though one short game leaves substantial rating uncertainty." } ], "cost_usd": 0.80232, "reasoning_effort": "medium", "max_tokens": 8192, "run_start": "2026-10-11", "run_end": "2026-10-11", "display_name": "GPT-6 Astra" }, { "model": "openai/gpt-5.6-sol", "summary": { "games": 10, "player_estimates": 20, "actual_rating_range": [ 1112, 2718 ], "mae": 338.55, "rmse": 439.9613051167114, "signed_bias": -195.65, "pearson": 0.7296257889874272, "spearman": 0.7451967961310768, "within": { "100": { "count": 5, "fraction": 0.25 }, "200": { "count": 8, "fraction": 0.4 }, "300": { "count": 12, "fraction": 0.6 } }, "interval_80_coverage": 0.55, "interval_80_mean_width": 550.0, "game_mean_rating_mae": 325.25, "mae_game_bootstrap_95_interval": [ 205.05, 499.15124999999995 ], "baseline_1500_mae": 506.05, "baseline_leave_one_game_out_mean_mae": 471.5666666666666, "games_beating_loo_baseline": 7, "history_count_range": [ 1136, 31199 ], "max_profile_rating_deviation": 59 }, "predictions": [ { "game": "game_01", "color": "white", "actual": 1363, "predicted": 1450, "error": 87, "absolute_error": 87, "low_80": 1200, "high_80": 1700, "interval_covers": true, "loo_baseline": 1916.1666666666667, "rapid_games": 4696, "rating_deviation": 46, "explanation": "White recognized the strong 15.Nd5 and 16.Nf6+ fork, then consolidated with 21.Nxf3 and 23.Bxf3. However, conversion was weak: 31.e5 allowed ...Bxe1, repeated checks did not control the h-pawn, and 40.Qb8+ miscalculated the promotion sequence, turning a materially favorable position into a loss. The passive opening and two major tactical errors suggest an established rating around the mid-1400s." }, { "game": "game_01", "color": "black", "actual": 1159, "predicted": 1450, "error": 291, "absolute_error": 291, "low_80": 1200, "high_80": 1700, "interval_covers": false, "loo_baseline": 1916.1666666666667, "rapid_games": 3260, "rating_deviation": 45, "explanation": "Black showed some sound opening intent with 2...d5 and later generated practical counterplay through ...e4 and the passed h-pawn. Nevertheless, 15...Nxg5 missed the elementary 16.Nf6+ queen fork, an extremely serious oversight. Black recovered well with 30...Bxg3, 31...Bxe1, active rooks, and accurate handling of 39...Re8 through 42...Rxe8. Overall this resembles a tactically alert but error-prone mid-1400s player." }, { "game": "game_02", "color": "white", "actual": 1575, "predicted": 1400, "error": -175, "absolute_error": 175, "low_80": 1150, "high_80": 1650, "interval_covers": true, "loo_baseline": 1834.888888888889, "rapid_games": 1928, "rating_deviation": 45, "explanation": "White showed some opening familiarity with 10.d4 and the Qd3/Bh6 attacking setup, but 9.c3 allowed Black to win the e4-pawn cleanly. The queen trade after 18.h4 left a difficult two-pawn-down position, and 23.a3 was too slow. Most tellingly, 25.Bg5 overlooked the immediate back-rank mate ...Rh1#. Overall this resembles an established lower-intermediate rapid player." }, { "game": "game_02", "color": "black", "actual": 2410, "predicted": 1650, "error": -760, "absolute_error": 760, "low_80": 1400, "high_80": 1900, "interval_covers": false, "loo_baseline": 1834.888888888889, "rapid_games": 2936, "rating_deviation": 45, "explanation": "Black played a coherent game: 9...Nxe4 exploited White’s loose center, 13...Na5 and 14...Nc4 improved the knight with tempo, and 15...g6 correctly answered the Qd3/Bc2 battery. The simplifying 18...Bxh4 was sound while two pawns ahead, and 21...Re2, 23...Rb3, and 24...Rxh3 formed an effective mating attack. The play was accurate, though White offered relatively little resistance." }, { "game": "game_03", "color": "white", "actual": 1782, "predicted": 1800, "error": 18, "absolute_error": 18, "low_80": 1500, "high_80": 2050, "interval_covers": true, "loo_baseline": 1851.6666666666667, "rapid_games": 18588, "rating_deviation": 45, "explanation": "White showed purposeful attacking play with 15.Ng6 and accurately exploited 16...Kf7 through 18.Be7 and 19.Bxf6. The conversion was tactically alert: 33.Qf5, 34.Ne7+, 37.Rd8, and the final mating sequence were all clean. Some opening and queen maneuvers were not especially refined, and the opponent supplied major errors, so this strong game alone does not establish expert-level consistency." }, { "game": "game_03", "color": "black", "actual": 1901, "predicted": 1450, "error": -451, "absolute_error": 451, "low_80": 1150, "high_80": 1750, "interval_covers": false, "loo_baseline": 1851.6666666666667, "rapid_games": 1366, "rating_deviation": 50, "explanation": "Black played a coherent Dutch setup and later sought activity with 28...Rc4, 31...f4, and 36...Rg7. However, 16...Kf7 inexplicably declined the available rook recapture on d6, losing substantial material. Further tactical oversights with 33...Nc8 and especially 37...Ne7 allowed decisive forks and the queen’s loss. The repeated one-move misses suggest a modest established rapid rating despite some positional competence." }, { "game": "game_04", "color": "white", "actual": 2508, "predicted": 1750, "error": -758, "absolute_error": 758, "low_80": 1450, "high_80": 2050, "interval_covers": false, "loo_baseline": 1765.9444444444443, "rapid_games": 1324, "rating_deviation": 47, "explanation": "White showed coherent Caro-Kann development and a purposeful queenside expansion with 14.b4, 16.a4, 18.Nc5 and 20.a5. However, 21.c4 and especially 22.b5 neglected Black’s central and kingside threats; 23.Qb2 failed to address the developing mating net, and 24.Bb4 allowed the decisive 24...Ne2+. Overall, decent strategic fluency but a serious tactical oversight." }, { "game": "game_04", "color": "black", "actual": 2718, "predicted": 1950, "error": -768, "absolute_error": 768, "low_80": 1650, "high_80": 2250, "interval_covers": false, "loo_baseline": 1765.9444444444443, "rapid_games": 1136, "rating_deviation": 59, "explanation": "Black handled the opening confidently, castled oppositely, centralized with 16...Ne4 and organized counterplay through 17...f5 and 19...g5. The sequence 22...Nd4, 23...Qxc5 and 24...Ne2+ showed strong tactical coordination, culminating in the attractive 25...Qg1+ 26.Rxg1 Nf2#. Some earlier choices were risky, but the attack was purposeful and accurately executed." }, { "game": "game_05", "color": "white", "actual": 1120, "predicted": 1450, "error": 330, "absolute_error": 330, "low_80": 1200, "high_80": 1750, "interval_covers": false, "loo_baseline": 1890.3333333333333, "rapid_games": 5115, "rating_deviation": 45, "explanation": "White showed a coherent development plan with 10.e5 and 14.Bb2, but tactical awareness was weak. The retreat 16.Be4? missed 16...Nxg2+ followed by ...fxe4, while 19.Nfg5 wasted time. Most significantly, 22.Bc3 and 23.Nd4 overlooked the clearance of the b7–h1 diagonal and loss of the rook. These are typical intermediate-level errors, though one game leaves substantial uncertainty." }, { "game": "game_05", "color": "black", "actual": 1867, "predicted": 1850, "error": -17, "absolute_error": 17, "low_80": 1550, "high_80": 2100, "interval_covers": true, "loo_baseline": 1890.3333333333333, "rapid_games": 31199, "rating_deviation": 45, "explanation": "Black played an ambitious but purposeful opening and calculated several connected tactics accurately. The sequence 16...Nxg2+ 17...fxe4 won material, while 20...Nb4 and 21...Nbd3 established a strong outpost. The clearance tactic 22...Bxe4 23...Bxh1 and coordinated mating sequence from 24...Qb7 through 29...Qxf3# show strong tactical vision. Some opening choices, especially 11...f5, were committal rather than clearly high-level." }, { "game": "game_06", "color": "white", "actual": 1112, "predicted": 1400, "error": 288, "absolute_error": 288, "low_80": 1100, "high_80": 1700, "interval_covers": true, "loo_baseline": 1917.7222222222222, "rapid_games": 3821, "rating_deviation": 45, "explanation": "White showed a coherent attacking idea with 6.Bb1, 7.Qc2, 8.h4 and found 24.Nf6+, but neglected development and king safety. The sequence 16.f4 through 20.Bxa1 allowed a costly rook-for-knight loss without sufficient compensation. After 25.Rf1 and 27.Ke1, White failed to preserve the attack or recognize Black’s mating net, though 28.Qxc7+ showed continued tactical awareness." }, { "game": "game_06", "color": "black", "actual": 1382, "predicted": 1600, "error": 218, "absolute_error": 218, "low_80": 1300, "high_80": 1900, "interval_covers": true, "loo_baseline": 1917.7222222222222, "rapid_games": 5055, "rating_deviation": 45, "explanation": "Black’s opening was somewhat slow, especially 3...Bd7 and the bishop redeployment 9...Bf5 10...Bd7, but the position was handled tactically well. Black exploited White’s setup with 13...Nd4, 17...Nc5, 18...Ncxb3 and 19...Nxa1, then defended accurately with 22...Ne6 and 23...Ba6. The conversion was decisive: 27...Qxf6 followed by 29...Rac8, 30...Rc1+ and 31...Qf2#." }, { "game": "game_07", "color": "white", "actual": 1862, "predicted": 1500, "error": -362, "absolute_error": 362, "low_80": 1250, "high_80": 1750, "interval_covers": false, "loo_baseline": 1864.3333333333333, "rapid_games": 7581, "rating_deviation": 45, "explanation": "White showed coherent Slav structure knowledge with 8.cxd5, 13.Ba3 and the queenside expansion 17.b4–19.b5. However, 18.Rb3 placed the rook on a bishop-fork line, and White failed to notice ...Bc2 for several moves; 21.Bf3 allowed it outright. The countershot 22.Bh5 was alert but could not prevent losing a rook after 22...Bxd1 23.Bxg6 Bxb3." }, { "game": "game_07", "color": "black", "actual": 1593, "predicted": 1650, "error": 57, "absolute_error": 57, "low_80": 1400, "high_80": 1900, "interval_covers": true, "loo_baseline": 1864.3333333333333, "rapid_games": 13102, "rating_deviation": 45, "explanation": "Black played a sound Slav setup, used 11...Ne4 actively, and handled the simplifying sequence through 16...Qxd7 cleanly. The queen maneuver 18...Qe6 and 20...Qg6 was purposeful, and 21...Bc2 plus 22...Bxd1 accurately calculated the fork and intermediate queen trade. Black nevertheless overlooked the already available ...Bc2 fork for multiple moves after 18.Rb3, limiting the estimate." }, { "game": "game_08", "color": "white", "actual": 1310, "predicted": 1450, "error": 140, "absolute_error": 140, "low_80": 1150, "high_80": 1750, "interval_covers": true, "loo_baseline": 1884.0, "rapid_games": 4221, "rating_deviation": 45, "explanation": "The opening setup through 14.Rfe1 was coherent but modest. The major error 15.Bxc5? simply surrendered a bishop for a pawn; 16.Ne5 and 17.Bxe4 did not recover the material. Later 22.Rd4 allowed ...Bxf3+ followed by a knight fork, and 24.h4 directly enabled ...Nf3+. Correct resignation awareness, but repeated short tactical oversights suggest a below-average established rapid rating." }, { "game": "game_08", "color": "black", "actual": 1791, "predicted": 1750, "error": -41, "absolute_error": 41, "low_80": 1450, "high_80": 2050, "interval_covers": true, "loo_baseline": 1884.0, "rapid_games": 4152, "rating_deviation": 49, "explanation": "Black played a purposeful central plan with 9...e4, 11...Re8, and 14...c5, immediately exploiting White’s awkward queen placement. After 15.Bxc5, Black accurately maintained the extra piece with 16...Qb6 and 17...Nxe4. The conversion was efficient: 19...f6, 20...Bg4, 21...Ng5, and 22...Bxf3+ created the decisive fork sequence ending in 24...Nf3+. The game shows solid tactical awareness, though the test was limited." }, { "game": "game_09", "color": "white", "actual": 2193, "predicted": 1900, "error": -293, "absolute_error": 293, "low_80": 1600, "high_80": 2200, "interval_covers": true, "loo_baseline": 1813.5, "rapid_games": 4525, "rating_deviation": 45, "explanation": "White showed solid opening familiarity and purposeful play with 12.Bd2, 19.f4, and the active 28.e5–32.Rxe7 sequence. The rook coordination after 37.Re8+ and 38.Rc8 was also competent. However, the pawn sacrifices lacked full justification, and 39.R8xc5 overlooked the direct ...Rh2+ and ...Rbg2+ attack, a serious tactical miss." }, { "game": "game_09", "color": "black", "actual": 2177, "predicted": 2050, "error": -127, "absolute_error": 127, "low_80": 1750, "high_80": 2300, "interval_covers": true, "loo_baseline": 1813.5, "rapid_games": 8936, "rating_deviation": 46, "explanation": "Black handled the Benoni-style structure confidently with 7...Na6, 10...b5, and the well-timed 19...Nd4. The sequence 20...Nxe2–23...Qxc3 efficiently exploited White's structure, while 25...Rb2 and 32...Qf4+ maintained activity. The liquidation with 35...Bxd5 and the decisive 39...Rh2+–41...f3 mating net showed strong tactical awareness and conversion." }, { "game": "game_10", "color": "white", "actual": 2689, "predicted": 1750, "error": -939, "absolute_error": 939, "low_80": 1500, "high_80": 2000, "interval_covers": false, "loo_baseline": 1767.9444444444443, "rapid_games": 9005, "rating_deviation": 45, "explanation": "White showed sound opening familiarity through 14.Qd2 and found active ideas with 20.Nb5 and 22.Rxc8. However, repeated queen moves (15.Qf4, 18.Qh4, 21.Qf4) lost time, 24.Nxa7 was a risky pawn grab, and 26.Be5?? completely missed the direct 25...Ba6 attack on the rook. Overall this resembles a competent club player with a significant tactical lapse." }, { "game": "game_10", "color": "black", "actual": 2501, "predicted": 1850, "error": -651, "absolute_error": 651, "low_80": 1600, "high_80": 2100, "interval_covers": false, "loo_baseline": 1767.9444444444443, "rapid_games": 3141, "rating_deviation": 45, "explanation": "Black handled the opening coherently, including 12...Nxc5, 14...Rfe8, and the structurally sensible 16...fxe6. The queen activation 21...Qg6 and 24...Qc2 created concrete pressure, while 25...Ba6 accurately exploited the rook’s vulnerable diagonal; 27...Ne8 also calmly consolidated after winning the exchange. The play was consistently purposeful, though the short game provides limited evidence of deeper endgame strength." } ], "cost_usd": 0.185174, "reasoning_effort": "medium", "max_tokens": 8192, "run_start": "2026-10-11", "run_end": "2026-10-11", "display_name": "GPT-5.6 Sol" }, { "model": "google/gemini-3.5-flash", "summary": { "games": 10, "player_estimates": 20, "actual_rating_range": [ 1112, 2718 ], "mae": 346.75, "rmse": 451.8195989551582, "signed_bias": -310.65, "pearson": 0.832907404849772, "spearman": 0.840760703194705, "within": { "100": { "count": 5, "fraction": 0.25 }, "200": { "count": 8, "fraction": 0.4 }, "300": { "count": 11, "fraction": 0.55 } }, "interval_80_coverage": 0.4, "interval_80_mean_width": 360.0, "game_mean_rating_mae": 321.25, "mae_game_bootstrap_95_interval": [ 201.8475, 506.30249999999995 ], "baseline_1500_mae": 506.05, "baseline_leave_one_game_out_mean_mae": 471.5666666666666, "games_beating_loo_baseline": 6, "history_count_range": [ 1136, 31199 ], "max_profile_rating_deviation": 59 }, "predictions": [ { "game": "game_01", "color": "white", "actual": 1363, "predicted": 1250, "error": -113, "absolute_error": 113, "low_80": 1050, "high_80": 1450, "interval_covers": true, "loo_baseline": 1916.1666666666667, "rapid_games": 4696, "rating_deviation": 46, "explanation": "White chose a passive opening (2. d3) but capitalized on Black's tactical blunder with 16. Nf6+. White navigated the middlegame reasonably well with a large advantage, but ultimately threw the game with the catastrophic blunder 40. Qb8+??, hanging the queen entirely. This combination of basic tactical awareness and sudden massive blunders is typical of a 1250-rated player." }, { "game": "game_01", "color": "black", "actual": 1159, "predicted": 1150, "error": -9, "absolute_error": 9, "low_80": 950, "high_80": 1350, "interval_covers": true, "loo_baseline": 1916.1666666666667, "rapid_games": 3260, "rating_deviation": 45, "explanation": "Black's opening was highly suspect, blocking their own light-squared bishop with 7...Qd7. Black then suffered a severe tactical oversight, blundering the queen with 15...Nxg5? to the simple fork 16. Nf6+. Although Black managed to win after White's late-game collapse, the overall positional and tactical errors indicate a rating of around 1150." }, { "game": "game_02", "color": "white", "actual": 1575, "predicted": 1350, "error": -225, "absolute_error": 225, "low_80": 1150, "high_80": 1550, "interval_covers": false, "loo_baseline": 1834.888888888889, "rapid_games": 1928, "rating_deviation": 45, "explanation": "White played a standard opening but quickly lost pawn control with 8. Nd5, allowing Nxe4. The attack starting with 15. Qd3 was superficial and easily countered. 18. h4 was a major tactical oversight that lost a crucial pawn and traded active pieces. White's passive defense after 21... Re2 failed to address king safety, leading to a quick checkmate on h1." }, { "game": "game_02", "color": "black", "actual": 2410, "predicted": 1500, "error": -910, "absolute_error": 910, "low_80": 1300, "high_80": 1700, "interval_covers": false, "loo_baseline": 1834.888888888889, "rapid_games": 2936, "rating_deviation": 45, "explanation": "Black showed solid tactical awareness and clean execution. They snatched the free e4-pawn cleanly, developed harmoniously, and executed active knight maneuvers with 13... Na5 and 14... Nc4. Black clinical punished White's blunder with 18... Bxh4 and maintained heavy pressure on the second rank. The conversion was direct, accurate, and capped with an efficient mating net." }, { "game": "game_03", "color": "white", "actual": 1782, "predicted": 1850, "error": 68, "absolute_error": 68, "low_80": 1700, "high_80": 2000, "interval_covers": true, "loo_baseline": 1851.6666666666667, "rapid_games": 18588, "rating_deviation": 45, "explanation": "White showed excellent tactical awareness and conversion. After building a reasonable position in a Dutch defence, White immediately pounced on Black's mistake with 16. Bxd6!, correctly calculating that the d6-bishop could not be recaptured due to the Ne7+ fork. Following this, White converted the extra piece flawlessly with highly efficient moves (e.g., 24. Qb3, 33. Qf5), demonstrating a level of positional and tactical control typical of a strong high-intermediate player." }, { "game": "game_03", "color": "black", "actual": 1901, "predicted": 1650, "error": -251, "absolute_error": 251, "low_80": 1500, "high_80": 1800, "interval_covers": false, "loo_baseline": 1851.6666666666667, "rapid_games": 1366, "rating_deviation": 50, "explanation": "Black played a reasonable Dutch defence until making a critical tactical oversight with 15... Rd8?, failing to foresee White's reply 16. Bxd6! which forced the loss of a minor piece because recapturing allows a devastating knight fork on e7. After this blunder, Black struggled to mount any counterplay, showing passive resistance and allowing White a straightforward mating attack. This level of tactical vulnerability is characteristic of a mid-tier intermediate player." }, { "game": "game_04", "color": "white", "actual": 2508, "predicted": 1800, "error": -708, "absolute_error": 708, "low_80": 1600, "high_80": 2000, "interval_covers": false, "loo_baseline": 1765.9444444444443, "rapid_games": 1324, "rating_deviation": 47, "explanation": "White demonstrates solid opening knowledge in the Caro-Kann Exchange, playing thematic queenside expansion moves like 14. b4 and 18. Nc5. However, White struggled to coordinate a defense against Black's kingside initiative. The decisive blunder 24. Bb4? completely overlooked the forced smothered mate sequence starting with 24... Ne2+, indicating tactical vulnerability under pressure." }, { "game": "game_04", "color": "black", "actual": 2718, "predicted": 1900, "error": -818, "absolute_error": 818, "low_80": 1700, "high_80": 2100, "interval_covers": false, "loo_baseline": 1765.9444444444443, "rapid_games": 1136, "rating_deviation": 59, "explanation": "Black played an excellent game, showing strong theoretical understanding in the opening (5... Qc7, 6... Bg4) and managing the transition to a sharp opposite-side castling middlegame well. Black initiated a well-timed kingside break with 19... g5 and capitalized beautifully on White's blunder, executing a classic smothered mate sequence starting with 24... Ne2+ and 25... Qg1+." }, { "game": "game_05", "color": "white", "actual": 1120, "predicted": 1150, "error": 30, "absolute_error": 30, "low_80": 1000, "high_80": 1300, "interval_covers": true, "loo_baseline": 1890.3333333333333, "rapid_games": 5115, "rating_deviation": 45, "explanation": "White's opening play is highly passive and repetitive, moving the light-squared bishop five times by move 12 (Bc4-Bd3-Be2-Bd3-Be4-Bd3). White fails to defend the kingside, blundering with 16. Be4? which allows 16... Nxg2+, and completely collapses to basic tactics, eventually walking into a simple mate." }, { "game": "game_05", "color": "black", "actual": 1867, "predicted": 1450, "error": -417, "absolute_error": 417, "low_80": 1300, "high_80": 1600, "interval_covers": false, "loo_baseline": 1890.3333333333333, "rapid_games": 31199, "rating_deviation": 45, "explanation": "Black demonstrates solid opening principles, expanding on the queenside and developing pieces actively. Black misses a tactical opportunity to win a pawn with 13... Nxc3, but quickly capitalizes on White's blunders starting with 15... Nf4 and 16... Nxg2+. The tactical clean-up (20... Nb4, 21... Nbd3, and the mating net with 25... Nh3+) shows comfortable tactical awareness." }, { "game": "game_06", "color": "white", "actual": 1112, "predicted": 1150, "error": 38, "absolute_error": 38, "low_80": 950, "high_80": 1350, "interval_covers": true, "loo_baseline": 1917.7222222222222, "rapid_games": 3821, "rating_deviation": 45, "explanation": "White's opening plan of 4.Bd3 and 6.Bb1 is highly passive and strategically flawed, aiming only for a basic battery. White struggles with tactical awareness, allowing the fork 18...Ncxb3 and later blundering a rook with 25.Rf1, followed by losing the pinned knight on f6. These errors indicate a beginner-to-intermediate rating around 1150." }, { "game": "game_06", "color": "black", "actual": 1382, "predicted": 1450, "error": 68, "absolute_error": 68, "low_80": 1250, "high_80": 1650, "interval_covers": true, "loo_baseline": 1917.7222222222222, "rapid_games": 5055, "rating_deviation": 45, "explanation": "Black defends soundly against White's direct threats with 7...g6 and 8...h5, then efficiently capitalizes on White's structural weaknesses with 13...Nd4 and 17...Nc5. Black cleanly spots the tactical wins starting with 18...Ncxb3 and 25...Bxf1, converting the material advantage into a swift checkmate. Black's play suggests a solid intermediate rating of around 1450." }, { "game": "game_07", "color": "white", "actual": 1862, "predicted": 1600, "error": -262, "absolute_error": 262, "low_80": 1400, "high_80": 1800, "interval_covers": false, "loo_baseline": 1864.3333333333333, "rapid_games": 7581, "rating_deviation": 45, "explanation": "White played a solid and logical game, organizing a standard queenside expansion with a4, b4, and b5. However, White exhibited a significant tactical blindspot by placing the rook on b3 and failing to anticipate Black's double threat on move 20, leading to a direct blunder of a full rook after 21. Bf3 Bc2." }, { "game": "game_07", "color": "black", "actual": 1593, "predicted": 1750, "error": 157, "absolute_error": 157, "low_80": 1550, "high_80": 1950, "interval_covers": true, "loo_baseline": 1864.3333333333333, "rapid_games": 13102, "rating_deviation": 45, "explanation": "Black demonstrated strong positional understanding, easily neutralizing White's queenside pressure. Black showed excellent tactical alertness with 20...Qg6, combining a mate threat on g2 with a fork of White's queen and undefended b3-rook via 21...Bc2. Black executed the winning sequence flawlessly, reflecting solid intermediate capability." }, { "game": "game_08", "color": "white", "actual": 1310, "predicted": 1200, "error": -110, "absolute_error": 110, "low_80": 1000, "high_80": 1400, "interval_covers": true, "loo_baseline": 1884.0, "rapid_games": 4221, "rating_deviation": 45, "explanation": "White's passive opening setup (1. Nc3, 2. d3, 3. e3) and multiple tactical oversights suggest a lower rating. They missed the pin on the f3-knight (8. O-O), allowed a simple piece loss with 14. Rfe1 c5, and fell into another pin/fork combination with 20. Nf3 Bg4 and 24. h4 Nf3+, showing weak tactical awareness." }, { "game": "game_08", "color": "black", "actual": 1791, "predicted": 1450, "error": -341, "absolute_error": 341, "low_80": 1250, "high_80": 1650, "interval_covers": false, "loo_baseline": 1884.0, "rapid_games": 4152, "rating_deviation": 49, "explanation": "Black played a solid game, developing naturally and punishing White's mistakes. They capitalized on the f3 pin with 9... e4, spotted the tactical win of a piece with 14... c5, and efficiently converted the endgame with precise tactics, including 20... Bg4 and the final fork 24... Nf3+. This level of tactical alertness is typical of an intermediate player." }, { "game": "game_09", "color": "white", "actual": 2193, "predicted": 1550, "error": -643, "absolute_error": 643, "low_80": 1400, "high_80": 1700, "interval_covers": false, "loo_baseline": 1813.5, "rapid_games": 4525, "rating_deviation": 45, "explanation": "White played a reasonable opening but quickly lost control. A tactical lapse with 15. Kh2 allowed Black's knight to active squares and capture the bishop pair. White then blundered a pawn with 21. Qxe2?, weakening the queenside. Under pressure, White failed to find active defensive resources and ultimately allowed a simple mating net in the rook endgame." }, { "game": "game_09", "color": "black", "actual": 2177, "predicted": 1750, "error": -427, "absolute_error": 427, "low_80": 1600, "high_80": 1900, "interval_covers": false, "loo_baseline": 1813.5, "rapid_games": 8936, "rating_deviation": 46, "explanation": "Black demonstrated strong positional understanding of the Benoni, executing the ...b5 queenside expansion effectively. Black punished White's passive play with energetic moves like 13...Ng4 and 15...Nc4. The conversion of the material advantage was highly efficient, particularly in the rook endgame where Black simplified with 35...Bxd5 and coordinated the rooks to force a quick mate." }, { "game": "game_10", "color": "white", "actual": 2689, "predicted": 1850, "error": -839, "absolute_error": 839, "low_80": 1700, "high_80": 2000, "interval_covers": false, "loo_baseline": 1767.9444444444443, "rapid_games": 9005, "rating_deviation": 45, "explanation": "White demonstrated comfortable opening knowledge in the English/Reti and handled the early middlegame logically (13. Nd4, 15. Qf4). However, 24. Nxa7 was a severe tactical oversight that completely missed the double attack 24... Qc2. This mix of solid positional play and a sudden, game-ending tactical blind spot is highly characteristic of an established player around the high-1800s." }, { "game": "game_10", "color": "black", "actual": 2501, "predicted": 2000, "error": -501, "absolute_error": 501, "low_80": 1850, "high_80": 2150, "interval_covers": false, "loo_baseline": 1767.9444444444443, "rapid_games": 3141, "rating_deviation": 45, "explanation": "Black played an exceptionally clean game. The opening and middlegame development were harmonic, managing the hanging pawns structure after 16... fxe6. Crucially, Black demonstrated sharp tactical vision by instantly exploiting White's error with 24... Qc2!, followed by the decisive 25... Ba6 and the precise finishing move 27... Ne8!. This level of tactical alertness and structural understanding points to a solid 2000-level player." } ], "cost_usd": 0.195573, "reasoning_effort": "medium", "max_tokens": 8192, "run_start": "2026-10-11", "run_end": "2026-10-11", "display_name": "Gemini 3.5 Flash" }, { "model": "openai/gpt-6-sol", "summary": { "games": 10, "player_estimates": 20, "actual_rating_range": [ 1112, 2718 ], "mae": 352.05, "rmse": 471.14323724319763, "signed_bias": -243.15, "pearson": 0.7144596214336087, "spearman": 0.7115597600182529, "within": { "100": { "count": 3, "fraction": 0.15 }, "200": { "count": 8, "fraction": 0.4 }, "300": { "count": 13, "fraction": 0.65 } }, "interval_80_coverage": 0.65, "interval_80_mean_width": 485.0, "game_mean_rating_mae": 312.85, "mae_game_bootstrap_95_interval": [ 192.24875, 545.2549999999999 ], "baseline_1500_mae": 506.05, "baseline_leave_one_game_out_mean_mae": 471.5666666666666, "games_beating_loo_baseline": 5, "history_count_range": [ 1136, 31199 ], "max_profile_rating_deviation": 59 }, "predictions": [ { "game": "game_01", "color": "white", "actual": 1363, "predicted": 1400, "error": 37, "absolute_error": 37, "low_80": 1150, "high_80": 1650, "interval_covers": true, "loo_baseline": 1916.1666666666667, "rapid_games": 4696, "rating_deviation": 46, "explanation": "White spotted the knight fork with 15.Nd5 and 16.Nf6+, then advanced the passed pawn with 38.e6 and 39.e7. Earlier bishop retreats and queen shuffling were less convincing, and 40.Qb8+ overlooked that the rook could take the queen. The mix suggests an intermediate player, though one game leaves substantial uncertainty." }, { "game": "game_01", "color": "black", "actual": 1159, "predicted": 1400, "error": 241, "absolute_error": 241, "low_80": 1150, "high_80": 1650, "interval_covers": true, "loo_baseline": 1916.1666666666667, "rapid_games": 3260, "rating_deviation": 45, "explanation": "Black played energetically with 4...f5 and 9...g5, but 15...Nxg5 allowed a straightforward knight fork that lost the queen. Black kept creating problems with 27...e4 and 30...Bxg3, then correctly exploited 40.Qb8+. Resourceful defence offsets the major tactical oversight, suggesting an intermediate rating." }, { "game": "game_02", "color": "white", "actual": 1575, "predicted": 1450, "error": -125, "absolute_error": 125, "low_80": 1200, "high_80": 1700, "interval_covers": true, "loo_baseline": 1834.888888888889, "rapid_games": 1928, "rating_deviation": 45, "explanation": "White developed sensibly and found the active 15.Qd3, but 18.h4 allowed Black to open lines toward the king. After 20.Qh3 Qxh3, White did not adequately address the rook invasion: 23.a3 and 24.Bc2 allowed ...Rb3 and ...Rxh3, and 25.Bg5 missed the mate. These lapses suggest an intermediate player, though one short game leaves considerable uncertainty." }, { "game": "game_02", "color": "black", "actual": 2410, "predicted": 1650, "error": -760, "absolute_error": 760, "low_80": 1400, "high_80": 1900, "interval_covers": false, "loo_baseline": 1834.888888888889, "rapid_games": 2936, "rating_deviation": 45, "explanation": "Black handled the opening coherently with 9...Nxe4 and 13...Na5–14...Nc4, then developed pressure with 16...Re8 and 17...Bb7. After 18.h4, Black used the open lines effectively: 21...Re2, 22...Rxb2 and 23...Rb3 led to a precise mating finish. The game shows good tactical awareness, but offers little evidence about Black's defensive or endgame strength." }, { "game": "game_03", "color": "white", "actual": 1782, "predicted": 1800, "error": 18, "absolute_error": 18, "low_80": 1550, "high_80": 2000, "interval_covers": true, "loo_baseline": 1851.6666666666667, "rapid_games": 18588, "rating_deviation": 45, "explanation": "White spotted the tactical opportunity with 16.Bxd6 and followed it with 17.Ne5+ and 18.Be7. Later, 34.Ne7+ used a discovered check to win the rook, and 37.Rd8 exploited the queen's position. The opening knight maneuvers, especially 9.Ng5 and 11.Nh3, cost time, but the sustained tactical awareness suggests an established rating around 1800." }, { "game": "game_03", "color": "black", "actual": 1901, "predicted": 1600, "error": -301, "absolute_error": 301, "low_80": 1350, "high_80": 1800, "interval_covers": false, "loo_baseline": 1851.6666666666667, "rapid_games": 1366, "rating_deviation": 50, "explanation": "Black developed coherently through 14...Bd6, but 15...Rd8 allowed a tactical sequence beginning 16.Bxd6. The defence with 16...Kf7 and 18...Re8 left White with a substantial advantage. Black remained active with 28...Rc4 and 31...f4, but missed the discovered-check tactic after 33...Nc8. Those decisions suggest roughly 1600, with considerable uncertainty from one game." }, { "game": "game_04", "color": "white", "actual": 2508, "predicted": 1650, "error": -858, "absolute_error": 858, "low_80": 1400, "high_80": 1900, "interval_covers": false, "loo_baseline": 1765.9444444444443, "rapid_games": 1324, "rating_deviation": 47, "explanation": "White showed a coherent opposite-side plan with 14.b4, 16.a4 and 18.Nc5. But 21.c4 allowed Black to loosen the kingside and take the c5-pawn, while 24.Bb4 overlooked a forcing mating sequence. One missed tactic in a 10-minute game leaves considerable uncertainty." }, { "game": "game_04", "color": "black", "actual": 2718, "predicted": 1800, "error": -918, "absolute_error": 918, "low_80": 1550, "high_80": 2050, "interval_covers": false, "loo_baseline": 1765.9444444444443, "rapid_games": 1136, "rating_deviation": 59, "explanation": "Black developed sensibly, met the queenside advance with 16...Ne4 and 20...a6, then used 21...gxf4 and 22...Nd4 to seize the initiative. The calculated 24...Ne2+ followed by 25...Qg1+ and 26...Nf2# is impressive, though one tactical finish alone cannot establish a precise rating." }, { "game": "game_05", "color": "white", "actual": 1120, "predicted": 1300, "error": 180, "absolute_error": 180, "low_80": 1050, "high_80": 1550, "interval_covers": true, "loo_baseline": 1890.3333333333333, "rapid_games": 5115, "rating_deviation": 45, "explanation": "White spent several opening moves retreating the bishop (5.Bd3, 6.Be2, 12.Bd3). More tellingly, 16.Be4 allowed ...fxe4 after ...Nxg2+, and 22.Bc3 left the e4-knight vulnerable. These repeated missed threats suggest a developing player, though one game leaves considerable uncertainty." }, { "game": "game_05", "color": "black", "actual": 1867, "predicted": 1700, "error": -167, "absolute_error": 167, "low_80": 1450, "high_80": 1950, "interval_covers": true, "loo_baseline": 1890.3333333333333, "rapid_games": 31199, "rating_deviation": 45, "explanation": "Black developed coherently with 14...Qc7 and used 15...Nf4 to create immediate threats. The sequence 16...Nxg2+ 17...fxe4 won material, while 20...Nb4 and 21...Nbd3 kept the initiative. Black spotted 22...Bxe4 and converted with forcing moves through 29...Qxf3#, although White offered substantial tactical opportunities." }, { "game": "game_06", "color": "white", "actual": 1112, "predicted": 1300, "error": 188, "absolute_error": 188, "low_80": 1050, "high_80": 1550, "interval_covers": true, "loo_baseline": 1917.7222222222222, "rapid_games": 3821, "rating_deviation": 45, "explanation": "White developed a plausible English setup, but 6.Bb1 and 15.f3 followed by 16.f4 cost time and loosened the position. After 18...Ncxb3, 19.f5 allowed the rook to fall. The active 24.Nf6+ showed tactical intent, but 25.Rf1 overlooked the bishop on a6, and 27.Ke1 left the knight vulnerable. One game leaves substantial uncertainty." }, { "game": "game_06", "color": "black", "actual": 1382, "predicted": 1600, "error": 218, "absolute_error": 218, "low_80": 1350, "high_80": 1800, "interval_covers": true, "loo_baseline": 1917.7222222222222, "rapid_games": 5055, "rating_deviation": 45, "explanation": "Black responded coherently to White's kingside play with 8...h5 and created useful pressure with 13...Nd4. The sequence 16...Bxg5, 17...Nc5 and 18...Ncxb3 exploited White's loose queenside; 23...Ba6 maintained pressure on f1. Black then removed the pawn guarding the knight with 26...Nxg5 before 27...Qxf6. The opening included some bishop retreats, so this is not evidence of consistently precise play." }, { "game": "game_07", "color": "white", "actual": 1862, "predicted": 1650, "error": -212, "absolute_error": 212, "low_80": 1400, "high_80": 1900, "interval_covers": true, "loo_baseline": 1864.3333333333333, "rapid_games": 7581, "rating_deviation": 45, "explanation": "White developed coherently and pursued a sensible queenside plan with 17.b4 and 19.b5. But 21.Bf3 allowed ...Bc2 to attack both the queen and rook; 22.Bh5 won Black's queen only at the cost of White's queen and rook. That tactical oversight weighs against a higher estimate, though one game leaves substantial uncertainty." }, { "game": "game_07", "color": "black", "actual": 1593, "predicted": 1800, "error": 207, "absolute_error": 207, "low_80": 1550, "high_80": 2050, "interval_covers": true, "loo_baseline": 1864.3333333333333, "rapid_games": 13102, "rating_deviation": 45, "explanation": "Black handled the Slav structure soundly, met the queenside expansion with 19...axb5, and found the decisive 21...Bc2, recognizing that 22.Bh5 could be answered by ...Bxd1 and ...Bxb3. The earlier play was competent rather than especially revealing, so the estimate remains broad." }, { "game": "game_08", "color": "white", "actual": 1310, "predicted": 1300, "error": -10, "absolute_error": 10, "low_80": 1100, "high_80": 1550, "interval_covers": true, "loo_baseline": 1884.0, "rapid_games": 4221, "rating_deviation": 45, "explanation": "White developed and castled sensibly, but 15.Bxc5 gave up a bishop for a pawn, and 18.Qxb6 simplified while behind in material. Later, 22.Rd4 and 24.h4 allowed the decisive ...Nf3+ fork. Those oversights suggest a lower-intermediate established rating, though one game leaves substantial uncertainty." }, { "game": "game_08", "color": "black", "actual": 1791, "predicted": 1600, "error": -191, "absolute_error": 191, "low_80": 1350, "high_80": 1800, "interval_covers": true, "loo_baseline": 1884.0, "rapid_games": 4152, "rating_deviation": 49, "explanation": "Black played a coherent central opening with 9...e4 and 14...c5, correctly recaptured after 15.Bxc5, and converted the advantage with 20...Bg4, 22...Bxf3+ and 24...Nf3+. The play was purposeful, but the short game against significant errors gives limited evidence of a higher rating." }, { "game": "game_09", "color": "white", "actual": 2193, "predicted": 1800, "error": -393, "absolute_error": 393, "low_80": 1550, "high_80": 2050, "interval_covers": false, "loo_baseline": 1813.5, "rapid_games": 4525, "rating_deviation": 45, "explanation": "White played a coherent Benoni setup and found active counterplay with 25.Rxa7 and 28.e5. But 19.f4 allowed ...Nd4 and the ensuing ...Bxc3 and queen raids; later 35.Rc1 let Black take the d5 pawn. The play shows tactical awareness, with recurring positional and defensive lapses." }, { "game": "game_09", "color": "black", "actual": 2177, "predicted": 1950, "error": -227, "absolute_error": 227, "low_80": 1750, "high_80": 2200, "interval_covers": true, "loo_baseline": 1813.5, "rapid_games": 8936, "rating_deviation": 46, "explanation": "Black handled the Benoni structure purposefully with 10...b5, then coordinated ...Nd4, ...Bxc3 and ...Qxa3 to exploit White's queenside. The bishop retreat 29...Bb1 and active rooks from 38...Rdd2 showed good calculation and conversion. One well-played game leaves substantial uncertainty about the established rating." }, { "game": "game_10", "color": "white", "actual": 2689, "predicted": 1650, "error": -1039, "absolute_error": 1039, "low_80": 1400, "high_80": 1900, "interval_covers": false, "loo_baseline": 1767.9444444444443, "rapid_games": 9005, "rating_deviation": 45, "explanation": "White handled the opening coherently through 14.Qd2 and found active play with 20.Nb5. But 24.Nxa7 prioritized a pawn while Black’s queen became active, and 25.Rf1 allowed ...Ba6 to target the rook. The game is too short to judge White’s endgame ability." }, { "game": "game_10", "color": "black", "actual": 2501, "predicted": 1750, "error": -751, "absolute_error": 751, "low_80": 1500, "high_80": 2000, "interval_covers": false, "loo_baseline": 1767.9444444444443, "rapid_games": 3141, "rating_deviation": 45, "explanation": "Black developed sensibly with 7...c5 and 11...Rc8, then used 18...Rf8 and 21...Qg6 to build pressure. The sequence 24...Qc2, 25...Ba6 and 26...Bxf1 shows good tactical awareness. The short game provides little evidence about endgame strength." } ], "cost_usd": 0.200374, "reasoning_effort": "medium", "max_tokens": 8192, "run_start": "2026-10-11", "run_end": "2026-10-11", "display_name": "GPT-6 Sol" }, { "model": "google/gemini-3.6-flash", "summary": { "games": 10, "player_estimates": 20, "actual_rating_range": [ 1112, 2718 ], "mae": 357.05, "rmse": 443.1319780832794, "signed_bias": -270.65, "pearson": 0.8177690873394186, "spearman": 0.7873158797062895, "within": { "100": { "count": 3, "fraction": 0.15 }, "200": { "count": 7, "fraction": 0.35 }, "300": { "count": 12, "fraction": 0.6 } }, "interval_80_coverage": 0.2, "interval_80_mean_width": 300.0, "game_mean_rating_mae": 350.35, "mae_game_bootstrap_95_interval": [ 230.7, 505.65 ], "baseline_1500_mae": 506.05, "baseline_leave_one_game_out_mean_mae": 471.5666666666666, "games_beating_loo_baseline": 7, "history_count_range": [ 1136, 31199 ], "max_profile_rating_deviation": 59 }, "predictions": [ { "game": "game_01", "color": "white", "actual": 1363, "predicted": 1400, "error": 37, "absolute_error": 37, "low_80": 1250, "high_80": 1550, "interval_covers": true, "loo_baseline": 1916.1666666666667, "rapid_games": 4696, "rating_deviation": 46, "explanation": "White played a passive opening (2. d3) but demonstrated good tactical vision with 16. Nf6+ to win Black's queen. White converted the advantage well through the endgame with strong pawn pushes (38. e6). However, White suffered a catastrophic hallucination or mouseslip with 40. Qb8+??, giving away the queen and game instantly. This mix of intermediate calculation and severe inconsistency suggests a mid-1400 Lichess rapid level." }, { "game": "game_01", "color": "black", "actual": 1159, "predicted": 1350, "error": 191, "absolute_error": 191, "low_80": 1200, "high_80": 1500, "interval_covers": false, "loo_baseline": 1916.1666666666667, "rapid_games": 3260, "rating_deviation": 45, "explanation": "Black played an overly aggressive, structurally weak opening (4... f5, 9... g5) and completely missed the tactical fork 16. Nf6+, losing the queen for a piece. Black struggled positionally while down material, though managed to push the h-pawn to create minor threats. Black won purely due to White's game-ending blunder 40. Qb8+??. The lack of tactical prophylaxis and dubious opening choices indicate a lower-intermediate rating." }, { "game": "game_02", "color": "white", "actual": 1575, "predicted": 1350, "error": -225, "absolute_error": 225, "low_80": 1200, "high_80": 1500, "interval_covers": false, "loo_baseline": 1834.888888888889, "rapid_games": 1928, "rating_deviation": 45, "explanation": "White selected a slightly passive line in the Ruy Lopez (5.Nc3) and gave away the e4-pawn with 9.c3. After decent midgame development, White made a tactical blunder with 18.h4 Bxh4, allowing Black to win material. White then failed to defend against Black's rook invasion, culminating in a complete oversight of the mating threat on h1 with 24.Bc2." }, { "game": "game_02", "color": "black", "actual": 2410, "predicted": 1650, "error": -760, "absolute_error": 760, "low_80": 1500, "high_80": 1800, "interval_covers": false, "loo_baseline": 1834.888888888889, "rapid_games": 2936, "rating_deviation": 45, "explanation": "Black demonstrated solid opening knowledge, correctly punishing White's pawn offering with 9...Nxe4. Black positioned pieces effectively (14...Nc4), promptly punished 18.h4 with 18...Bxh4, and converted the advantage smoothly. Black executed an active rook endgame (21...Re2, 23...Rb3) and spotted the forced checkmate immediately." }, { "game": "game_03", "color": "white", "actual": 1782, "predicted": 1550, "error": -232, "absolute_error": 232, "low_80": 1400, "high_80": 1700, "interval_covers": false, "loo_baseline": 1851.6666666666667, "rapid_games": 18588, "rating_deviation": 45, "explanation": "White plays a solid setup and demonstrates strong tactical alertness with 16.Bxd6! exploiting Black's tactical oversight. White converts the advantage smoothly with energetic moves like 24.Qb3! and 34.Ne7+, showing consistent conversion technique." }, { "game": "game_03", "color": "black", "actual": 1901, "predicted": 1300, "error": -601, "absolute_error": 601, "low_80": 1150, "high_80": 1450, "interval_covers": false, "loo_baseline": 1851.6666666666667, "rapid_games": 1366, "rating_deviation": 50, "explanation": "Black develops logically in a Dutch structure but suffers from severe tactical blind spots. 15...Rd8? overlooks the sequence starting with 16.Bxd6!, and 22...e5? drops the d5-pawn cleanly, leading to a swift defeat." }, { "game": "game_04", "color": "white", "actual": 2508, "predicted": 1800, "error": -708, "absolute_error": 708, "low_80": 1650, "high_80": 1950, "interval_covers": false, "loo_baseline": 1765.9444444444443, "rapid_games": 1324, "rating_deviation": 47, "explanation": "White demonstrated solid opening knowledge in the Caro-Kann Exchange (5.c3, 8.Qe1) and launched a coherent queenside pawn storm (14.b4, 20.a5). However, White failed to manage Black's central counterplay and blundered fatally with 24.Bb4??, walking directly into a forced smothered mate." }, { "game": "game_04", "color": "black", "actual": 2718, "predicted": 1850, "error": -868, "absolute_error": 868, "low_80": 1700, "high_80": 2000, "interval_covers": false, "loo_baseline": 1765.9444444444443, "rapid_games": 1136, "rating_deviation": 59, "explanation": "Black played a very clean game, combining solid opening technique with strong positional sense (16...Ne4, 17...f5). Black correctly initiated active counterplay with 19...g5 and spotlessly executed the classic tactical sequence 24...Ne2+ 25...Qg1+!! to deliver smothered mate." }, { "game": "game_05", "color": "white", "actual": 1120, "predicted": 1400, "error": 280, "absolute_error": 280, "low_80": 1250, "high_80": 1550, "interval_covers": false, "loo_baseline": 1890.3333333333333, "rapid_games": 5115, "rating_deviation": 45, "explanation": "White mismanaged the opening by moving the light-squared bishop four times by move 12 (Bc4-Bd3-Be2-Bd3), granting Black easy space. 13. b3 created weaknesses around the c3-pawn, and 16. Be4?? completely blundered material while unravelling White's kingside. White showed poor opening planning and weak tactical vision under pressure." }, { "game": "game_05", "color": "black", "actual": 1867, "predicted": 1800, "error": -67, "absolute_error": 67, "low_80": 1650, "high_80": 1950, "interval_covers": true, "loo_baseline": 1890.3333333333333, "rapid_games": 31199, "rating_deviation": 45, "explanation": "Black played a very strong game, efficiently punishing White's wasted tempi with ...b5 and ...c4. Black demonstrated sharp tactical alertness with 15...Nf4! and 16...Nxg2+, and seamlessly conducted a textbook attack (20...Nb4!, 24...Qb7, 25...Nh3+) to force checkmate in 29 moves." }, { "game": "game_06", "color": "white", "actual": 1112, "predicted": 1300, "error": 188, "absolute_error": 188, "low_80": 1150, "high_80": 1450, "interval_covers": false, "loo_baseline": 1917.7222222222222, "rapid_games": 3821, "rating_deviation": 45, "explanation": "White demonstrated weak opening principles with passive, waste-of-tempo bishop maneuvers (4. Bd3 followed by 6. Bb1, and 3. e3 followed by 10. e4). Their kingside pawn pushes left structural weaknesses, and tactical awareness was poor, capped by hanging the f1 rook with 25. Rf1 and allowing a straightforward checkmate sequence." }, { "game": "game_06", "color": "black", "actual": 1382, "predicted": 1550, "error": 168, "absolute_error": 168, "low_80": 1400, "high_80": 1700, "interval_covers": false, "loo_baseline": 1917.7222222222222, "rapid_games": 5055, "rating_deviation": 45, "explanation": "After an unusual 3...Bd7, Black quickly capitalized on White's lack of development. Black showed good tactical sharpness with 13...Nd4, 17...Nc5, and 18...Ncxb3, winning heavy material cleanly. Black converted the advantage ruthlessly, accepting the free rook on f1 and delivering a decisive checkmate with 29...Rac8, 30...Rc1+, and 31...Qf2#." }, { "game": "game_07", "color": "white", "actual": 1862, "predicted": 1450, "error": -412, "absolute_error": 412, "low_80": 1300, "high_80": 1600, "interval_covers": false, "loo_baseline": 1864.3333333333333, "rapid_games": 7581, "rating_deviation": 45, "explanation": "White played a solid but passive opening (3. e3, 5. Nbd2) and maintained a balanced position through move 19. However, White overlooked Black's dual threats with 20...Qg6 (threatening mate on g2 and 21...Bc2 targeting the undefended b3-rook). After 21. Bf3 Bc2 22. Bh5 Bxd1, White lost a full rook and resigned." }, { "game": "game_07", "color": "black", "actual": 1593, "predicted": 1550, "error": -43, "absolute_error": 43, "low_80": 1400, "high_80": 1700, "interval_covers": true, "loo_baseline": 1864.3333333333333, "rapid_games": 13102, "rating_deviation": 45, "explanation": "Black played a very clean Slav Defence, achieving easy equality with natural piece placement. Black seized the initiative with 20...Qg6!, creating a dual threat of mate on g2 and winning material with ...Bc2. Black accurately calculated the tactical sequence 21...Bc2 22. Bh5 Bxd1 23. Bxg6 Bxb3 to emerge a whole rook ahead." }, { "game": "game_08", "color": "white", "actual": 1310, "predicted": 1200, "error": -110, "absolute_error": 110, "low_80": 1050, "high_80": 1350, "interval_covers": true, "loo_baseline": 1884.0, "rapid_games": 4221, "rating_deviation": 45, "explanation": "White played a passive, planless opening (1. Nc3, 2. d3, 3. e3). White struggled with pins after 9...e4, then blundered a piece to the simple fork 14...c5!. White continued to miss basic tactics, allowing 21...Ng5 and 22. Rd4?? which led to 22...Bxf3+ and the decisive fork 24...Nf3+. These frequent tactical errors reflect a lower-intermediate level." }, { "game": "game_08", "color": "black", "actual": 1791, "predicted": 1500, "error": -291, "absolute_error": 291, "low_80": 1350, "high_80": 1650, "interval_covers": false, "loo_baseline": 1884.0, "rapid_games": 4152, "rating_deviation": 49, "explanation": "Black seized the center effectively with 2...d5 and 9...e4!. Black demonstrated good tactical alertness, immediately punishing White's mistake with 14...c5! to win a piece. After trading queens, Black maintained control efficiently, using 21...Ng5! to set up the winning tactical sequence ending in 24...Nf3+. Black played cleanly without blunders, displaying solid intermediate skill." }, { "game": "game_09", "color": "white", "actual": 2193, "predicted": 1850, "error": -343, "absolute_error": 343, "low_80": 1700, "high_80": 2000, "interval_covers": false, "loo_baseline": 1813.5, "rapid_games": 4525, "rating_deviation": 45, "explanation": "White handled the opening passively with 6. g3 and 8. a3, allowing Black easy queenside expansion. 19. f4 weakened the position, and 21. Qxe2 allowed 21...Bxc3, dropping a pawn. White struggled under tactical pressure after 25...Rb2, leaving pieces passive with 27. Rg1 and collapsing into a forced checkmate network by 41...f3." }, { "game": "game_09", "color": "black", "actual": 2177, "predicted": 1950, "error": -227, "absolute_error": 227, "low_80": 1800, "high_80": 2100, "interval_covers": false, "loo_baseline": 1813.5, "rapid_games": 8936, "rating_deviation": 46, "explanation": "Black demonstrated strong opening understanding with 10...b5 and excellent tactical awareness. Maneuvers like 14...Ne5 and 15...Nc4 gained full control of the dark squares, followed by 21...Bxc3 to net a pawn. Black's attack with 26...Qd2 and conversion via 34...Ba2 and 41...f3 showed accurate calculations typical of an advanced player." }, { "game": "game_10", "color": "white", "actual": 2689, "predicted": 1850, "error": -839, "absolute_error": 839, "low_80": 1700, "high_80": 2000, "interval_covers": false, "loo_baseline": 1767.9444444444443, "rapid_games": 9005, "rating_deviation": 45, "explanation": "White showed strong positional understanding of double-fianchetto Reti/Catalan structures through 14. Qd2. However, after standard maneuvering, White suffered a fatal tactical oversight with 24. Nxa7?, leaving b2 undefended and allowing Black's crushing double attack 24...Qc2." }, { "game": "game_10", "color": "black", "actual": 2501, "predicted": 1950, "error": -551, "absolute_error": 551, "low_80": 1800, "high_80": 2100, "interval_covers": false, "loo_baseline": 1767.9444444444443, "rapid_games": 3141, "rating_deviation": 45, "explanation": "Black demonstrated high opening competence (1...e6 to 12...Nxc5) and sound positional play. When White miscalculated, Black punished the mistake with great tactical sharpness, finding 24...Qc2! and following up accurately with 25...Ba6, 26...Bxf1, and 27...Ne8! to force White's resignation." } ], "cost_usd": 0.061208250000000006, "reasoning_effort": "medium", "max_tokens": 8192, "run_start": "2026-10-11", "run_end": "2026-10-11", "display_name": "Gemini 3.6 Flash" }, { "model": "anthropic/claude-sonnet-5.5", "summary": { "games": 10, "player_estimates": 20, "actual_rating_range": [ 1112, 2718 ], "mae": 433.55, "rmse": 539.4496732782401, "signed_bias": -253.15, "pearson": 0.4510479466597261, "spearman": 0.4581770424786477, "within": { "100": { "count": 2, "fraction": 0.1 }, "200": { "count": 5, "fraction": 0.25 }, "300": { "count": 9, "fraction": 0.45 } }, "interval_80_coverage": 0.35, "interval_80_mean_width": 510.0, "game_mean_rating_mae": 392.85, "mae_game_bootstrap_95_interval": [ 275.6, 629.30125 ], "baseline_1500_mae": 506.05, "baseline_leave_one_game_out_mean_mae": 471.5666666666666, "games_beating_loo_baseline": 4, "history_count_range": [ 1136, 31199 ], "max_profile_rating_deviation": 59 }, "predictions": [ { "game": "game_01", "color": "white", "actual": 1363, "predicted": 1550, "error": 187, "absolute_error": 187, "low_80": 1300, "high_80": 1800, "interval_covers": true, "loo_baseline": 1916.1666666666667, "rapid_games": 4696, "rating_deviation": 46, "explanation": "Passive opening with 2.d3. Weakening h4 and Nh2 ideas (moves 11-12), then Nd5 and Nxd7 traded the queen for a knight (moves 15-17). Played on to the end, but 40.Qb8+ was a desperate and losing move. Mostly club-level play." }, { "game": "game_01", "color": "black", "actual": 1159, "predicted": 1600, "error": 441, "absolute_error": 441, "low_80": 1350, "high_80": 1850, "interval_covers": false, "loo_baseline": 1916.1666666666667, "rapid_games": 3260, "rating_deviation": 45, "explanation": "Took a strong center and kept up the pressure with g4 and h5. Won the queen with 15...Nxg5 and 16...Ke7, then converted well. Some king walking and inaccuracies (the Ke6 and Kd7 moves), but the 30...Bxg3 tactic and the endgame finish were accurate." }, { "game": "game_02", "color": "white", "actual": 1575, "predicted": 1450, "error": -125, "absolute_error": 125, "low_80": 1200, "high_80": 1700, "interval_covers": true, "loo_baseline": 1834.888888888889, "rapid_games": 1928, "rating_deviation": 45, "explanation": "Reasonable Ruy Lopez opening through move 12, but 18.h4 dropped a pawn and the knight. Allowing 19...Qxh4 and then 21.gxh3 wrecked the king, with 22-25 passive. Weak tactical awareness suggests mid-range club level." }, { "game": "game_02", "color": "black", "actual": 2410, "predicted": 1550, "error": -860, "absolute_error": 860, "low_80": 1300, "high_80": 1800, "interval_covers": false, "loo_baseline": 1834.888888888889, "rapid_games": 2936, "rating_deviation": 45, "explanation": "Knew the Open Spanish line with 9...Nxe4 and 14...Nc4 was active. Spotted 18...Bxh4 and 19...Qxh4, then converted cleanly with 21...Re2, 22...Rxb2 and a mating finish at 25...Rh1#. Solid but not exceptional." }, { "game": "game_03", "color": "white", "actual": 1782, "predicted": 1700, "error": -82, "absolute_error": 82, "low_80": 1450, "high_80": 1950, "interval_covers": true, "loo_baseline": 1851.6666666666667, "rapid_games": 18588, "rating_deviation": 45, "explanation": "White handled the opening sensibly and found tactics: Bxd6/Be7 (moves 16-19), Nxd5 (23) and Qxc5 (27) won material. Some inaccuracy, e.g. h4 and Ng6 knight wanderings, and the 31.c4 phase was fine. Converted cleanly with Nxg8, Qxf6+ and Re8# (46)." }, { "game": "game_03", "color": "black", "actual": 1901, "predicted": 1650, "error": -251, "absolute_error": 251, "low_80": 1400, "high_80": 1900, "interval_covers": false, "loo_baseline": 1851.6666666666667, "rapid_games": 1366, "rating_deviation": 50, "explanation": "Black's Dutch setup was playable, but 14...Bd6 allowed the Ng6 fork ideas and the Bxd6 exchange sacrifice (15-19). 22...e5 and 23...Nd7 lost a pawn to Nxd5 tactics. Later 28...Rc4 and 33...Nc8 were passive, and 34...Kh8 lost the exchange. Overall roughly club level." }, { "game": "game_04", "color": "white", "actual": 2508, "predicted": 1500, "error": -1008, "absolute_error": 1008, "low_80": 1250, "high_80": 1750, "interval_covers": false, "loo_baseline": 1765.9444444444443, "rapid_games": 1324, "rating_deviation": 47, "explanation": "White drifted in a Caro-Advance Exchange line. Qe1 and f4 (move 10) were loosening, and the queenside pawn pushes (b4, a4, a5 on moves 14-20) ignored Black's central play. Allowing ...Nxf2 tactics with Kh1 and Rxg1 (move 26) is a serious blunder. Opening knowledge looks adequate, tactical awareness weak." }, { "game": "game_04", "color": "black", "actual": 2718, "predicted": 1600, "error": -1118, "absolute_error": 1118, "low_80": 1350, "high_80": 1850, "interval_covers": false, "loo_baseline": 1765.9444444444443, "rapid_games": 1136, "rating_deviation": 59, "explanation": "Black played the opening soundly with ...Nc6, ...Qc7 and ...Bg4 (moves 4-6). The ...Ne4 and ...f5 build-up (moves 16-17) was good, and ...Nd4 (move 22) and ...Qxc5 won material. The finish with ...Ne2+, ...Qg1+ and ...Nf2# was accurate. Few mistakes, but one short game gives limited evidence." }, { "game": "game_05", "color": "white", "actual": 1120, "predicted": 1450, "error": 330, "absolute_error": 330, "low_80": 1150, "high_80": 1750, "interval_covers": false, "loo_baseline": 1890.3333333333333, "rapid_games": 5115, "rating_deviation": 45, "explanation": "Passive opening play with Bc4-d3-e2 moves, then 15.c4 and 16.Be4 allowed Nxg2+ winning a pawn and wrecking the king. 22.Bc3 and 23.Nd4 lost material to 22...Bxe4 and 23...Bxh1. Missed the threats until the mate on move 29." }, { "game": "game_05", "color": "black", "actual": 1867, "predicted": 1600, "error": -267, "absolute_error": 267, "low_80": 1300, "high_80": 1900, "interval_covers": true, "loo_baseline": 1890.3333333333333, "rapid_games": 31199, "rating_deviation": 45, "explanation": "Solid Sicilian setup with ...b5 and ...c4 gaining space. Found 15...Nf4 and 16...Nxg2+ tactic, then kept up the attack with 22...Bxe4 and 23...Bxh1, finishing with 27...Be4+ and 29...Qxf3#. Tactically sharp, though the opening was a little loose." }, { "game": "game_06", "color": "white", "actual": 1112, "predicted": 1550, "error": 438, "absolute_error": 438, "low_80": 1300, "high_80": 1800, "interval_covers": false, "loo_baseline": 1917.7222222222222, "rapid_games": 3821, "rating_deviation": 45, "explanation": "Odd opening play with Bb1 and Bb2 (moves 4-11), then missed Nd4 and b3 tactics (13-19); lost a rook on a1 and pawns by moves 18-20. Ignored threats, and 28.Qxc7+ and 30.Qxb7 walked into mate on move 31." }, { "game": "game_06", "color": "black", "actual": 1382, "predicted": 1650, "error": 268, "absolute_error": 268, "low_80": 1400, "high_80": 1900, "interval_covers": false, "loo_baseline": 1917.7222222222222, "rapid_games": 5055, "rating_deviation": 45, "explanation": "Solid setup, then good tactical play: 13...Nd4, 18...Ncxb3 and 19...Nxa1 won material. 25...Bxf1 and 27...Qxf6 were sound, and the finish with 30...Rc1+ and 31...Qf2# was clean. Some inaccuracy, but a convincing win." }, { "game": "game_07", "color": "white", "actual": 1862, "predicted": 1500, "error": -362, "absolute_error": 362, "low_80": 1250, "high_80": 1750, "interval_covers": false, "loo_baseline": 1864.3333333333333, "rapid_games": 7581, "rating_deviation": 45, "explanation": "Played a standard Slow Slav setup, but 21.Bf3 allowed 21...Bc2 and 22.Bh5 Bxd1 23.Bxg6 Bxb3 lost the queen for a bishop and the exchange, effectively losing. Earlier play (Ba3, Ne5 trades, b4 push) was reasonable but passive." }, { "game": "game_07", "color": "black", "actual": 1593, "predicted": 1550, "error": -43, "absolute_error": 43, "low_80": 1300, "high_80": 1800, "interval_covers": true, "loo_baseline": 1864.3333333333333, "rapid_games": 13102, "rating_deviation": 45, "explanation": "Solid opening with Bf5, ...Ne4 and the bishop trade at 13. Spotted the tactical 21...Bc2, then 22...Bxd1 and 23...Bxb3 winning material. Qg6 was well placed. Sound but not brilliant play; one game gives uncertainty." }, { "game": "game_08", "color": "white", "actual": 1310, "predicted": 1450, "error": 140, "absolute_error": 140, "low_80": 1200, "high_80": 1700, "interval_covers": true, "loo_baseline": 1884.0, "rapid_games": 4221, "rating_deviation": 45, "explanation": "White played a passive opening (Nc3, d3, e3) and gave up the bishop pair with 8.O-O Bxc3. At move 15 Bxc5 and 16.Ne5 were inconsistent. The knight fork was missed at 22.Rd4 and 24.h4 walked into Nf3+ mate. Several tactical oversights." }, { "game": "game_08", "color": "black", "actual": 1791, "predicted": 1550, "error": -241, "absolute_error": 241, "low_80": 1300, "high_80": 1800, "interval_covers": true, "loo_baseline": 1884.0, "rapid_games": 4152, "rating_deviation": 49, "explanation": "Black developed sensibly with e5, d5 and Bb4, and found 9...e4 and 10...Nxe4. The queen trade at 17-18 was handled fine. 20...Bg4, 22...Bxf3+ and 24...Nf3+ were good tactics. Solid but not highly precise play." }, { "game": "game_09", "color": "white", "actual": 2193, "predicted": 1750, "error": -443, "absolute_error": 443, "low_80": 1500, "high_80": 2000, "interval_covers": false, "loo_baseline": 1813.5, "rapid_games": 4525, "rating_deviation": 45, "explanation": "White got a playable Benoni-style setup but lost the thread: 12.Bd2 and 13.Qc1 were passive, 19.f4 and 21.Qxe2 allowed Bxc3 and Qxa3, winning pawns. 28.e5 and 30.f5 were speculative. Later 33-41 defence was passive and lost to the rook invasion. Mixed play of roughly club level." }, { "game": "game_09", "color": "black", "actual": 2177, "predicted": 1850, "error": -327, "absolute_error": 327, "low_80": 1600, "high_80": 2100, "interval_covers": false, "loo_baseline": 1813.5, "rapid_games": 8936, "rating_deviation": 46, "explanation": "Black handled the Benoni well with ...Na6-c7, ...b5 (10) and active knight play (13-16). Took pawns with 21...Bxc3 and 22...Qxa3, then 28...Bf5 and 35...Bxd5 kept the initiative. The finish with doubled rooks on the 2nd rank (38-41) was accurate. Solid tactics, though the game was not flawless." }, { "game": "game_10", "color": "white", "actual": 2689, "predicted": 1700, "error": -989, "absolute_error": 989, "low_80": 1450, "high_80": 1950, "interval_covers": false, "loo_baseline": 1767.9444444444443, "rapid_games": 9005, "rating_deviation": 45, "explanation": "Normal opening setup through move 15, but the middlegame drifted. Nb5 and Nxa7 (moves 20, 24) grabbed pawns while leaving the queen offside. Rxc8 on move 22 and Qc7 on move 23 let Black's pieces become active. Be5 on move 26 allowed Bxf1 and the loss of material." }, { "game": "game_10", "color": "black", "actual": 2501, "predicted": 1750, "error": -751, "absolute_error": 751, "low_80": 1500, "high_80": 2000, "interval_covers": false, "loo_baseline": 1767.9444444444443, "rapid_games": 3141, "rating_deviation": 45, "explanation": "Solid opening with ...d5, ...b6 and ...Bb7, and a sound ...c5 break. Black handled the exchanges well, and Qc2 on move 24 took advantage of White's greedy knight. Bxf1 on move 26 was accurate, and ...Ne8 on move 27 defends well. Play was fairly clean, without strong sharpness." } ], "cost_usd": 0.055644, "reasoning_effort": "medium", "max_tokens": 8192, "run_start": "2026-10-11", "run_end": "2026-10-11", "display_name": "Claude Sonnet 5.5" }, { "model": "openai/gpt-5.6-luna", "summary": { "games": 10, "player_estimates": 20, "actual_rating_range": [ 1112, 2718 ], "mae": 433.55, "rmse": 565.71719966782, "signed_bias": -348.15, "pearson": 0.5290527441898335, "spearman": 0.5244396025511762, "within": { "100": { "count": 5, "fraction": 0.25 }, "200": { "count": 9, "fraction": 0.45 }, "300": { "count": 10, "fraction": 0.5 } }, "interval_80_coverage": 0.55, "interval_80_mean_width": 775.0, "game_mean_rating_mae": 412.85, "mae_game_bootstrap_95_interval": [ 242.495, 651.4087499999998 ], "baseline_1500_mae": 506.05, "baseline_leave_one_game_out_mean_mae": 471.5666666666666, "games_beating_loo_baseline": 4, "history_count_range": [ 1136, 31199 ], "max_profile_rating_deviation": 59 }, "predictions": [ { "game": "game_01", "color": "white", "actual": 1363, "predicted": 1500, "error": 137, "absolute_error": 137, "low_80": 1100, "high_80": 1900, "interval_covers": true, "loo_baseline": 1916.1666666666667, "rapid_games": 4696, "rating_deviation": 46, "explanation": "White showed useful tactical awareness with 16.Nf6+ and 17.Nxd7, correctly exploiting Black’s 15...Nxg5. However, the opening was loose (2.d3 and 10.O-O-O), and White later overlooked 31...Bxe1, surrendering a rook. The decisive 40.Qb8+ allowed ...Rxb8 and the promotion was then easily neutralized. This suggests intermediate strength with uneven calculation and conversion." }, { "game": "game_01", "color": "black", "actual": 1159, "predicted": 1450, "error": 291, "absolute_error": 291, "low_80": 1050, "high_80": 1850, "interval_covers": true, "loo_baseline": 1916.1666666666667, "rapid_games": 3260, "rating_deviation": 45, "explanation": "Black played the central opening actively with ...d5 and ...f5, but 15...Nxg5? allowed the major tactical sequence losing the queen on d7. Black then found practical counterplay, especially 30...Bxg3 and 31...Bxe1, and defended the advanced e-pawn effectively after 40.Qb8+. The large tactical swings indicate intermediate, inconsistent play rather than advanced calculation." }, { "game": "game_02", "color": "white", "actual": 1575, "predicted": 1200, "error": -375, "absolute_error": 375, "low_80": 850, "high_80": 1650, "interval_covers": true, "loo_baseline": 1834.888888888889, "rapid_games": 1928, "rating_deviation": 45, "explanation": "White showed reasonable opening familiarity through 8.Nd5 and 10.d4, but the middlegame contained major tactical oversights. After 18.h4? Bxh4!, White allowed the bishop sacrifice and queen invasion, then 20.Qh3 led to a materially inferior position. The later 24.Bc2?? allowed ...Rxh3, and 25.Bg5?? missed the immediate ...Rh1 mate. This looks substantially below advanced club level, though the opening prevents a very low estimate." }, { "game": "game_02", "color": "black", "actual": 2410, "predicted": 1750, "error": -660, "absolute_error": 660, "low_80": 1350, "high_80": 2150, "interval_covers": false, "loo_baseline": 1834.888888888889, "rapid_games": 2936, "rating_deviation": 45, "explanation": "Black handled the opening coherently, especially 9...Nxe4 and the consolidation with ...Nf6, ...g6, and ...Re8. The sequence 18...Bxh4! 19...Qxh4 showed effective tactical calculation, and 21...Re2 followed by ...Rxb2 and ...Rxh3 demonstrated good attacking coordination. Black’s finish with 25...Rh1# was alert and accurate. The game mainly tests tactics against a weak defense, so the rating estimate remains broad." }, { "game": "game_03", "color": "white", "actual": 1782, "predicted": 1750, "error": -32, "absolute_error": 32, "low_80": 1400, "high_80": 2100, "interval_covers": true, "loo_baseline": 1851.6666666666667, "rapid_games": 18588, "rating_deviation": 45, "explanation": "White showed good tactical awareness with 15.Ng6, 16.Bxd6, and the forcing 17.Ne5+. The attack was handled coherently, and 23.Nxd5, 27.Qxc5, and especially 37.Rd8 exploited Black’s weaknesses. The conversion after winning the queen was clean, though 8.h4 and 9.Ng5 were speculative opening choices." }, { "game": "game_03", "color": "black", "actual": 1901, "predicted": 1350, "error": -551, "absolute_error": 551, "low_80": 1000, "high_80": 1750, "interval_covers": false, "loo_baseline": 1851.6666666666667, "rapid_games": 1366, "rating_deviation": 50, "explanation": "Black developed reasonably but played several strategically risky moves, notably 16...Kf7 and 18...Re8, leaving the king exposed. The defensive decisions around 23...Nd7 and 26...Nb6 ceded control, and 37...Ne7 overlooked the decisive rook attack on the queen at g8. Some active ideas appeared, but the tactical oversight and king handling suggest a lower established rating." }, { "game": "game_04", "color": "white", "actual": 2508, "predicted": 1500, "error": -1008, "absolute_error": 1008, "low_80": 1100, "high_80": 1900, "interval_covers": false, "loo_baseline": 1765.9444444444443, "rapid_games": 1324, "rating_deviation": 47, "explanation": "White showed a coherent aggressive plan with 5.c3, 10.f4, and queenside expansion by 14.b4 and 16.a4, but the attack was not well coordinated. After 21.c4 and 22.b5, White allowed ...Nd4 and then missed the decisive tactical sequence with 23.Qb2, 24.Bb4, and 25...Qg1#." }, { "game": "game_04", "color": "black", "actual": 2718, "predicted": 1750, "error": -968, "absolute_error": 968, "low_80": 1350, "high_80": 2150, "interval_covers": false, "loo_baseline": 1765.9444444444443, "rapid_games": 1136, "rating_deviation": 59, "explanation": "Black demonstrated solid Caro-Kann development and purposeful counterplay, including 13...O-O-O, 15...Rc8, and 17...f5. The sequence 21...gxf4, 22...Nd4, and especially 24...Ne2+ followed by 25...Qg1# shows good tactical awareness, though the game was short and does not establish consistently strong endgame or defensive skill." }, { "game": "game_05", "color": "white", "actual": 1120, "predicted": 1300, "error": 180, "absolute_error": 180, "low_80": 1000, "high_80": 1650, "interval_covers": true, "loo_baseline": 1890.3333333333333, "rapid_games": 5115, "rating_deviation": 45, "explanation": "White showed limited opening efficiency with 2.Bc4, 5.Bd3, and 6.Be2, then allowed Black strong activity after 15.c4. The tactical oversight 22.Bc3? Bxe4 23.Nd4?? Bxh1 lost a rook, followed by inadequate king defense and mate on move 29. There were some reasonable developing moves, but the tactical and defensive errors suggest a lower-intermediate established rapid rating." }, { "game": "game_05", "color": "black", "actual": 1867, "predicted": 1750, "error": -117, "absolute_error": 117, "low_80": 1450, "high_80": 2050, "interval_covers": true, "loo_baseline": 1890.3333333333333, "rapid_games": 31199, "rating_deviation": 45, "explanation": "Black developed a coherent attacking plan despite the unusual opening: ...b5, ...c4, ...f5, and ...Nf4 created sustained pressure. The sequence 16...Nxg2+, 21...Nbd3, 22...Bxe4, and 23...Bxh1 showed good tactical awareness, while the final ...Nh3+ and ...Qxf3# demonstrated effective mating-net calculation. Some early moves were committal, so the game supports solid intermediate strength rather than an expert-level estimate." }, { "game": "game_06", "color": "white", "actual": 1112, "predicted": 1200, "error": 88, "absolute_error": 88, "low_80": 800, "high_80": 1600, "interval_covers": true, "loo_baseline": 1917.7222222222222, "rapid_games": 3821, "rating_deviation": 45, "explanation": "White showed some opening and attacking ideas with 10.e4, 16.f4, and 24.Nf6+, but repeatedly neglected concrete threats. In particular, 18...Ncxb3 followed by 19...Nxa1 lost a rook for a knight, and 25.Rf1 allowed ...Bxf1. The final 27.Ke1 and 28.Qxc7 did not address the mating danger, suggesting intermediate club strength despite active play." }, { "game": "game_06", "color": "black", "actual": 1382, "predicted": 1450, "error": 68, "absolute_error": 68, "low_80": 1050, "high_80": 1850, "interval_covers": true, "loo_baseline": 1917.7222222222222, "rapid_games": 5055, "rating_deviation": 45, "explanation": "Black demonstrated useful tactical awareness: 18...Nc5 and ...Ncxb3 exploited White’s loose queenside, while ...Nxa1 and 23...Ba6 created the decisive bishop route to f1. The sequence 25...Bxf1, 26...Nxg5, and 27...Qxf6 efficiently converted material and the final ...Rc1+/...Qf2# showed mating-pattern recognition. Opening play was somewhat imprecise, so the game does not indicate advanced strength." }, { "game": "game_07", "color": "white", "actual": 1862, "predicted": 1250, "error": -612, "absolute_error": 612, "low_80": 900, "high_80": 1700, "interval_covers": false, "loo_baseline": 1864.3333333333333, "rapid_games": 7581, "rating_deviation": 45, "explanation": "White showed reasonable opening and middlegame structure with 8.cxd5, 11.Ne5, and active queenside play via 17.b4 and 19.b5. However, 21.Bf3 allowed the clear bishop fork 21...Bc2, and 22.Bh5 ignored attacks on both the queen and rook. The sequence 22...Bxd1 23.Bxg6 Bxb3 is a major tactical oversight." }, { "game": "game_07", "color": "black", "actual": 1593, "predicted": 1500, "error": -93, "absolute_error": 93, "low_80": 1150, "high_80": 1900, "interval_covers": true, "loo_baseline": 1864.3333333333333, "rapid_games": 13102, "rating_deviation": 45, "explanation": "Black handled the opening sensibly, developed efficiently, and found the important tactical idea 21...Bc2, forking the queen on d1 and rook on b3. The conversion 22...Bxd1 23...Bxb3 was accurate and decisive. Earlier moves such as 11...Ne4 and 18...Qe6 were coherent, though this short game gives limited evidence about deeper positional or endgame strength." }, { "game": "game_08", "color": "white", "actual": 1310, "predicted": 1400, "error": 90, "absolute_error": 90, "low_80": 1000, "high_80": 1800, "interval_covers": true, "loo_baseline": 1884.0, "rapid_games": 4221, "rating_deviation": 45, "explanation": "The opening choices (1 Nc3, 2 d3, 4 Bd2) are passive and White allowed ...e4 on move 9 and ...c5 on move 14. White made some sensible developing and simplifying moves, including 16 Ne5 and 18 Qxb6, but 22 Rd4 allowed ...Bxf3+ and 24 h4 overlooked ...Nf3+, winning the rook. This suggests modest strength, with substantial uncertainty from one game." }, { "game": "game_08", "color": "black", "actual": 1791, "predicted": 1650, "error": -141, "absolute_error": 141, "low_80": 1250, "high_80": 2050, "interval_covers": true, "loo_baseline": 1884.0, "rapid_games": 4152, "rating_deviation": 49, "explanation": "Black developed naturally, castled early, and used the central breaks ...e4 and ...c5 effectively. The sequence 14...c5, 17...Nxe4, and 18...axb6 was coherent, while 22...Bxf3+ exploited the exposed knight and 24...Nf3+ created a decisive rook fork. The play is competent and tactically alert, though the short game gives limited evidence about deeper endgame skill." }, { "game": "game_09", "color": "white", "actual": 2193, "predicted": 1450, "error": -743, "absolute_error": 743, "low_80": 1100, "high_80": 1850, "interval_covers": false, "loo_baseline": 1813.5, "rapid_games": 4525, "rating_deviation": 45, "explanation": "White showed reasonable opening development, but 19.f4 and especially 20.Qd3 allowed ...Nd4, ...Nxe2, and ...Bxc3, leaving Black materially ahead. The rook activity on moves 25, 32, and 37 was resourceful, yet 39.R8xc5 permitted a decisive kingside attack. This looks like competent tactical awareness mixed with substantial calculation and defensive shortcomings." }, { "game": "game_09", "color": "black", "actual": 2177, "predicted": 1750, "error": -427, "absolute_error": 427, "low_80": 1400, "high_80": 2150, "interval_covers": false, "loo_baseline": 1813.5, "rapid_games": 8936, "rating_deviation": 46, "explanation": "Black handled the opening actively with 10...b5 and coordinated the tactical sequence 16...Nxd2, 19...Nd4, 20...Nxe2, and 21...Bxc3. The conversion was generally purposeful, using queen penetration on moves 22–26 and rook activity from move 38 onward. Some opportunities were not maximally precise, but the strategic and tactical execution was clearly stronger than White’s." }, { "game": "game_10", "color": "white", "actual": 2689, "predicted": 1450, "error": -1239, "absolute_error": 1239, "low_80": 1050, "high_80": 1900, "interval_covers": false, "loo_baseline": 1767.9444444444443, "rapid_games": 9005, "rating_deviation": 45, "explanation": "White handled the opening sensibly through 15 Qf4 and found the useful exchange on c8 with 22 Rxc8. However, 24 Nxa7 was greedy, and after 25...Ba6, 26 Be5?? failed to notice the diagonal attack on the rook at f1, losing a full exchange to 26...Bxf1. The generally coherent play is consistent with an intermediate established rating, but the decisive tactical oversight widens the uncertainty." }, { "game": "game_10", "color": "black", "actual": 2501, "predicted": 1650, "error": -851, "absolute_error": 851, "low_80": 1200, "high_80": 2100, "interval_covers": false, "loo_baseline": 1767.9444444444443, "rapid_games": 3141, "rating_deviation": 45, "explanation": "Black developed naturally and maintained sound central play, with 16...fxe6, 17...Bd6, and active queen maneuvers. The tactical sequence 24...Qc2 followed by 25...Ba6 exploited the vulnerability of f1, and 26...Bxf1 was well spotted. There is too little evidence of deeper endgame or defensive skill from this short game, so the estimate remains broad despite Black’s effective tactical conversion." } ], "cost_usd": 0.0169628, "reasoning_effort": "medium", "max_tokens": 8192, "run_start": "2026-10-11", "run_end": "2026-10-11", "display_name": "GPT-5.6 Luna" }, { "model": "z-ai/glm-5.3-flash", "summary": { "games": 10, "player_estimates": 20, "actual_rating_range": [ 1112, 2718 ], "mae": 456.95, "rmse": 567.3851866236904, "signed_bias": -235.65, "pearson": 0.14977740274435578, "spearman": 0.08771407590923147, "within": { "100": { "count": 1, "fraction": 0.05 }, "200": { "count": 6, "fraction": 0.3 }, "300": { "count": 9, "fraction": 0.45 } }, "interval_80_coverage": 0.35, "interval_80_mean_width": 470.0, "game_mean_rating_mae": 445.25, "mae_game_bootstrap_95_interval": [ 290.89875, 661.0099999999999 ], "baseline_1500_mae": 506.05, "baseline_leave_one_game_out_mean_mae": 471.5666666666666, "games_beating_loo_baseline": 4, "history_count_range": [ 1136, 31199 ], "max_profile_rating_deviation": 59 }, "predictions": [ { "game": "game_01", "color": "white", "actual": 1363, "predicted": 1650, "error": 287, "absolute_error": 287, "low_80": 1450, "high_80": 1850, "interval_covers": false, "loo_baseline": 1916.1666666666667, "rapid_games": 4696, "rating_deviation": 46, "explanation": "Reasonable opening handling but 8.Bg5/9.Bd2 wastes tempi; 19.f3 and 21.Nxf3 avoid material loss yet White drifts. After 30.dxe4, 40.Qb8+ Rxb8 41.e8=Q shows tactical awareness in a lost position, but overall imprecise. Likely intermediate club level, wide interval from one game." }, { "game": "game_01", "color": "black", "actual": 1159, "predicted": 1750, "error": 591, "absolute_error": 591, "low_80": 1550, "high_80": 1950, "interval_covers": false, "loo_baseline": 1916.1666666666667, "rapid_games": 3260, "rating_deviation": 45, "explanation": "Black's 9...g5/11...g4 kingside clamp shows decent planning, and 15...Nxg5 16.Nf6+ wins material. Weaker moments: 22...Rab8 slow, and 30...Bxg3 31.e5 Bxe1 trades bishop for rook carelessly. Nonetheless converts the exchange-up endgame methodically with h4-h3 and Re8, finally landing the 40...Rxb8/41...Rxe8 sequence. Solid club strength." }, { "game": "game_02", "color": "white", "actual": 1575, "predicted": 1450, "error": -125, "absolute_error": 125, "low_80": 1250, "high_80": 1700, "interval_covers": true, "loo_baseline": 1834.888888888889, "rapid_games": 1928, "rating_deviation": 45, "explanation": "White knows mainstream Ruy Lopez theory to move 17, but 18.h4?? loses a piece to 18...Bxh4 19.Nxh4 Qxh4, and 20.Qh3 also drops the queen. Missing Black's mating net (25...Rh1#) after 21.gxh3 shows weak tactical vision. Solid book knowledge but poor calculation suggests a mid-level amateur, roughly 1400-1500." }, { "game": "game_02", "color": "black", "actual": 2410, "predicted": 1700, "error": -710, "absolute_error": 710, "low_80": 1500, "high_80": 1950, "interval_covers": false, "loo_baseline": 1834.888888888889, "rapid_games": 2936, "rating_deviation": 45, "explanation": "Black chose a sharp theoretical line (8...d6, 9...Nxe4), knew the 18...Bxh4 trap, converted cleanly with 20...Qxh3, 21...Rxe2 grabbing material, and finished with the precise rook mate 25...Rh1#. No errors and deep trap knowledge indicate a strong club player, roughly 1650-1800." }, { "game": "game_03", "color": "white", "actual": 1782, "predicted": 1650, "error": -132, "absolute_error": 132, "low_80": 1450, "high_80": 1850, "interval_covers": true, "loo_baseline": 1851.6666666666667, "rapid_games": 18588, "rating_deviation": 45, "explanation": "White knew setup ideas (h4-h5, Ng5/Nh3 maneuvers, 15.Ng6-winning exchange trick vs 14...Bd6). Tactical conversion 25-38 (deflection 37.Rd8, mating net 45.Re8) was clean, with decent technique simplifying to a winning endgame. Some looseness: 22.Nhf4 allowed 22...e5 counterplay. Overall solid club-level tactical player; single-game uncertainty keeps range wide." }, { "game": "game_03", "color": "black", "actual": 1901, "predicted": 1450, "error": -451, "absolute_error": 451, "low_80": 1250, "high_80": 1650, "interval_covers": false, "loo_baseline": 1851.6666666666667, "rapid_games": 1366, "rating_deviation": 50, "explanation": "Black plays a recognizable Stonewall/Lennon setup but misses tactics: 14...Bd6 loses a piece to 15.Ng6; 18...Re8?/19...gxf6 wrecked kingside; 45...Qxg6? allowed mate (46.Re8#) after 44...Kh8 opened the wrong diagonal—should contest the e-file. Some defensive resourcefulness (26...Nb6, 27...Rac8, 28...Rc4 counterplay) shows club competence but calculation is weak." }, { "game": "game_04", "color": "white", "actual": 2508, "predicted": 1450, "error": -1058, "absolute_error": 1058, "low_80": 1200, "high_80": 1750, "interval_covers": false, "loo_baseline": 1765.9444444444443, "rapid_games": 1324, "rating_deviation": 47, "explanation": "Solid Caro-Kann structure but several soft moves: 8.Qe1 and 13.Bh4 lacked purpose, 17.Be1 retreats awkwardly, and White allowed Black's knight into e4/d4. Missing the 24...Ne2+ 25...Qg1 26.Nf2# mate pattern after 25.Kh1 (instead of 25.Bg1/Kh1 alternatives addressing f2) is a serious oversight for ~1500 level." }, { "game": "game_04", "color": "black", "actual": 2718, "predicted": 1600, "error": -1118, "absolute_error": 1118, "low_80": 1350, "high_80": 1850, "interval_covers": false, "loo_baseline": 1765.9444444444443, "rapid_games": 1136, "rating_deviation": 59, "explanation": "Good opening understanding: 6...Bg4/9...Bh5, castling long with counterplay, and accurate tactics — 19...g5! undermining, 21...gxf4, then the clean mating combination 22...Nd4, 24...Ne2+, 25...Qg1+ 26.Rxg1 Nf2# (a Philidor-style mate). Some quieter moments (e.g., 20...a6, 17...f5) were less precise, suggesting ~1600." }, { "game": "game_05", "color": "white", "actual": 1120, "predicted": 1300, "error": 180, "absolute_error": 180, "low_80": 1000, "high_80": 1650, "interval_covers": true, "loo_baseline": 1890.3333333333333, "rapid_games": 5115, "rating_deviation": 45, "explanation": "White mishandled the opening (2.Bc4 vs Sicilian, allowed 5...c4), then blundered tactically: 15.c4? (16.Be4 Nxg2+ loses a piece to 17...fxe4), and passively allowed Black's knight-infiltration 21...Nbd3. Defenses 25.f3 and 26.Kf1 were forced but weak; White never coordinated. Mediocre calculation and repeated tactical misses suggest a lower-intermediate rapid rating." }, { "game": "game_05", "color": "black", "actual": 1867, "predicted": 1750, "error": -117, "absolute_error": 117, "low_80": 1450, "high_80": 2050, "interval_covers": true, "loo_baseline": 1890.3333333333333, "rapid_games": 31199, "rating_deviation": 45, "explanation": "Black showed strong tactical vision: 11...f5! gaining space, 15...Nf4 and the brilliant 16...Nxg2+ sacrifice, followed by 17...fxe4 winning material. Quiet consolidation 18...Nf4, regrouping 20...Nb4/21...Nbd3, and a clean forced sequence 25...Nh3+ 26...Bxf3 27...Be4+ 29...Qxf3# show good calculation and attacking technique, though the opening (4...b5?! 5...c4) was slightly eccentric." }, { "game": "game_06", "color": "white", "actual": 1112, "predicted": 1350, "error": 238, "absolute_error": 238, "low_80": 1100, "high_80": 1600, "interval_covers": true, "loo_baseline": 1917.7222222222222, "rapid_games": 3821, "rating_deviation": 45, "explanation": "White lost a piece outright to 13...Nd4 (13.Ng5 Nd4, then 17...Nc5 18.Qe3 Ncxb3 trading into 19...Nxa1 winning material). Later errors: 19.f5? drops ideas, 24.Nf6+?? throws the knight (27...Qxf6), and defence is passive (25.Rf1 Bxf1, 30.Qxb7 missing 31...Qf2#). Basic opening ideas okay (English setup), tactics clearly weak—solidly sub-1500, likely ~1350." }, { "game": "game_06", "color": "black", "actual": 1382, "predicted": 1650, "error": 268, "absolute_error": 268, "low_80": 1400, "high_80": 1850, "interval_covers": false, "loo_baseline": 1917.7222222222222, "rapid_games": 5055, "rating_deviation": 45, "explanation": "Black spotted the 13...Nd4/17...Nc5/19...Nxa1 tactic, played solidly thereafter: 20...Bxc4, 22...Ne6 and 23...Ba6 developing, converted with 24...Kg7, 27...Qxf6, then 28...Kh6, 29...Rac8 and a clean mating net (30...Rc1+ 31...Qf2#). Some looseness early (8...h5, 9...Bf5-10...Bd7), and the win took many moves versus a weaker opponent—suggests ~1650, not much higher." }, { "game": "game_07", "color": "white", "actual": 1862, "predicted": 1500, "error": -362, "absolute_error": 362, "low_80": 1300, "high_80": 1750, "interval_covers": false, "loo_baseline": 1864.3333333333333, "rapid_games": 7581, "rating_deviation": 45, "explanation": "Played a plausible Slav setup (8.cxd5, 13.Ba3 trade, 17.b4-b5 plan) but overlooked 21...Bc2 hitting Rb3, then blundered 22.Bh5 allowing 22...Bxd1/23...Bxb3 netting a rook. In a roughly equal/defensible material balance (24.Bd3, two rooks vs queen) White still lost 0-1, suggesting flagging/poor calculation. Competent opening but shaky tactics/time management point near 1500." }, { "game": "game_07", "color": "black", "actual": 1593, "predicted": 1550, "error": -43, "absolute_error": 43, "low_80": 1350, "high_80": 1800, "interval_covers": true, "loo_baseline": 1864.3333333333333, "rapid_games": 13102, "rating_deviation": 45, "explanation": "Solid Slav development (3...Bf5, 9...Bd6, 11...Ne4) with standard trades; 15...Nd7 and 18...Qe6 show routine understanding. Best moment: 21...Bc2 striking Rb3 and d1, exploiting White's 22.Bh5 with 22...Bxd1 and 23...Bxb3. Minor slow moves (5...h6, 9...Bd6) and no deep endgame shown; converting with tactical shots but against an erring opponent suggests mid-1500s rapid strength." }, { "game": "game_08", "color": "white", "actual": 1310, "predicted": 1850, "error": 540, "absolute_error": 540, "low_80": 1600, "high_80": 2100, "interval_covers": false, "loo_baseline": 1884.0, "rapid_games": 4221, "rating_deviation": 45, "explanation": "White's opening setup was coherent (King's Indian-style with g3, Bd2). Move 11 Bb4 allowed ...Be6 and 12.Qd4 invited ...Nd7 hitting the bishop. 14.Rfe1 was slow; 19.a3 f6 then 20.Nf3 lost time. Errors at 22.Rd4 (pinning own issues) and 24.h4 blundered into ...Nf3+ fork winning Rd4. Solid but tactically careless; ~1850." }, { "game": "game_08", "color": "black", "actual": 1791, "predicted": 1900, "error": 109, "absolute_error": 109, "low_80": 1650, "high_80": 2150, "interval_covers": true, "loo_baseline": 1884.0, "rapid_games": 4152, "rating_deviation": 49, "explanation": "Black handled the opening well: 3...Bb4, 8...Bxc3 with 9...e4 gaining space and bishop pair. Clean tactics: 12...Be6, 14...c5, 15...Ndxc5, 16...Qb6 creating threats. 19...f6 and 21...Ng5, 22...Bxf3+, 23...Be4 set up a fork; 24...Nf3+ won the exchange and the game. Confident technique, ~1900." }, { "game": "game_09", "color": "white", "actual": 2193, "predicted": 1600, "error": -593, "absolute_error": 593, "low_80": 1400, "high_80": 1850, "interval_covers": false, "loo_baseline": 1813.5, "rapid_games": 4525, "rating_deviation": 45, "explanation": "White plays a reasonable Benko setup but errs repeatedly: 23.Ra1? allows 23...Qxc3 winning material, 25.Rxa7 grabs a pawn while letting 25...Rb2, and 26.Qf3 loses to 26...Qd2. Later 37.Rc8/39.R8xc5 trades rooks passively in a lost position. Solid opening understanding and basic tactics, but weak defensive awareness suggest ~1600, with wide uncertainty from a single game." }, { "game": "game_09", "color": "black", "actual": 2177, "predicted": 1850, "error": -327, "absolute_error": 327, "low_80": 1600, "high_80": 2100, "interval_covers": false, "loo_baseline": 1813.5, "rapid_games": 8936, "rating_deviation": 46, "explanation": "Black handles the Benko/Grunfeld structure well: purposeful regrouping 9...Nc7 and 10...b5, clean conversion 21...Bxc3/22...Qxa3/23...Qxc3, active rook play 25...Rb2, and accurate 32...Qf4+ liquidation into a winning rook endgame. Minor imprecisions (28...Bf5, 30...dxe5 allowed fxg6) keep it below expert level. Overall confident, near-blunder-free play suggests ~1850." }, { "game": "game_10", "color": "white", "actual": 2689, "predicted": 1600, "error": -1089, "absolute_error": 1089, "low_80": 1400, "high_80": 1800, "interval_covers": false, "loo_baseline": 1767.9444444444443, "rapid_games": 9005, "rating_deviation": 45, "explanation": "Solid English opening setup, but White loses the plot: 18.Qh4 allows ...Rf8 hitting the queen, 23.Qc7 walks into trouble, and 24.Nxa7 grabs a pawn while letting Black's 24...Qc2 infiltrate. 26.Be5 overlooks 26...Bxf1 winning the exchange; White resigns a piece down with no compensation. Reasonable development but weak tactics and defense suggest ~1600." }, { "game": "game_10", "color": "black", "actual": 2501, "predicted": 1700, "error": -801, "absolute_error": 801, "low_80": 1500, "high_80": 1900, "interval_covers": false, "loo_baseline": 1767.9444444444443, "rapid_games": 3141, "rating_deviation": 45, "explanation": "Comfortable Queen's Indian structure with good piece placement (...Nc5, ...Ne6 regrouping). Exploits White's lapses: 24...Qc2 invades the second rank, 25...Ba6 and 26...Bxf1 simplifies into a winning material-up endgame. Steady, accurate conversion after winning the exchange, though few tactics were needed. Consistent club-level play estimates ~1700." } ], "cost_usd": 0.002859448, "reasoning_effort": "medium", "max_tokens": 8192, "run_start": "2026-10-11", "run_end": "2026-10-11", "display_name": "GLM-5.3 Flash" }, { "model": "anthropic/claude-opus-5.5", "summary": { "games": 10, "player_estimates": 20, "actual_rating_range": [ 1112, 2718 ], "mae": 463.35, "rmse": 594.2356014242163, "signed_bias": -435.65, "pearson": 0.7027141076986881, "spearman": 0.70695589972226, "within": { "100": { "count": 4, "fraction": 0.2 }, "200": { "count": 6, "fraction": 0.3 }, "300": { "count": 8, "fraction": 0.4 } }, "interval_80_coverage": 0.4, "interval_80_mean_width": 502.5, "game_mean_rating_mae": 456.25, "mae_game_bootstrap_95_interval": [ 271.2975, 680.9575 ], "baseline_1500_mae": 506.05, "baseline_leave_one_game_out_mean_mae": 471.5666666666666, "games_beating_loo_baseline": 4, "history_count_range": [ 1136, 31199 ], "max_profile_rating_deviation": 59 }, "predictions": [ { "game": "game_01", "color": "white", "actual": 1363, "predicted": 1250, "error": -113, "absolute_error": 113, "low_80": 1000, "high_80": 1500, "interval_covers": true, "loo_baseline": 1916.1666666666667, "rapid_games": 4696, "rating_deviation": 46, "explanation": "White's opening was passive (2.d3), but he exploited 14...Nh7? with 15.Nd5 and 16.Nf6+, winning the queen. He then squandered it: 19.f3 and 21.Nxf3 gave back material, 30.dxe4 allowed ...Bxg3 forking the rooks, and 40.Qb8+?? threw away the queen to a simple recapture before 42.Rxe8 lost everything. These are repeated tactical blindness and poor technique, typical of club level." }, { "game": "game_01", "color": "black", "actual": 1159, "predicted": 1200, "error": 41, "absolute_error": 41, "low_80": 950, "high_80": 1450, "interval_covers": true, "loo_baseline": 1916.1666666666667, "rapid_games": 3260, "rating_deviation": 45, "explanation": "Black's opening was loose and over-aggressive (4...f5, 9...g5, 12...h5). 14...Nh7?? dropped the queen to the Nd5-Nf6+ fork. Black did fight back well, using the g-file pawns (20...gxf3) and spotting 30...Bxg3 to win a rook. He also punished 40.Qb8+?? correctly. The queen blunder and the shaky play overall suggest a rating similar to White's." }, { "game": "game_02", "color": "white", "actual": 1575, "predicted": 1200, "error": -375, "absolute_error": 375, "low_80": 950, "high_80": 1450, "interval_covers": false, "loo_baseline": 1834.888888888889, "rapid_games": 1928, "rating_deviation": 45, "explanation": "Reasonable Ruy Lopez setup, but 8.Nd5 and 9.c3 drop the e4 pawn. 18.h4?? gives away a piece for nothing after 18...Bxh4 19.Nxh4 Qxh4. 20.Qh3 trades into a lost ending and wrecks the kingside. White then drifted with 23.a3 and 25.Bg5, missing a back-rank/h-file mate in one. These are multiple serious tactical oversights." }, { "game": "game_02", "color": "black", "actual": 2410, "predicted": 1350, "error": -1060, "absolute_error": 1060, "low_80": 1100, "high_80": 1600, "interval_covers": false, "loo_baseline": 1834.888888888889, "rapid_games": 2936, "rating_deviation": 45, "explanation": "Solid opening, and Black grabbed e4 correctly with 9...Nxe4. 13...Na5 and 14...Nc4 are sensible. Black spotted 18...Bxh4, then activated the rook with 21...Re2 and collected pawns with 22...Rxb2 and 24...Rxh3 before mating with 25...Rh1#. Mostly simple punishment of errors; little was tested against strong resistance." }, { "game": "game_03", "color": "white", "actual": 1782, "predicted": 1550, "error": -232, "absolute_error": 232, "low_80": 1300, "high_80": 1800, "interval_covers": true, "loo_baseline": 1851.6666666666667, "rapid_games": 18588, "rating_deviation": 45, "explanation": "Reasonable London setup, but 8.h4 and 9.Ng5 are aimless. White capitalised on Black's errors: 16.Bxd6 winning a piece (Black forgot recapture), 23.Nxd5 picking up material, 34.Ne7+ fork winning the exchange, and a clean finish with 43.Bxg6+ and 46.Re8#. Some inaccuracies, like 18.Be7 and 19.Bxf6 giving back material, but overall a solid conversion typical of club level." }, { "game": "game_03", "color": "black", "actual": 1901, "predicted": 1300, "error": -601, "absolute_error": 601, "low_80": 1050, "high_80": 1550, "interval_covers": false, "loo_baseline": 1851.6666666666667, "rapid_games": 1366, "rating_deviation": 50, "explanation": "Sensible Dutch development, but 13...d5 and 14...Bd6 allowed 15.Ng6 and 16.Bxd6, losing a piece because the queen couldn't recapture; 15...Rd8 ignored the threat. 22...e5 dropped the d5 pawn, 29...Rg8 and 33...Nc8 allowed the Ne7+ fork, and 44...Kh8 hung the rook. Repeated tactical oversights suggest a lower club level." }, { "game": "game_04", "color": "white", "actual": 2508, "predicted": 1450, "error": -1058, "absolute_error": 1058, "low_80": 1200, "high_80": 1700, "interval_covers": false, "loo_baseline": 1765.9444444444443, "rapid_games": 1324, "rating_deviation": 47, "explanation": "Reasonable Caro-Kann Exchange setup through move 9, but 8.Qe1 is odd and 10.f4 weakens e3/e4. 18.Nc5 dxc5 gave up structure, and 21.c4 and 22.b5 left the king exposed. 23.Qb2 dropped c5. The decisive blunder was 24.Bb4?? (or 25.Kh1), walking into a classic smothered mate via Qg1+ and Nf2#." }, { "game": "game_04", "color": "black", "actual": 2718, "predicted": 1550, "error": -1168, "absolute_error": 1168, "low_80": 1300, "high_80": 1800, "interval_covers": false, "loo_baseline": 1765.9444444444443, "rapid_games": 1136, "rating_deviation": 59, "explanation": "Solid development with 6...Bg4 and 10...Bxe2. Queenside castling and 14...Kb8 were sensible, and 16...Ne4 and 17...f5 built a strong outpost. 21...gxf4 and 22...Nd4 were active, and 23...Qxc5 won a pawn. Black spotted the smothered mate 24...Ne2+ 25...Qg1+ 26...Nf2#. Some loose choices earlier (12...h6, 18...Bxc5) keep the estimate moderate." }, { "game": "game_05", "color": "white", "actual": 1120, "predicted": 1150, "error": 30, "absolute_error": 30, "low_80": 900, "high_80": 1400, "interval_covers": true, "loo_baseline": 1890.3333333333333, "rapid_games": 5115, "rating_deviation": 45, "explanation": "White's opening was aimless: 2.Bc4, then 5.Bd3 and 6.Be2 wasted tempi. 15.c4 allowed ...Nf4, and 16.Be4?? walked into Nxg2+, losing the king's safety. Then 19.Nfg5 and 20.Nf3 shuffled pointlessly. 23.Nd4 dropped the rook on h1, and 27.Qxd3 and 28.Nf3 hastened mate. These are repeated tactical oversights typical of a low-rated player." }, { "game": "game_05", "color": "black", "actual": 1867, "predicted": 1400, "error": -467, "absolute_error": 467, "low_80": 1150, "high_80": 1700, "interval_covers": false, "loo_baseline": 1890.3333333333333, "rapid_games": 31199, "rating_deviation": 45, "explanation": "Black's queenside expansion (3...a6, 4...b5, 5...c4) was sensible. Black spotted 16...Nxg2+ and 22...Bxe4, collected material efficiently and finished with a clean mating net (25...Nh3+, 29...Qxf3#). However, 11...f5 and the loose knights went unpunished only because the opposition was weak. Solid tactics overall, but untested." }, { "game": "game_06", "color": "white", "actual": 1112, "predicted": 1250, "error": 138, "absolute_error": 138, "low_80": 1000, "high_80": 1500, "interval_covers": true, "loo_baseline": 1917.7222222222222, "rapid_games": 3821, "rating_deviation": 45, "explanation": "Odd opening (6.Bb1, 8.h4) without development purpose. 13.Ng5 into a trade allowing hxg5 was fine, but 18.Qe3?? hung b3 and the Ra1 fork (18...Ncxb3). 25.Rf1?? dropped the exchange, and 27.Ke1 left Nf6 hanging. Plenty of activity, but tactical awareness is poor and king safety was ignored, ending in 31.Ke2 allowing mate." }, { "game": "game_06", "color": "black", "actual": 1382, "predicted": 1450, "error": 68, "absolute_error": 68, "low_80": 1200, "high_80": 1700, "interval_covers": true, "loo_baseline": 1917.7222222222222, "rapid_games": 5055, "rating_deviation": 45, "explanation": "Solid but slow opening with bishop wandering (Bd7-f5-d7-g4-e6). 16...Bxg5 gave the h-file, yet 18...Ncxb3 spotted the fork and 20...Bxc4 won a pawn. 25...Bxf1, 26...Nxg5 and 27...Qxf6 punished errors, and 30...Rc1+ 31...Qf2# finished cleanly. Allowing 24.Nf6+ shows some looseness." }, { "game": "game_07", "color": "white", "actual": 1862, "predicted": 1200, "error": -662, "absolute_error": 662, "low_80": 950, "high_80": 1450, "interval_covers": false, "loo_baseline": 1864.3333333333333, "rapid_games": 7581, "rating_deviation": 45, "explanation": "Opening play was sensible through move 14, but 15.Ne5 and 16.Nxd7 traded pieces aimlessly, and the pawn pushes on moves 17-20 had no clear plan. 21.Bf3?? allowed 21...Bc2, hitting the queen and rook. 22.Bh5?? then dropped the queen outright, and White resigned at move 24. These are severe tactical oversights typical of lower-club play." }, { "game": "game_07", "color": "black", "actual": 1593, "predicted": 1350, "error": -243, "absolute_error": 243, "low_80": 1100, "high_80": 1600, "interval_covers": true, "loo_baseline": 1864.3333333333333, "rapid_games": 13102, "rating_deviation": 45, "explanation": "Black played a solid Slav setup with ...Bf5, ...h6 and ...Nbd7. 11...Ne4 and the bishop trades on moves 12-13 were reasonable. After 18...Qe6 and 20...Qg6, Black spotted 21...Bc2 and took the queen with 22...Bxd1 for a winning material edge. Otherwise the play was unremarkable and not seriously tested, so the estimate is uncertain." }, { "game": "game_08", "color": "white", "actual": 1310, "predicted": 1300, "error": -10, "absolute_error": 10, "low_80": 1050, "high_80": 1550, "interval_covers": true, "loo_baseline": 1884.0, "rapid_games": 4221, "rating_deviation": 45, "explanation": "Odd opening (1.Nc3, d3, e3) but playable. 15.Bxc5 drops a piece to a simple recapture tactic after 14...c5 (Ndxc5 defends). 21.Kg2?? walks into Ng5 ideas, and 23.Kg1?? rather than Kxf3 loses further; 24.h4 allows Nf3+ forking king and Re1. Repeated tactical oversights suggest a lower-club player." }, { "game": "game_08", "color": "black", "actual": 1791, "predicted": 1450, "error": -341, "absolute_error": 341, "low_80": 1200, "high_80": 1700, "interval_covers": false, "loo_baseline": 1884.0, "rapid_games": 4152, "rating_deviation": 49, "explanation": "Solid development with 3...Bb4, 8...Bxc3 and 9...e4 grabbing space. 14...c5 cleverly exploited the Bb4/Qd4 setup; 15...Ndxc5 won material. 17...Nxe4 and the queen trade were sensible. 22...Bxf3+ and 24...Nf3+ punished errors. Some passivity (12...Be6, 13...Nd7 allowing Qd4 pressure) but generally clean and tactically alert." }, { "game": "game_09", "color": "white", "actual": 2193, "predicted": 1700, "error": -493, "absolute_error": 493, "low_80": 1450, "high_80": 1950, "interval_covers": false, "loo_baseline": 1813.5, "rapid_games": 4525, "rating_deviation": 45, "explanation": "Reasonable Benoni/KID setup through move 12, but 12.Bd2 and 13.Qc1 allowed ...Ng4-e5-c4 winning the bishop pair. 20.Qd3 then 21.Qxe2 let Black win pawns with ...Bxc3 and ...Qxa3. Later play was active (25.Rxa7, 28.e5, 30.f5), but 34.Kh1 and 38.Rc8 ignored back-rank and mating threats. After 39.R8xc5 he was mated by doubled rooks on the second rank." }, { "game": "game_09", "color": "black", "actual": 2177, "predicted": 1800, "error": -377, "absolute_error": 377, "low_80": 1550, "high_80": 2050, "interval_covers": false, "loo_baseline": 1813.5, "rapid_games": 8936, "rating_deviation": 46, "explanation": "Played a solid Benoni with the ...b5 break on moves 10-11. The knight maneuver ...Ng4-e5-c4 (13-16) won the bishop pair. Good tactics with 21...Bxc3 and 22...Qxa3 won material. Some looseness followed: 27...Rd8 and 28...Bf5 allowed counterplay, and 31...hxg6 let 32.Rxe7 in. Converted cleanly with the queen trade (32...Qf4+), ...Bxd5, doubled rooks, and the mating 41...f3." }, { "game": "game_10", "color": "white", "actual": 2689, "predicted": 1650, "error": -1039, "absolute_error": 1039, "low_80": 1400, "high_80": 1900, "interval_covers": false, "loo_baseline": 1767.9444444444443, "rapid_games": 9005, "rating_deviation": 45, "explanation": "Solid Catalan/English setup through move 15, but 16.Nxe6 helped Black's structure. 24.Nxa7 greedily misplaced the knight and queen. After 24...Qc2 the reply 25.Rf1? allowed 25...Ba6 to win the exchange. White then resigned after 27...Ne8 drove the queen. That suggests weak tactical vigilance despite reasonable opening knowledge." }, { "game": "game_10", "color": "black", "actual": 2501, "predicted": 1750, "error": -751, "absolute_error": 751, "low_80": 1500, "high_80": 2000, "interval_covers": false, "loo_baseline": 1767.9444444444443, "rapid_games": 3141, "rating_deviation": 45, "explanation": "Sound opening setup with 7...c5 and a hanging-pawn structure. 13...Qd7 and 15...Ne6 were reasonable. 24...Qc2 was active and hit White's loose pieces. 25...Ba6 won the exchange, and 27...Ne8 harassed the queen. Some passivity around moves 17-20 (Bd6, Rf8, Be7 shuffling) keeps the estimate moderate." } ], "cost_usd": 0.134368, "reasoning_effort": "medium", "max_tokens": 8192, "run_start": "2026-10-11", "run_end": "2026-10-11", "display_name": "Claude Opus 5.5" }, { "model": "anthropic/claude-fable-5.1", "summary": { "games": 10, "player_estimates": 20, "actual_rating_range": [ 1112, 2718 ], "mae": 519.75, "rmse": 633.1752916846962, "signed_bias": -515.65, "pearson": 0.8213365112640489, "spearman": 0.8016840111947131, "within": { "100": { "count": 3, "fraction": 0.15 }, "200": { "count": 5, "fraction": 0.25 }, "300": { "count": 6, "fraction": 0.3 } }, "interval_80_coverage": 0.3, "interval_80_mean_width": 495.0, "game_mean_rating_mae": 515.65, "mae_game_bootstrap_95_interval": [ 332.94874999999996, 724.5099999999999 ], "baseline_1500_mae": 506.05, "baseline_leave_one_game_out_mean_mae": 471.5666666666666, "games_beating_loo_baseline": 4, "history_count_range": [ 1136, 31199 ], "max_profile_rating_deviation": 59 }, "predictions": [ { "game": "game_01", "color": "white", "actual": 1363, "predicted": 1300, "error": -63, "absolute_error": 63, "low_80": 1050, "high_80": 1550, "interval_covers": true, "loo_baseline": 1916.1666666666667, "rapid_games": 4696, "rating_deviation": 46, "explanation": "Found the Nd5/Nf6+ fork winning the queen (moves 15-17) and consolidated reasonably (19-24), but opening play (2.d3, 7.Bg5 then 9.Bd2) was aimless, conversion of a queen-up position was slow (26-37), 38.e6 allowed h3 counterplay, and 40.Qb8+?? hallucinated a mate/promotion and threw away a trivially won game. Typical low-intermediate pattern.\"\"" }, { "game": "game_01", "color": "black", "actual": 1159, "predicted": 1200, "error": 41, "absolute_error": 41, "low_80": 950, "high_80": 1450, "interval_covers": true, "loo_baseline": 1916.1666666666667, "rapid_games": 3260, "rating_deviation": 45, "explanation": "Unorthodox 2...d5/4...f5 and 7...Qd7 showed shaky opening understanding; 14...Nh7? overlooked the simple fork 15.Nd5 and lost the queen. Afterwards Black kept pieces active (Nd4, Bxg3xe1 trick at 30-31, h-pawn push) and showed decent resourcefulness, but the win came only from White's blunder at move 40." }, { "game": "game_02", "color": "white", "actual": 1575, "predicted": 1250, "error": -325, "absolute_error": 325, "low_80": 1000, "high_80": 1500, "interval_covers": false, "loo_baseline": 1834.888888888889, "rapid_games": 1928, "rating_deviation": 45, "explanation": "Standard Ruy Lopez setup but 8.Nd5 lost a pawn to 9...Nxe4 (unprotected e4 after 8.Nd5/9.c3). 18.h4?? dropped a second piece-level material, 20.Qh3? allowed a queen trade into a lost position, then 21...Re2 was ignored; 22.Bd3 and 24.Bc2 left the rook free to mate on h1. Repeated oversight of simple tactics suggests a low-intermediate rapid player. " }, { "game": "game_02", "color": "black", "actual": 2410, "predicted": 1450, "error": -960, "absolute_error": 960, "low_80": 1200, "high_80": 1700, "interval_covers": false, "loo_baseline": 1834.888888888889, "rapid_games": 2936, "rating_deviation": 45, "explanation": "Sound Ruy Lopez play, grabbed the free pawn with 9...Nxe4, exchanged accurately (11...Nf6, 12...Bxf6), won material via 18...Bxh4 and 19...Qxh4, simplified with 20...Qxh3, and executed a clean rook raid 21...Re2–Rxb2–Rb3–Rxh3–Rh1#. Competent but opponent offered no resistance; 13...Na5/14...Nc4 was reasonable. Mid-intermediate level." }, { "game": "game_03", "color": "white", "actual": 1782, "predicted": 1450, "error": -332, "absolute_error": 332, "low_80": 1200, "high_80": 1700, "interval_covers": false, "loo_baseline": 1851.6666666666667, "rapid_games": 18588, "rating_deviation": 45, "explanation": "White played a London setup with the odd 8.h4/9.Ng5 and 11.Nh3 retreat, but found decent knight manoeuvres (13.Nc4, 14.Ne5, 20.Ng6) and converted material cleanly after 23.Nxd5 and 24.Qb3. Technique from 33.Qf5 onwards (34.Ne7+, 37.Rd8, 46.Re8#) was tidy. Several imprecise moves suggest club level around 1400-1500.\n" }, { "game": "game_03", "color": "black", "actual": 1901, "predicted": 1200, "error": -701, "absolute_error": 701, "low_80": 950, "high_80": 1450, "interval_covers": false, "loo_baseline": 1851.6666666666667, "rapid_games": 1366, "rating_deviation": 50, "explanation": "Black's Dutch setup was reasonable, but 13...d5 and 14...Bd6 lost material to 15.Ng6/16.Bxd6; 16...Kf7 and 18...Re8 were aimless. 23...Nd7 left d5 and c5 hanging, 29...Rg8 and 33...Nc8 blundered further. Repeated tactical oversights indicate a low-intermediate player, roughly 1100-1300." }, { "game": "game_04", "color": "white", "actual": 2508, "predicted": 1450, "error": -1058, "absolute_error": 1058, "low_80": 1200, "high_80": 1700, "interval_covers": false, "loo_baseline": 1765.9444444444443, "rapid_games": 1324, "rating_deviation": 47, "explanation": "White knows the Caro-Kann Exchange setup (4.Bd3, 5.c3, 6.Ne2) but misplays it: 8.Qe1 and 10.f4 are awkward, 18.Nc5 trades a good piece, 21.c4 ignores Black's kingside play, 23.Qb2 allows Nd4-e2+ ideas, and 25.Kh1?? permits a smothered mate instead of 25.Kh1-alternatives like Kh1... actually Kh1 was forced-ish; 24.Bb4 was the real blunder. Typical intermediate club play.\n" }, { "game": "game_04", "color": "black", "actual": 2718, "predicted": 1600, "error": -1118, "absolute_error": 1118, "low_80": 1350, "high_80": 1850, "interval_covers": false, "loo_baseline": 1765.9444444444443, "rapid_games": 1136, "rating_deviation": 59, "explanation": "Black handled the Exchange Caro-Kann sensibly (Bg4, Qc7, long castle) and generated a coherent kingside pawn storm with 17...f5, 19...g5, 21...gxf4. The finish 22...Nd4, 23...Qxc5, 24...Ne2+ and 25...Qg1+ 26...Nf2# is a clean smothered mate pattern showing solid tactical awareness, though some moves (15...Rc8, 16...Ne4) were slightly loose." }, { "game": "game_05", "color": "white", "actual": 1120, "predicted": 1000, "error": -120, "absolute_error": 120, "low_80": 750, "high_80": 1250, "interval_covers": true, "loo_baseline": 1890.3333333333333, "rapid_games": 5115, "rating_deviation": 45, "explanation": "White's opening (2.Bc4, 5.Bd3 blocking the d-pawn, 7.d3 recapturing) shows weak understanding. 16.Be4?? drops material to the g2 fork; 19.Nfg5 then 20.Nf3 wastes tempi; 21.a3 ignores threats; 25.f3 and 27.Qxd3 collapse into mate. Repeated tactical oversights suggest a low-rated player.\n" }, { "game": "game_05", "color": "black", "actual": 1867, "predicted": 1250, "error": -617, "absolute_error": 617, "low_80": 1000, "high_80": 1500, "interval_covers": false, "loo_baseline": 1890.3333333333333, "rapid_games": 31199, "rating_deviation": 45, "explanation": "Black played a sensible Sicilian setup (4...b5, 6...Bb7) and punished 16.Be4 with 16...Nxg2+ and 17...fxe4, then found the strong 21...Nbd3 and 23...Bxh1. The finish 25...Nh3+ to 29...Qxf3# was accurate. Earlier 5...c4 and 11...f5 are a bit loose, and the opponent's weakness inflates the result, so moderate rating.\n" }, { "game": "game_06", "color": "white", "actual": 1112, "predicted": 900, "error": -212, "absolute_error": 212, "low_80": 700, "high_80": 1150, "interval_covers": true, "loo_baseline": 1917.7222222222222, "rapid_games": 3821, "rating_deviation": 45, "explanation": "White's opening is incoherent: 4.Bd3 then 6.Bb1 and 7.Qc2 waste tempi, 8.h4 is premature, 13.Ng5 drops material to 13...Nd4/14.Qd3 awkwardness, 18...Ncxb3 loses the rook and 20...Bxc4 a pawn. 24.Nf6+ and 25.Rf1 sacrifice everything for nothing, and 30.Qxb7 allows a two-move mate. Consistent with a low-rated rapid player.\n" }, { "game": "game_06", "color": "black", "actual": 1382, "predicted": 1050, "error": -332, "absolute_error": 332, "low_80": 850, "high_80": 1300, "interval_covers": false, "loo_baseline": 1917.7222222222222, "rapid_games": 5055, "rating_deviation": 45, "explanation": "Black's setup is passive and clumsy (3...Bd7, 9...Bf5 10...Bd7, 12...Bg4 then 15...Be6) but exploits White's errors reliably: 16...Bxg5, 18...Ncxb3 winning the exchange, 22...Ne6 defending, 26...Nxg5 and 27...Qxf6 collecting material, and the finish 30...Rc1+ 31...Qf2#. Tactical awareness exceeds White but opening play suggests a modest rating." }, { "game": "game_07", "color": "white", "actual": 1862, "predicted": 1250, "error": -612, "absolute_error": 612, "low_80": 1000, "high_80": 1500, "interval_covers": false, "loo_baseline": 1864.3333333333333, "rapid_games": 7581, "rating_deviation": 45, "explanation": "Passive Slav setup (3.e3, 5.Nbd2, 13.Ba3 trading into awkward Ra3) and a queenside push (17.b4, 19.b5) that ignored tactics. 21.Bf3 hung the rook to Bc2, then 22.Bh5?? lost the queen for nothing; resigning after 24.Bd3 down a full rook plus queen-for-rook. Typical of low-intermediate rapid play with weak tactical awareness." }, { "game": "game_07", "color": "black", "actual": 1593, "predicted": 1400, "error": -193, "absolute_error": 193, "low_80": 1150, "high_80": 1650, "interval_covers": true, "loo_baseline": 1864.3333333333333, "rapid_games": 13102, "rating_deviation": 45, "explanation": "Sound Slav development (2...c6, 3...Bf5, 9...Bd6, 10...O-O), reasonable simplification with 11...Ne4 and 13...Bxa3, and alert tactics: 21...Bc2 forked queen and rook, then 22...Bxd1 and 23...Bxb3 cashed in cleanly. Few hard tests were posed by White, so the ceiling is uncertain, but no errors visible." }, { "game": "game_08", "color": "white", "actual": 1310, "predicted": 1250, "error": -60, "absolute_error": 60, "low_80": 1000, "high_80": 1500, "interval_covers": true, "loo_baseline": 1884.0, "rapid_games": 4221, "rating_deviation": 45, "explanation": "Passive, unprincipled setup (1.Nc3, 2.d3, 3.e3, 5.g3) and 11.Bb4 misplaces the bishop. 14...c5 trapped the bishop, and 15.Bxc5 simply loses a piece for pawn. Later 17.Bxe4 and 18.Qxb6 conceded without resistance; 21.Kg2 and 22.Rd4 walked into a fork and 24.h4 allowed Nf3+ losing the rook. Weak tactical awareness.pulling rating down.pulling rating down." }, { "game": "game_08", "color": "black", "actual": 1791, "predicted": 1400, "error": -391, "absolute_error": 391, "low_80": 1150, "high_80": 1650, "interval_covers": false, "loo_baseline": 1884.0, "rapid_games": 4152, "rating_deviation": 49, "explanation": "Sensible development (3...Bb4, 7...Bg4, 8...Bxc3) and a good trap with 14...c5 winning a piece. Converted with simple trades (16...Qb6, 17...Nxe4) and finished with the fork 24...Nf3+. Solid but opponent offered little resistance; 11...Re8 and 20...Bg4 are competent club-level moves." }, { "game": "game_09", "color": "white", "actual": 2193, "predicted": 1500, "error": -693, "absolute_error": 693, "low_80": 1250, "high_80": 1750, "interval_covers": false, "loo_baseline": 1813.5, "rapid_games": 4525, "rating_deviation": 45, "explanation": "Reasonable Benoni setup with g3/Bg2, but passive and inaccurate middlegame: 12.Bd2/13.Qc1 let Black's knight tour win material (Nc4xd2) and 22.bxc3 dropped a3 and c3 pawns. 25.Rxa7 and 28.e5 showed activity but 29.g4/30.f5 loosened the king, allowing the mating net after 32...Qf4+. Typical club-level planning errors.\"},\"black\":{\"rating\":1650,\"low_80\":1400,\"high_80\":1900,\"explanation\":\"Knew Benoni ideas: Na6-c7, b5 break, Rb8, the knight manoeuvre Ng4-e5-c4 winning the bishop pair, and Nd4/Bxc3 to collect pawns 22-23. Conversion was purposeful (28...Bf5, 32...Qf4+ simplifying, 34...Ba2, 38...Rdd2 doubling on the 2nd rank for mate). Slight inaccuracies like 26...Qd2 allowing Rxe7 counterplay, but solid intermediate-level play." }, { "game": "game_09", "color": "black", "actual": 2177, "predicted": 1650, "error": -527, "absolute_error": 527, "low_80": 1400, "high_80": 1900, "interval_covers": false, "loo_baseline": 1813.5, "rapid_games": 8936, "rating_deviation": 46, "explanation": "Knew Benoni ideas: Na6-c7, b5 break, Rb8, the knight manoeuvre Ng4-e5-c4 winning the bishop pair, and Nd4/Bxc3 to collect pawns 22-23. Conversion was purposeful (28...Bf5, 32...Qf4+ simplifying, 34...Ba2, 38...Rdd2 doubling on the 2nd rank for mate). Slight inaccuracies like 26...Qd2 allowing Rxe7 counterplay, but solid intermediate-level play." }, { "game": "game_10", "color": "white", "actual": 2689, "predicted": 1500, "error": -1189, "absolute_error": 1189, "low_80": 1250, "high_80": 1750, "interval_covers": false, "loo_baseline": 1767.9444444444443, "rapid_games": 9005, "rating_deviation": 45, "explanation": "Reasonable English/Réti setup through move 12, but aimless queen maneuvers (14.Qd2, 15.Qf4, 18.Qh4, 21.Qf4) and 20.Nb5 show weak planning. 23.Qc7 and 24.Nxa7 grabbed a pawn while ignoring 24...Qc2, then 25.Rf1 and 26.Be5 collapsed, dropping the exchange and a piece. Resignation after 27...Ne8 was accurate but the tactical oversight is typical of club level.​" }, { "game": "game_10", "color": "black", "actual": 2501, "predicted": 1650, "error": -851, "absolute_error": 851, "low_80": 1400, "high_80": 1900, "interval_covers": false, "loo_baseline": 1767.9444444444443, "rapid_games": 3141, "rating_deviation": 45, "explanation": "Solid Tarrasch-style development (7...c5, 9...b6, 11...Rc8) and sensible regrouping (17...Bd6, 19...Qf7). 21...Qg6 slightly passive, 23...Bc5 allowed Nxa7, but 24...Qc2 was a strong infiltration and 25...Ba6 and 27...Ne8 cleanly won material and trapped the queen's play. Good tactical alertness, modest strategic depth." } ], "cost_usd": 0.39707, "reasoning_effort": "medium", "max_tokens": 8192, "run_start": "2026-10-11", "run_end": "2026-10-11", "display_name": "Claude Fable 5.1" }, { "model": "qwen/qwen3.8-max-prime", "summary": { "games": 10, "player_estimates": 20, "actual_rating_range": [ 1112, 2718 ], "mae": 523.05, "rmse": 682.5473976215865, "signed_bias": -440.65, "pearson": 0.12188521660072832, "spearman": 0.13157111386384723, "within": { "100": { "count": 2, "fraction": 0.1 }, "200": { "count": 4, "fraction": 0.2 }, "300": { "count": 10, "fraction": 0.5 } }, "interval_80_coverage": 0.15, "interval_80_mean_width": 300.0, "game_mean_rating_mae": 482.55, "mae_game_bootstrap_95_interval": [ 292.34625, 790.2025 ], "baseline_1500_mae": 506.05, "baseline_leave_one_game_out_mean_mae": 471.5666666666666, "games_beating_loo_baseline": 4, "history_count_range": [ 1136, 31199 ], "max_profile_rating_deviation": 59 }, "predictions": [ { "game": "game_01", "color": "white", "actual": 1363, "predicted": 1150, "error": -213, "absolute_error": 213, "low_80": 1000, "high_80": 1300, "interval_covers": false, "loo_baseline": 1916.1666666666667, "rapid_games": 4696, "rating_deviation": 46, "explanation": "Opening is playable but passive (2.d3). Tactical misstep at 15.Nd5? allows 15...Nxg5 16.Nf6+ Ke7, losing material for insufficient compensation (16...Kxd7). Positional drift continues with 12.Nh2 and 13.g3, yielding Black easy targets. Later, 26.Qc6 aims at counterplay but the 30.dxe4 sequence collapses the structure. The perpetual attempt (33–37) is the only resourceful idea; overall execution sits near 1100–1200." }, { "game": "game_01", "color": "black", "actual": 1159, "predicted": 1300, "error": 141, "absolute_error": 141, "low_80": 1150, "high_80": 1450, "interval_covers": true, "loo_baseline": 1916.1666666666667, "rapid_games": 3260, "rating_deviation": 45, "explanation": "Solid central opening (2...d5, 3...dxe4, 6...Bd6) and active kingside expansion (9...g5, 11...g4). Wins material cleanly after 15...Nxg5 16...Ke7, converting via simplification (19...Nd4, 21...Nxf3). Endgame technique is competent: 27...e4 breaks, 30...Bxg3 removes the key pawn, and the h-pawn advance (38...h3) forces a winning rook trade. Slight wobble with 34...Kb7 allowing perpetual chances, but 36...b5 closes it. Mid-1200 to low-1400 range." }, { "game": "game_02", "color": "white", "actual": 1575, "predicted": 1200, "error": -375, "absolute_error": 375, "low_80": 1100, "high_80": 1350, "interval_covers": false, "loo_baseline": 1834.888888888889, "rapid_games": 1928, "rating_deviation": 45, "explanation": "Opening was sound until 9.Nxe4? (a knight fork was missed: 9...Nxe4). Later 16.Bh6?? lost the h4-pawn, and 18.Qh3 traded queens, exposing g2. The final collapse (24...Rh1#) stemmed from 22.Rb3 and 24.Bg5??, ignoring the rook lift to g1." }, { "game": "game_02", "color": "black", "actual": 2410, "predicted": 1350, "error": -1060, "absolute_error": 1060, "low_80": 1250, "high_80": 1500, "interval_covers": false, "loo_baseline": 1834.888888888889, "rapid_games": 2936, "rating_deviation": 45, "explanation": "Standard Ruy Lopez: 8...d6, 9...Nxe4, 12...Bf6. Correctly identified 14...Nc4, then 17...Bb7. Won a pawn via 17...Bxh4 and doubled rooks on the 7th (21...Re2, 22...Rb3, 23...Rxh3). Finished cleanly with 24...Rh1#." }, { "game": "game_03", "color": "white", "actual": 1782, "predicted": 1500, "error": -282, "absolute_error": 282, "low_80": 1400, "high_80": 1600, "interval_covers": false, "loo_baseline": 1851.6666666666667, "rapid_games": 18588, "rating_deviation": 45, "explanation": "Solid London-system moves (1–7) but the h4/Ng5 sortie (8–10) is unfocused. Good knight repositioning (15–17, 23) and central pressure, yet the unnecessary king walk allowed Black back (16–17). The conversion after 23.Nxd5 is competent but not sparkling (25–38), with a slightly clumsy 39.Qd4/43.Bxg6+ sequence." }, { "game": "game_03", "color": "black", "actual": 1901, "predicted": 1200, "error": -701, "absolute_error": 701, "low_80": 1100, "high_80": 1300, "interval_covers": false, "loo_baseline": 1851.6666666666667, "rapid_games": 1366, "rating_deviation": 50, "explanation": "Standard Dutch setup (1–6) but 8…d6 and 9…Qc8 are passive. After 12.dxc5 Black fails to recapture dynamically and drifts with 14…Bd6 and 15…Rd8. 19…gxf6 weakens the structure badly. Poor defense afterward (27–30, 33) and the rook exchange (35…Qxg8) collapses immediately into mate (46.Re8#). No resistance after move 33." }, { "game": "game_04", "color": "white", "actual": 2508, "predicted": 1300, "error": -1208, "absolute_error": 1208, "low_80": 1150, "high_80": 1450, "interval_covers": false, "loo_baseline": 1765.9444444444443, "rapid_games": 1324, "rating_deviation": 47, "explanation": "Passive opening (8.Qe1, 9.Bg5), slow queenside play (14.b4, 15.Nb3) with no concrete threat. Allowed Black to dominate the center with Ne4/f5. Fatal errors: 21.c4 opens lines, 22.b5 abandons defense, 23.Qb2 ignores Ne2+ and Qg1# patterns. Missed all counterplay chances." }, { "game": "game_04", "color": "black", "actual": 2718, "predicted": 1550, "error": -1168, "absolute_error": 1168, "low_80": 1400, "high_80": 1700, "interval_covers": false, "loo_baseline": 1765.9444444444443, "rapid_games": 1136, "rating_deviation": 59, "explanation": "Solid Caro-Kann development, castled long, coordinated pieces. 16…Ne4, 17…f5 seized space. 21…gxf4, 22…Nd4 exploited white's weaknesses. Calculated the final tactic precisely: 23…Qxc5, 24…Ne2+, 25…Qg1+, 26…Nf2#. Clean conversion with no inaccuracies in the attack." }, { "game": "game_05", "color": "white", "actual": 1120, "predicted": 1350, "error": 230, "absolute_error": 230, "low_80": 1200, "high_80": 1500, "interval_covers": false, "loo_baseline": 1890.3333333333333, "rapid_games": 5115, "rating_deviation": 45, "explanation": "The 2.Bc4 anti-Sicilian with a slow c3/d3 plan is common at this level. Moves 5–8 shuffle the bishop awkwardly, but real damage comes from 11.e5?? (losing the d5 square and eventually the e4 pawn), then 16.Kf1 forfeiting castling, and finally 23.Nd4?? overlooking Bxh1. Opening ideas are present; tactical vision and king safety judgment are not." }, { "game": "game_05", "color": "black", "actual": 1867, "predicted": 1650, "error": -217, "absolute_error": 217, "low_80": 1500, "high_80": 1800, "interval_covers": false, "loo_baseline": 1890.3333333333333, "rapid_games": 31199, "rating_deviation": 45, "explanation": "Black plays a competent Taimanov/Kan setup (…a6, …b5, …Bb7) and converts smoothly. 15…Nf4 followed by 16…Nxg2+ is accurate sacrificial calculation, and 21…Nbd3 dominates. 25…Nh3+ and 26…Bxf3 show clean technique to finish with 29…Qxf3#. The play is purposeful and largely error-free, though it mostly punished White's self-inflicted wounds rather than creating deep complications independently." }, { "game": "game_06", "color": "white", "actual": 1112, "predicted": 1400, "error": 288, "absolute_error": 288, "low_80": 1300, "high_80": 1550, "interval_covers": false, "loo_baseline": 1917.7222222222222, "rapid_games": 3821, "rating_deviation": 45, "explanation": "The Bb1-Qc2 battery is known but a3 and h4 are dubious; 8.h4 invites Black's comfortable structure. Move 10 e4 wastes time after 3.e3. Losing the a1-rook via Nxa1 on move 19 and the queen-ending collapse show tactical gaps above roughly 1500, while avoiding immediate disaster earlier keeps the rating near low-intermediate." }, { "game": "game_06", "color": "black", "actual": 1382, "predicted": 1500, "error": 118, "absolute_error": 118, "low_80": 1400, "high_80": 1600, "interval_covers": false, "loo_baseline": 1917.7222222222222, "rapid_games": 5055, "rating_deviation": 45, "explanation": "Sound opening (moves 1-8) and good timing with Nd4/Be6. Winning the rook on 18-19 and grabbing material on 24-27 show alert calculation. Missed 27...Nf3+ but still won cleanly. Allowing 28.Qxc7+ gave some counterplay, yet 29...Rac8 and 30...Rc1# finish efficiently. Practical 10+0 play at roughly 1500 level." }, { "game": "game_07", "color": "white", "actual": 1862, "predicted": 1400, "error": -462, "absolute_error": 462, "low_80": 1200, "high_80": 1600, "interval_covers": false, "loo_baseline": 1864.3333333333333, "rapid_games": 7581, "rating_deviation": 45, "explanation": "Moves 1–10 (cxd5, h3, b3, a4) show solid but slow Slav development, yet White never found active play. The catastrophic 21.Bf3?? allows Black to grab the rook and bishop in two moves (22…Bxd1, 23…Bxb3). This isn't a mere oversight but a complete failure to see an obvious pin/tactic, capping the estimate well below club level." }, { "game": "game_07", "color": "black", "actual": 1593, "predicted": 1600, "error": 7, "absolute_error": 7, "low_80": 1450, "high_80": 1800, "interval_covers": true, "loo_baseline": 1864.3333333333333, "rapid_games": 13102, "rating_deviation": 45, "explanation": "The Slav setup (2…c6, 3…Bf5) and 11…Ne4 trade sequence are solid and thematic. 20…Qg6 is a reasonable repositioning. However, the game-winning 21…Bc2!! is a strong double-attack, yet the line is only two moves deep (22…Bxd1, 23…Bxb3), suggesting the combination was more a natural tactic than deep calculation. White resigned after 23 moves, so no endgame was reached." }, { "game": "game_08", "color": "white", "actual": 1310, "predicted": 1350, "error": 40, "absolute_error": 40, "low_80": 1200, "high_80": 1500, "interval_covers": true, "loo_baseline": 1884.0, "rapid_games": 4221, "rating_deviation": 45, "explanation": "White's setup (Nc3/d3/e3/g3) is passive but structurally coherent. Missed the Bb4 pin after 8…Bxc3, then 14…c5 15.Bxc5?? dropped a piece. Later 21.Kg2?? walked into the decisive knight fork. Opening and general maneuvering look near-club level, but the two outright tactical oversights cap the estimate well below 1500." }, { "game": "game_08", "color": "black", "actual": 1791, "predicted": 1550, "error": -241, "absolute_error": 241, "low_80": 1400, "high_80": 1700, "interval_covers": false, "loo_baseline": 1884.0, "rapid_games": 4152, "rating_deviation": 49, "explanation": "Black played sensible rapid development (…d5, …Bb4, …O-O) and calmly kept the pin after 15.Bxc5 Nxc5, trusting the fork motif. The finishing 21…Ng5 23…Be4 24…Nf3+ was precise. However, 8…Bxc3 concedes the bishop pair for minimal gain, and 12…Be6/13…Nd7 are workmanlike rather than ambitious, suggesting solid mid-club strength rather than expert." }, { "game": "game_09", "color": "white", "actual": 2193, "predicted": 1450, "error": -743, "absolute_error": 743, "low_80": 1250, "high_80": 1650, "interval_covers": false, "loo_baseline": 1813.5, "rapid_games": 4525, "rating_deviation": 45, "explanation": "Reached a standard KID setup (g3, Bg2) and managed the Benoni transition to move 13 without structural error. Then drifted: 15.Kh2, 21.bxc3, and 28.e5 all cede material or key squares without a clear plan. After 22…Qxa3 White never generated counterplay and lost coordination in the rook endgame. Solid opening, weak middlegame direction and defence—consistent with a mid-intermediate rapid player." }, { "game": "game_09", "color": "black", "actual": 2177, "predicted": 1800, "error": -377, "absolute_error": 377, "low_80": 1600, "high_80": 2000, "interval_covers": false, "loo_baseline": 1813.5, "rapid_games": 8936, "rating_deviation": 46, "explanation": "Handled the Benko-style gambit accurately (8…Bd7, 10…b5) and exploited the c-file and b-file after 13…Ng4 and 19…Nd4. Converted the pawn-up endgame cleanly: 29…Bb1, 33…exf4, 38…Rdd2, and the mating net 40…Rbg2+ 41.f3 showed precise calculation under time pressure. No visible inaccuracies; opening knowledge, piece activity, and technique point to a strong club-level player." }, { "game": "game_10", "color": "white", "actual": 2689, "predicted": 1200, "error": -1489, "absolute_error": 1489, "low_80": 1050, "high_80": 1400, "interval_covers": false, "loo_baseline": 1767.9444444444443, "rapid_games": 9005, "rating_deviation": 45, "explanation": "The opening was solid but the plan with Qf4–Qh4 showed little purpose (moves 15–19). Exchanging on e6 opened the f-file for Black (16.Nxe6). 24.Nxa7 was a fatal blunder, losing the knight after Qc2 and Ba6, and 26.Be5 failed to stop the rook loss. Competent opening but poor middlegame judgement and a decisive tactical oversight." }, { "game": "game_10", "color": "black", "actual": 2501, "predicted": 1400, "error": -1101, "absolute_error": 1101, "low_80": 1250, "high_80": 1600, "interval_covers": false, "loo_baseline": 1767.9444444444443, "rapid_games": 3141, "rating_deviation": 45, "explanation": "Black developed soundly and achieved comfortable equality after the opening. After 16.Nxe6, 16…fxe6 gave Black a strong semi-open f-file. 18…Rf8 and 19…Qf7 showed good piece coordination. 24…Qc2 and 25…Ba6 was a clean tactical sequence exploiting the knight on a7, forcing decisive material gain. Solid, patient play with one clear, well-executed tactical strike." } ], "cost_usd": 0.39306, "reasoning_effort": "medium", "max_tokens": 8192, "run_start": "2026-10-11", "run_end": "2026-10-11", "display_name": "Qwen3.8 Max Prime" } ], "partial_models": [ { "model": "deepseek/deepseek-v4.1-flash", "completed_games": 0, "attempted_games": 1, "errors": [ "429 Client Error: Too Many Requests for url: https://openrouter.ai/api/v1/chat/completions" ], "display_name": "DeepSeek V4.1 Flash" }, { "model": "x-ai/grok-4.7", "completed_games": 0, "attempted_games": 1, "errors": [ "Billing exceeded reservation; investigate before further calls" ], "display_name": "Grok 4.7" }, { "model": "openai/gpt-5.6-sol-pro", "completed_games": 2, "attempted_games": 3, "errors": [ "Provider exceeded the requested output token cap" ], "display_name": "GPT-5.6 Pro Sol" }, { "model": "openai/gpt-5.6-luna-pro", "completed_games": 0, "attempted_games": 1, "errors": [ "Provider exceeded the requested output token cap" ], "display_name": "GPT-5.6 Pro Luna" } ], "cost_usd": 3.3258107479999994, "reported_api_cost_usd": 3.313394748, "unresolved_reservations_usd": 0.012416, "classification_cost_usd": 0.001749, "project_accounted_usd": 3.327559747999999, "cost_fully_reported": false, "project_budget_remaining_usd": 16.672440252, "sample_sha256": "d0470a6203766c1e59b0de80dccda132a2758cf397664c5fef5afe9699c7df50", "prompt_sha256": "1be15031720751e5cfa751722c517c4b30bca7e5bcd207bafc1557ef14c3d484", "methodology": { "id": "lichess-rating", "title": "Guess the chess rating", "summary": "Estimate each player's established Lichess rapid rating from an anonymized game.", "about": [ "Ten rated 10+0 games, twenty distinct players. Every player had at least 1,000 rapid games, a non-provisional rating and rating deviation at most 60 at collection. Games were at most seven days old. Targets are the pre-game Lichess rapid Glicko-2 ratings, often informally called Elo.", "The same frozen prompt tells every model exactly which rating to predict: Lichess rapid, not FIDE or Chess.com, separately for White and Black. It sees legal mainline moves, result and time control. Names, ratings, dates, game links, engine analysis and clock annotations are hidden. Models have no tools or cross-game context.", "One call per game, medium reasoning requested and an 8,192-token output cap including reasoning. Provider meanings of medium can differ. Point estimates are rounded to 50; mean absolute error uses all twenty estimates. The original Opus pilot is reused unchanged." ], "metric": "MAE (rating points)", "precision": 2, "lower_is_better": true, "items": 20, "run_start": "2026-10-11", "run_end": "2026-10-11", "cost_usd": 3.3258107479999994, "rows": [ { "model": "Gemini 3 Flash (Preview)", "score": 255.85, "n": 20, "reasoning_effort": "medium", "setting": "medium reasoning · 10 games" }, { "model": "Gemini 3.1 Pro (Preview)", "score": 263.95, "n": 20, "reasoning_effort": "medium", "setting": "medium reasoning · 10 games" }, { "model": "Gemini 3.8 Flash", "score": 280.85, "n": 20, "reasoning_effort": "medium", "setting": "medium reasoning · 10 games" }, { "model": "GPT-6.1 Sol", "score": 298.35, "n": 20, "reasoning_effort": "medium", "setting": "medium reasoning · 10 games" }, { "model": "Gemini 3.7 Flash", "score": 300.55, "n": 20, "reasoning_effort": "medium", "setting": "medium reasoning · 10 games" }, { "model": "GPT-6 Astra", "score": 302.35, "n": 20, "reasoning_effort": "medium", "setting": "medium reasoning · 10 games" }, { "model": "GPT-5.6 Sol", "score": 338.55, "n": 20, "reasoning_effort": "medium", "setting": "medium reasoning · 10 games" }, { "model": "Gemini 3.5 Flash", "score": 346.75, "n": 20, "reasoning_effort": "medium", "setting": "medium reasoning · 10 games" }, { "model": "GPT-6 Sol", "score": 352.05, "n": 20, "reasoning_effort": "medium", "setting": "medium reasoning · 10 games" }, { "model": "Gemini 3.6 Flash", "score": 357.05, "n": 20, "reasoning_effort": "medium", "setting": "medium reasoning · 10 games" }, { "model": "Claude Sonnet 5.5", "score": 433.55, "n": 20, "reasoning_effort": "medium", "setting": "medium reasoning · 10 games" }, { "model": "GPT-5.6 Luna", "score": 433.55, "n": 20, "reasoning_effort": "medium", "setting": "medium reasoning · 10 games" }, { "model": "GLM-5.3 Flash", "score": 456.95, "n": 20, "reasoning_effort": "medium", "setting": "medium reasoning · 10 games" }, { "model": "Claude Opus 5.5", "score": 463.35, "n": 20, "reasoning_effort": "medium", "setting": "medium reasoning · 10 games" }, { "model": "Claude Fable 5.1", "score": 519.75, "n": 20, "reasoning_effort": "medium", "setting": "medium reasoning · 10 games" }, { "model": "Qwen3.8 Max Prime", "score": 523.05, "n": 20, "reasoning_effort": "medium", "setting": "medium reasoning · 10 games" }, { "model": "Other nine games' mean rating", "score": 471.57, "reference": true }, { "model": "Always predict 1500", "score": 506.05, "reference": true } ], "notes": [ "Pilot: 10 independent games, 20 player estimates. Lower mean absolute error is better. Only models completing all ten games are ranked.", "Two games per mean-rating band; this broad convenience sample is not representative of Lichess players. Both estimates from a game are related, and close scores do not establish a reliable ordering.", "Reference predictors are unranked. The held-out mean uses the other nine games' ratings and sees neither player's rating from the held-out game.", "Recorded evaluation API cost: $3.3134. The displayed total also retains $0.0124 for requests without reported billing.", "Not ranked: DeepSeek V4.1 Flash completed 0 of 10 games; see the evidence for the recorded request failure.", "Not ranked: Grok 4.7 completed 0 of 10 games; see the evidence for the recorded request failure.", "Not ranked: GPT-5.6 Pro Sol completed 2 of 10 games; see the evidence for the recorded request failure.", "Not ranked: GPT-5.6 Pro Luna completed 0 of 10 games; see the evidence for the recorded request failure." ], "link": { "href": "/benchmarks/lichess-rating-pilot.html", "label": "Games, exact prompt and every prediction" } } }