File "SM06log.txt", by Kenneth W. Regan, 10/6/06. Game source---Irina Krush's US Chess Online blog entry at http://beta.uschess.org/frontend/news_7_129.php [White: E. Sukhovsky] [Black: V. Mikhalevsky] [Event: Ashdod International] [Date: 9/22/06] [Result: 1-0] 1.e4 e5 2.Nf3 Nc6 3.Bb5 a6 4.Ba4 Nf6 5.O-O Nxe4 6.d4 b5 7.Bb3 d5 8.dxe5 Be6 9.Nbd2 Nc5 10.c3 d4 11.Bxe6 Nxe6 12.cxd4 Ncxd4 13.a4 Bb4 Only 12 games in BD2006 from here, so start Fritzing. Evals taken between 13/13 and 14/14 14 Nxd4 0.88 14 axb5 0.84 14 Nb3 0.26 14 Ne4 0.09 14 Kh1 -0.03 NON-MATCH, delta = -0.04 14.axb5 Nxb5 15 Qa4 0.74 15 Ne4 0.69 15 Qb3 0.68 15 Qe2 0.66 15 Qc2 0.66 NON-MATCH, delta = -0.05 15.Qb3 Bxd2 16 Bxd2 0.73 16 Rd1 0.69 16 Nxd2 0.29 16 Rxa6 0.15 16 Qc2 -1.71 at 14/14. 0.71 Bxd2, same Rd1, Nxd2 0.25, Rxa6 still 0.15 NON-MATCH, delta = -0.56 (still prep?) 16.Rxa6 Rb8 17 Rxe6+ 0.13 17 Nxd2 -0.25 17 Bxd2 -0.55 17 Rd1 =0.70 17 Qc4 -2.46 MATCH, delta2 = 0.38 17.Rxe6+ fxe6 18 Qxe6+ 0.18 (alreeady after 14/14) (obvious move to me) 18 Bxd2 -0.49 18 Rd1 -0.57 18 Nxd2 -1.14 18 Qc2 -2.87 MATCH, delta2 = 0.67 18.Qxe6+ Qe7 19 Qc6+ 0.06 19 Qxe7+ 0.00 19 Qb3 -1.96 Qc6 up to +0.16, Qxe7+ to -0.06 after 14/14. MATCH, delta = 0.06 19.Qc6+ Qd7 20 Qxd7+ 0.15 20 Qc5 -2.15 MATCH, forced move, delta = 2.30 20.Qxd7+ Kxd7 21 Bxd2 0.14 21 Rd1 -0.20 21 Nxd2 -0.21 21 e6+ -0.76 21 h4 -3.23 MATCH, delta = 0.36 21.Bxd2 Ke6 22 h3 0.22 22 Rc1 0.17 22 Be3 0.17 22 h4 0.05 22 Re1 0.01 22 Ra1 -0.10 22 Bb4 -0.12 22 Bb4 -0.12 22 b4 -0.12 22 Ng5+ -0.13 Wide choice! NON-MATCH, delta2 = -0.05 22 h3 still first by 0.01 after 14 ply. 22.Rc1 Rhd8 23 Be3 0.22 23 Ba5 -0.07 23 h3 -0.07 23 Ng5+ -0.10 23 Rc4 -0.10 23 g3 -0.12 23 Kf1 -0.12 23 Rc2 -0.16 23 Bg5 -0.16 23 b4 -0.17 Wide choice again MATCH, delta = 0.29 23.Be3 Rd3 24 Kf1 0.20 24 h4 0.18 24 h3 0.13 24 g4 0.11 24 g3 0.01 24 Rc6+ 0.00 24 Kh1 0.00 (Wide choice, though this move is ridiculous) 24 Bc5 -0.10 24 Re1 -0.29 24 Ra1 -0.48 NON-MATCH, delta2 = -0.09 24.g4 Rb3 25 Rc6+ 0.17 25 Rc5 0.17 25 Rc2 0.16 25 Nd2 0.00 25 Rd1 0.00 25 Ng5+ -0.21 25 Kg2 -0.25 25 h3 -0.27 25 h4 -0.29 25 Re1 -0.30 TIE, delta = 0.0 25.Rc6+ Kf7 26 Ng5+ 0.41 26 Rc2 0.33 26 Kg2 0.30 26 h3 0.27 26 Kf1 0.16 26 Kh1 0.09 26 h4 0.07 26 e6+ 0.06 26 Ra6 0.06 26 Rc5 0.00 Ng5+ still ahead 0.39 to 0.34 after 14/14. NON-MATCH, delta2 = -0.05 26.Rc2 Rd8 27 Nd2 0.49 27 Ng5+ 0.42 27 Kg2 0.42 27 h4 0.37 27 Kf1 0.24 27 h3 0.12 27 Kh1 0.12 27 Re2 0.01 27 Rc6 0.00 27 g5 -0.07 NON-MATCH, delta2 = -0.07 Wide choices each time! 27.Kg2 Rd5 28 h4 0.40 28 h3 0.40 28 Kg3 0.21 28 Bc5 0.18 28 Ba7 0.18 28 Kf1 0.16 28 Re2 0.13 28 Kh3 0.08 28 Bc1 0.04 28 Ng5+ 0.04 After 14 ply, h4 0.40, h3 0.38, which would match, but by rule TIE, delta = 0.00 28.h4 h6 29 h5 0.54 29 Bc5 0.47 29 Kg3 0.31 29 Kh2 0.29 29 Rc6 0.15 29 g5 0.04 29 Nd2 0.00 29 Ba7 0.00 29 Kh3 -0.01 29 Bd2 -0.07 MATCH, delta = 0.07 (h5 led by 0.01 at 14/43). Note--comments referred to "the strategy to fixing Black's pawns" 29.h5 Kg8 (?!) 30 Bc5 0.99 30 e6 0.86 30 Kg3 0.77 30 Rc6 0.71 30 Kh2 0.50 30 g5 0.49 30 Kh3 0.41 30 Bc1 0.38 30 Nd2 0.30 30 Kf1 0.30 MATCH, delta = 0.14 30.Bc5 Kf7 31 g5 +0.99 31 Rc4 0.25 31 Ba3 0.08 31 Ba7 0.00 31 Be3 0.00 31 Nd2 0.00 31 Rc1 -0.03 (these are 14-ply scores now, Fritz is too fast!) 31 Ng1 -0.32 31 Nh4 -0.34 31 Kg3 -0.36 WOW---31 Be3 gave away all of White's advantage! But this means eralier evals were contingent on a single (unexpected!?) vigorous followup. Hence the players probably had the illusion that the game and individual move differences were closer than Fritz reports! NON-MATCH, emphatic!, delta2 = -0.99? WAIT---White offered a repetition. Time pressure? Non-repeat eval after 31 Be3 is 0.48, so maybe delta2 = -0.51 31.Be3 Rb4 32 Kg3 0.60 32 g5 0.57 32 Kh3 0.30 32 Nh2 -0.45 32 e6+ -0.57 32 Nh4 -0.69 32 Kh2 -0.84 4 next similar Now after 14/14, g5 takes over 0.60 to Kg3 0.47, but by rule, MATCH, delta = 0.03 32.Kg3 Rb3 33 g5 0.87 33 Nd2 0.66 33 Rc6 0.61 22 Kh2 0.31 33 Kh4 0.24 33 Rc4 0.21 33 Rc1 0.20 33 Kh3 0.00 33 Kf4 0.00 33 Kg2 0.00 NON-MATCH, delta2 = -0.26 (after 14/14 similar) 33.Rc6 Rxb2 34 Rg6 +0.38 (g5 led before 13/13) 24 g5 +0.33 34 Ra6 0.32 34 Kh3 0.12 34 Kg2 0.07 34 Kh2 0.07 34 Rc4 0.04 34 e6+ 0.02 34 Kh4 0.00 34 Kf4 0.00 g5 leads after 14/14, 0.62 to 0.38 for Rg6, but by rule, MATCH, delta = 0.05 34.Rg6 Kf8 35 e6 0.32 35 Nh4 0.31 35 Rc6 0.00 35 Ra6 0.00 35 Kh3 -0.04 35 g5 -0.07 35 Kg2 -0.08 35 Kh4 -0.10 35 Kf4 -0.11 35 Re6 -0.11 MATCH, delta = 0.01 only. But after 14 ply, match by 0.15. 35.e6 Nc3 36 Kh4 0.15 36 Kg2 0.12 36 e7+ 0.12 (if Black doesn't take!) 36 Kh3 0.01 36 Kh2 0.00 36 Nh4 -0.07 14/14 now 36 g5 -0.17 36 Kf4 -0.27 36 Ne1 -0.66 36 Ng1 -0.91 after 14/14, e7+ 0.11 to Kh4 0.09. But by rule, NON-MATCH, delta2 = -0.03 36.e7+ Kxe7 (?) 37 Rxg7+ 0.89 (obvious) 37 Kg2 0.01 37 Kh3 0.00 37 Kh4 -0.05 37 g5 -0.37 37 Nh4 -0.40 37 Kh2 -0.82 (etc., with high delta these don't matter) MATCH, delta = 0.88 37.Rxg7+ Kd8 38 g5 1.15 (was 0.00 at 11 ply) 38 Rh7 0.83 38 Rg6 0.82 38 Rg8+ 0.70 38 Nh4 0.48 38 Kg2 0.31 38 Kh4 0.28 38 Rxh6 0.20 38 Kh3 0.11 38 Kh2 0.01 MATCH, delta = 0.32 38.g5 Ne4+ 39 Kg2 1.13 39 Kg4 0.00 39 Kh3 -0.05 39 Kf4 -0.11 39 Kh2 -0.14 39 Kh4 -0.21 MATCH, delta = 1.13 (obvious move) 39.Kg2 hxg5 (? time pressure?) 40 h6 2.87 40 Rg8+ 1.28 40 Nxg5 0.28 40 Kg1 0.00 40 Ne1 0.0---all these don't matter. MATCH, delta = 1.59 (obvious even at Move 40?) 40.h6 g4 41 Ng5 3.81 41 Rxg4 2.46 41 Rg8+ 1.53 41 Nd2 0.65 etc. MATCH, delta = 1.35. Here's a flaw---this "delta" is high enough to say the move is "obvious", but is it really? Rxg4 wins handily too, so is there much between them? Maybe deltas should scale once the eval itself is high. 41.Ng5 Nf6 42 Ne6+ 4.59 42 Rf7 3.81 42 Bd4 1.88 42 h7 1.75 42 Kg3 1.71 42 Bf4 1.36 42 Rg8+ 1.20 42 Rg6 1.20 (do these matter?) 42 Bc1 1.02 42 Kh2 0.95 MATCH, delta = 0.78 42.Ne6+ Kc8 43 Rc7+ 4.54 43 h7 2.60 43 Bd4 2.09 MATCH, delta = 1.94 43.Rxc7+ Kb8 44 Rf7 4.76 44 Bf4 3.60 44 h7 2.63 44 Rg7 2.62 44 Ba7+ 2.53 MATCH, delta = 1.16 44.Rf7 Ne4 45 Bf4+ 6.54 45 Ba7+ 6.53 45 Re7 3.33 45 Bd4 3.31 45 Rf4 2.96 etc. HA! By rule, this is a MATCH, delta = 0.01! Is this kind of thing significant? (Well, Bf4+ is the more obvious kill.) 45.Bf4+ Rd6 46 h7 #7! 46 Bxd6+ 10.30 Meaningless. 46.Bxd6+ Nxd6 Trivial match, but ignore. 47.Rf8+ 1-0 Hamming score: 32 moves 20 MATCH (last 9 in a row!---but after game broken open) 10 NON-MATCH 2 TIE Match %: 21/32 = 66% As high as alleged for Kramnik. For deltas under 0.75 pawn (the boundary from += to +/- in Fritz) 12 MATCH 10 NON-MATCH (including Move 31) 2 TIE This is 13/24 = 54%. *But*, notice that the MATCH-es were among much larger sets of similarly-rated moves than Kramnik's MATCHes in game 6! Perhaps Topalov plays in more-forcing situations? For this reason, my idea of a metric based on predisposed probabilities might find a higher similarity score for Sukhovsky here than for Kramnik in any of the WC games... The results (allowing that Be3 was a deliberate repetition, egged on by Black's King moves?) support Irina Krush's remark that White's play was of a high order.