search

official-stockfish / Stockfish

A free and strong UCI chess engine
https://stockfishchess.org/
GNU General Public License v3.0
11.56k stars 2.27k forks source link

Stockfish does not see the right move #408

Closed braich closed 9 years ago

braich commented 9 years ago

rn3k1r/p2nqpp1/b1p1p2p/3pP3/1p1P1N2/3B2Q1/PPP1NPP1/2KR3R w - - 2 15 In this position the latest Stockfish with 5-men syzygy bases does not see the right move: Ng6+

mcostalba commented 9 years ago

This is not an issue. Please don't open an issue if you are not a developer.

braich commented 9 years ago

@mcostalba Why not? In this postilion Stockfish analysis is wrong. By the way, this position is from this game: http://tcec.chessdom.com/archive.php?se=8&st=1a&ga=46

mcostalba commented 9 years ago

@braich this is for developers not for users. Issues are about specifc code, not SF behaviour in games.

Behaviour in games is very statistically based because the way SF search is strongly statistically based, in some position SF will see the move in others will not, this is absolutely normal and always will be like this. What is important is that on average SF sees more times the good move than it misses it.

braich commented 9 years ago

@mcostalba

I think that this is rather strange to change position evaluation from 0.0: Analysis by Stockfish 220815 64 POPCNT: 15.Kc1-b1 c6-c5 16.Nf4-h5 Qe7-g5 17.Qg3xg5 h6xg5 18.Nh5-f6 Rh8xh1 19.Rd1xh1 Nd7xf6 20.e5xf6 Kf8-g8 21.Bd3-h7+ Kg8-f8 22.Bh7-d3 = (0.00) Depth: 36/45 00:33:44 1725MN, tb=7

to 0.83: rn3k1r/p2nqpp1/b1p1p1Np/3pP3/1p1P4/3B2Q1/PPP1NPP1/2KR3R b - - 0 1 Analysis by Stockfish 220815 64 POPCNT: 15...f7xg6 16.Ne2-f4 Kf8-g8 17.Nf4xg6 Qe7-f7 18.Ng6xh8 Kg8xh8 19.Bd3-g6 Qf7-e7 20.Rh1-h4 Nd7-f8 21.Rh4-f4 Nb8-d7 22.Rf4-f7 Qe7-g5+ 23.Qg3xg5 h6xg5 24.Bg6-h5 Nd7-b6 25.Rf7-c7 g7-g6 26.Bh5-g4 Ba6-b5 27.b2-b3 Kh8-g8 28.Kc1-d2 a7-a5 29.Rc7-e7 Nb6-c8 30.Bg4xe6+ Nf8xe6 31.Re7xe6 Kg8-f7 32.Re6-f6+ Kf7-g7 33.Rd1-a1 a5-a4 34.Ra1-h1 Nc8-e7 35.g2-g4 Ra8-a7 36.Rf6-d6 Ne7-c8 37.Rd6-d8 Nc8-e7 38.f2-f3 a4xb3 +/- (0.83) Depth: 37/64 00:06:21 366MN, tb=24

in one (not the first line) move.

braich commented 9 years ago

@mcostalba But, again, if you do not bother, no problem.