illogical eval from SF? BBKBPK +- 2.00 then BBKBK = 0.00

Discussion of anything and everything relating to chess playing software and machines.

Moderator: Ras

bob
Posts: 20943
Joined: Mon Feb 27, 2006 7:30 pm
Location: Birmingham, AL

Re: illogical eval from SF? BBKBPK +- 2.00 then BBKBK = 0.00

Post by bob »

Eelco de Groot wrote:
mohzus wrote:Wow Eelco this is insane!!! I tested it and the position gives +76 with the latest SF as you said!!! This must be close to a world record seriously :D
However the syzygy (5 pieces at least) fix entirely the problem and SF returns +0.00 from depth 1 and onward.
It is almost funny is it not :) But Stockfish evaluation is not completely insane, he just does not know that he does not have a real bishop pair...

With a real bishop pair, the "known win" eval would have been right, because the knight can be captured by the the stronger side. Shredderchess tablebases online integrated in the GUI, if you don't have Syzygy, give proof:

[d]1BK5/8/4n3/5k2/8/8/8/1B6 b - -
1... Kg4 -M62
1... Kf6 -M61
1...Kg5 -M60
[www.shredderchess.com]
I can't think of any possible way this could happen except for a simple gross bug. Under-promote to a same-color bishop rather than a queen where the search can see winning within reasonable time limit?
Top
bob
Posts: 20943
Joined: Mon Feb 27, 2006 7:30 pm
Location: Birmingham, AL

Re: illogical eval from SF? BBKBPK +- 2.00 then BBKBK = 0.00

Post by bob »

zullil wrote:
Eelco de Groot wrote:Talking about wrong bishop-endgame evals, here is a real testgame played yesterday where one Stockfish on move 93 promotes to a bishop leading to a same coloured bishop pair. The eval is +76.79.
[d] 2K5/1P6/4n3/5k2/8/8/8/B7 w - - 3 93

This needs fixing. Stockfish doesn't switch to the obvious promotion to a queen until depth 29? :shock: (At least it eventually gets the correct evaluation of mate-in-11.)

Code: Select all

Searching: 2K5/1P6/4n3/5k2/8/8/8/B7 w - - 3 93
infinite: 1 ponder: 0 time: 0 increment: 0 moves to go: 0
 1  +76.89   00:00      16  b8=B 
 2  +76.72   00:00      48  b8=B Ke4 
 3  +76.82   00:00     121  b8=B Ke4 Kb7 
 4  +76.74   00:00     237  b8=B Ke4 Kb7 Nd4 
 5  +76.74   00:00     362  b8=B Ke4 Kb7 Nd4 Bb2 
 6  +76.74   00:00     535  b8=B Ke4 Kb7 Nd4 Bb2 Nf3 
 7  +76.74   00:00     988  b8=B Ke4 Kb7 Nd8+ Kb6 Nc6 Bg3 Nd4 
 8  +76.74   00:00    2666  b8=B Ke4 Bb2 Nd4 Kd7 Nc6 Bg3 Nd4 
 9  +76.72   00:00    5198  b8=B Ke4 Bb2 Nd4 Kd7 Kf3 Bc3 Ne2 Bh8 
10  +76.74   00:00    9729  b8=B Ke4 Bh8 Nc5 Bh2 Ne6 Kb7 Nd4 Bg3 Nf3 
11  +76.74   00:00   13775  b8=B Ke4 Bh8 Nd4 Kb7 Kf3 Bd6 Ke4 Bh2 Nf3 Bg3 
12  +76.74   00:00   15875  b8=B Ke4 Bh8 Nd4 Kb7 Kf3 Bd6 Ke4 Bh2 Nf3 Bg3 Nd4 
13  +76.74   00:00   45706  b8=B Nc5 Bb2 Ke4 Bh8 Nd3 Kb7 Ne5 Bc7 Kf3 Ba5 Ke4 Be1 
14  +76.84   00:00   69404  b8=B Nc5 Kd8 Ke4 Ke7 Nd3 Kd6 Nf4 Bc7 Nd3 Ba5 Ne5 Ke6 
                            Nd3 
15  +76.84   00:00   87355  b8=B Nc5 Kd8 Ke4 Ke7 Nd3 Kd6 Nf4 Bc7 Nd3 Ba5 Ne5 Ke6 
                            Nd3 Bb6 
16  +76.82   00:00  132155  b8=B Nc5 Kd8 Nd7 Bc7 Nb6 Bh2 Ke4 Kc7 Nc4 Kd7 Ne5+ 
                            Ke6 Ng4 Bc7 Kf3 
17  +76.82   00:00  162026  b8=B Nc5 Kd8 Nd7 Bc7 Nb6 Bh2 Ke4 Kc7 Nc4 Kd7 Ne3 Bb8 
                            Nc2 Bb2 Ne3 Ke6 Kf3 Bh8 Kg2 
18  +76.82   00:00  204554  b8=B Nc5 Kd8 Nd7 Bc7 Nb6 Bh2 Ke4 Kc7 Nc4 Kd7 Ne3 Bb8 
                            Nc2 Bb2 Ne3 Ke6 Kf3 Bh8 Kg2 
19  +76.84   00:00  343335  b8=B Nc5 Kd8 Nd7 Bc7 Nb6 Ke8 Nd5 Bb8 Ke4 Kf7 Nf4 Kf6 
                            Nd3 Ke6 Ne5 Ba7 Nd3 Kd6 Ne5 Ke6 
20  +76.84   00:00  481409  b8=B Nc5 Kd8 Nd7 Bc7 Nb6 Ke8 Nd5 Bb8 Ke4 Kf7 Nf4 Kf6 
                            Nd3 Ke6 Ne5 Ba7 Nd3 Kd6 Ne5 Bb8 Nf7+ Ke6 
21  +76.84   00:00  745326  b8=B Nc5 Kd8 Nd7 Bc7 Nb6 Ke8 Nd5 Bb8 Ke4 Kf7 Nf4 Kf6 
                            Nd3 Ke6 Ne5 Ba7 Nd3 Kd6 Nf4 Bb8 Nd3 Bc7 Ne5 Ke6 Nd3 
                            Bb8 
22  +76.82   00:00   1561K  b8=B Nc5 Kd8 Ke6 Bh2 Nb3 Bb2 Nc5 Bg1 Nd3 Ba1 Ne5 Ke8 
                            Kf5 Ke7 Nc6+ Kd7 Nd4 Bh2 Ke4 Bb2 Kf3 Bc7 Ke4 Bb6 Kf3 
                            Kd6 Nf5+ Kc6 Ne3 Ba1 Kg2 Bh8 Kf3 Bc7 Kg2 Bb6 
23  +76.82   00:00   1640K  b8=B Nc5 Kd8 Ke6 Bh2 Nb3 Bb2 Nc5 Bg1 Nd3 Ba1 Ne5 Ke8 
                            Kf5 Ke7 Nc6+ Kd7 Nd4 Bh2 Ke4 Bb2 Kf3 Bc7 Ke4 Bb6 Kf3 
                            Kd6 Nf5+ Kc6 Ne3 Ba1 Kg2 Bh8 Kf3 Bc7 Kg2 Bb6 
24  +76.82   00:01   1929K  b8=B Nc5 Kd8 Ke6 Bh2 Nd7 Bc7 Nc5 Ke8 Nb3 Bb2 Nc5 Bh2 
                            Nd3 Ba1 Ne5 Bg1 Nf3 Bb6 Nd4 Bb2 Nf5 Ba7 Nd6+ Kd8 Nc4 
                            Bh8 Ne5 Kc7 Kf5 Bg7 Ke6 Bb8 Kd5 
25  +76.82   00:01   2121K  b8=B Nc5 Kd8 Ke6 Bh2 Nd7 Bc7 Nc5 Ke8 Nb3 Bb2 Nc5 Bh2 
                            Nd3 Ba1 Ne5 Bg1 Nf3 Bb6 Nd4 Bb2 Nf5 Kd8 Nd4 Bc7 Nb5 
                            Bb8 Nd6 Ba1 Nf7+ Kc8 Ne5 Bc7 Kf5 Kd8 Kg4 Ke7 Kh3 Ke6 
                            Kh2 Kd5 
26  +76.82   00:01   2541K  b8=B Nc5 Kd8 Ke6 Bh2 Nd7 Bc7 Nc5 Ke8 Nb3 Bb2 Nc5 Bh2 
                            Nd3 Ba1 Ne5 Bg1 Nf3 Bb6 Nd4 Bb2 Nf5 Kd8 Ne3 Ba5 Kf5 
27  +76.84   00:01   3189K  b8=B Nc5 Kd8 Ke6 Bh2 Nd7 Bc7 Nc5 Ke8 Nb3 Bb2 Nc5 Bh2 
                            Nd3 Ba1 Ne5 Bg3 Kf5 Ke7 Ke4 Kf6 Kf3 Be1 Kg2 Kf5 Nf3 
                            Ba5 Nd4+ Ke4 Nb3 
28  +76.82   00:01   3899K  b8=B Nc5 Kd8 Ke6 Bh2 Nd7 Bc7 Nc5 Ke8 Nb3 Bb2 Nc5 Bh2 
                            Nd7 Ba1 Nf6+ Kf8 Nd7+ Kg7 Ne5 Bg1 Kf5 Bgd4 Nf3 Bdc3 
                            Ne5 Bd4 
29     #11   00:25 290513K  b8=Q Nf4 Qe5+ Kg4 Qe4 Kg5 Bf6+ Kxf6 Qxf4+ Ke6 Kc7 
                            Kd5 Qe3 Kc4 Kc6 Kb4 Qd3 Ka4 Kc5 Ka5 Qb5# 
30     #11   00:29 318053K  b8=Q Nf4 Qe5+ Kg4 Qe4 Kg5 Bf6+ Kxf6 Qxf4+ Ke6 Kc7 
                            Kd5 Qe3 Kc4 Kc6 Kb4 Qd3 Ka4 Kc5 Ka5 Qb5# 
31     #11   00:33 352238K  b8=Q Nf4 Qe5+ Kg4 Qe4 Kg5 Bf6+ Kxf6 Qxf4+ Ke6 Kc7 
                            Kd5 Qe3 Kc4 Kc6 Kb4 Qd3 Ka4 Kc5 Ka5 Qb5# 
32     #11   00:39 395492K  b8=Q Nf4 Qe5+ Kg4 Qe4 Kg5 Bf6+ Kxf6 Qxf4+ Ke6 Kc7 
                            Kd5 Qe3 Kc4 Kc6 Kb4 Qd3 Ka4 Kc5 Ka5 Qb5# 
33     #11   00:43 419812K  b8=Q Nf4 Qe5+ Kg4 Qe4 Kg5 Bf6+ Kxf6 Qxf4+ Ke6 Kc7 
                            Kd5 Qe3 Kc4 Kc6 Kb4 Qd3 Ka4 Kc5 Ka5 Qb5# 
Nodes: 419812094
Nodes/second: 9733193
Best move: b8=Q
Ponder move: Nf4
I could think of a couple of easy ways.

1. K+B+B vs KN is not a win if the two B's are the same color, so that's a flaw.

2. K+B+Q vs KN is just as winnable as KBB vs KN so including it in the "recognized wins" would also solve this.

As it is, the large (wrong) KBBvsKN score is making it impossible to find the queen promotion until the search is deep enough that the actual mate score is seen making it better than that +76 nonsensical score...
Top
User avatar
Eelco de Groot
Posts: 4664
Joined: Sun Mar 12, 2006 2:40 am
Full name:   Eelco de Groot

Re: illogical eval from SF? BBKBPK +- 2.00 then BBKBK = 0.00

Post by Eelco de Groot »

zullil wrote:
Eelco de Groot wrote:Talking about wrong bishop-endgame evals, here is a real testgame played yesterday where one Stockfish on move 93 promotes to a bishop leading to a same coloured bishop pair. The eval is +76.79.
[d] 2K5/1P6/4n3/5k2/8/8/8/B7 w - - 3 93

This needs fixing. Stockfish doesn't switch to the obvious promotion to a queen until depth 29? :shock: (At least it eventually gets the correct evaluation of mate-in-11.)

Code: Select all

Searching: 2K5/1P6/4n3/5k2/8/8/8/B7 w - - 3 93
infinite: 1 ponder: 0 time: 0 increment: 0 moves to go: 0
 1  +76.89   00:00      16  b8=B 
 2  +76.72   00:00      48  b8=B Ke4 
 3  +76.82   00:00     121  b8=B Ke4 Kb7 
 4  +76.74   00:00     237  b8=B Ke4 Kb7 Nd4 
 5  +76.74   00:00     362  b8=B Ke4 Kb7 Nd4 Bb2 
 6  +76.74   00:00     535  b8=B Ke4 Kb7 Nd4 Bb2 Nf3 
 7  +76.74   00:00     988  b8=B Ke4 Kb7 Nd8+ Kb6 Nc6 Bg3 Nd4 
 8  +76.74   00:00    2666  b8=B Ke4 Bb2 Nd4 Kd7 Nc6 Bg3 Nd4 
 9  +76.72   00:00    5198  b8=B Ke4 Bb2 Nd4 Kd7 Kf3 Bc3 Ne2 Bh8 
10  +76.74   00:00    9729  b8=B Ke4 Bh8 Nc5 Bh2 Ne6 Kb7 Nd4 Bg3 Nf3 
11  +76.74   00:00   13775  b8=B Ke4 Bh8 Nd4 Kb7 Kf3 Bd6 Ke4 Bh2 Nf3 Bg3 
12  +76.74   00:00   15875  b8=B Ke4 Bh8 Nd4 Kb7 Kf3 Bd6 Ke4 Bh2 Nf3 Bg3 Nd4 
13  +76.74   00:00   45706  b8=B Nc5 Bb2 Ke4 Bh8 Nd3 Kb7 Ne5 Bc7 Kf3 Ba5 Ke4 Be1 
14  +76.84   00:00   69404  b8=B Nc5 Kd8 Ke4 Ke7 Nd3 Kd6 Nf4 Bc7 Nd3 Ba5 Ne5 Ke6 
                            Nd3 
15  +76.84   00:00   87355  b8=B Nc5 Kd8 Ke4 Ke7 Nd3 Kd6 Nf4 Bc7 Nd3 Ba5 Ne5 Ke6 
                            Nd3 Bb6 
16  +76.82   00:00  132155  b8=B Nc5 Kd8 Nd7 Bc7 Nb6 Bh2 Ke4 Kc7 Nc4 Kd7 Ne5+ 
                            Ke6 Ng4 Bc7 Kf3 
17  +76.82   00:00  162026  b8=B Nc5 Kd8 Nd7 Bc7 Nb6 Bh2 Ke4 Kc7 Nc4 Kd7 Ne3 Bb8 
                            Nc2 Bb2 Ne3 Ke6 Kf3 Bh8 Kg2 
18  +76.82   00:00  204554  b8=B Nc5 Kd8 Nd7 Bc7 Nb6 Bh2 Ke4 Kc7 Nc4 Kd7 Ne3 Bb8 
                            Nc2 Bb2 Ne3 Ke6 Kf3 Bh8 Kg2 
19  +76.84   00:00  343335  b8=B Nc5 Kd8 Nd7 Bc7 Nb6 Ke8 Nd5 Bb8 Ke4 Kf7 Nf4 Kf6 
                            Nd3 Ke6 Ne5 Ba7 Nd3 Kd6 Ne5 Ke6 
20  +76.84   00:00  481409  b8=B Nc5 Kd8 Nd7 Bc7 Nb6 Ke8 Nd5 Bb8 Ke4 Kf7 Nf4 Kf6 
                            Nd3 Ke6 Ne5 Ba7 Nd3 Kd6 Ne5 Bb8 Nf7+ Ke6 
21  +76.84   00:00  745326  b8=B Nc5 Kd8 Nd7 Bc7 Nb6 Ke8 Nd5 Bb8 Ke4 Kf7 Nf4 Kf6 
                            Nd3 Ke6 Ne5 Ba7 Nd3 Kd6 Nf4 Bb8 Nd3 Bc7 Ne5 Ke6 Nd3 
                            Bb8 
22  +76.82   00:00   1561K  b8=B Nc5 Kd8 Ke6 Bh2 Nb3 Bb2 Nc5 Bg1 Nd3 Ba1 Ne5 Ke8 
                            Kf5 Ke7 Nc6+ Kd7 Nd4 Bh2 Ke4 Bb2 Kf3 Bc7 Ke4 Bb6 Kf3 
                            Kd6 Nf5+ Kc6 Ne3 Ba1 Kg2 Bh8 Kf3 Bc7 Kg2 Bb6 
23  +76.82   00:00   1640K  b8=B Nc5 Kd8 Ke6 Bh2 Nb3 Bb2 Nc5 Bg1 Nd3 Ba1 Ne5 Ke8 
                            Kf5 Ke7 Nc6+ Kd7 Nd4 Bh2 Ke4 Bb2 Kf3 Bc7 Ke4 Bb6 Kf3 
                            Kd6 Nf5+ Kc6 Ne3 Ba1 Kg2 Bh8 Kf3 Bc7 Kg2 Bb6 
24  +76.82   00:01   1929K  b8=B Nc5 Kd8 Ke6 Bh2 Nd7 Bc7 Nc5 Ke8 Nb3 Bb2 Nc5 Bh2 
                            Nd3 Ba1 Ne5 Bg1 Nf3 Bb6 Nd4 Bb2 Nf5 Ba7 Nd6+ Kd8 Nc4 
                            Bh8 Ne5 Kc7 Kf5 Bg7 Ke6 Bb8 Kd5 
25  +76.82   00:01   2121K  b8=B Nc5 Kd8 Ke6 Bh2 Nd7 Bc7 Nc5 Ke8 Nb3 Bb2 Nc5 Bh2 
                            Nd3 Ba1 Ne5 Bg1 Nf3 Bb6 Nd4 Bb2 Nf5 Kd8 Nd4 Bc7 Nb5 
                            Bb8 Nd6 Ba1 Nf7+ Kc8 Ne5 Bc7 Kf5 Kd8 Kg4 Ke7 Kh3 Ke6 
                            Kh2 Kd5 
26  +76.82   00:01   2541K  b8=B Nc5 Kd8 Ke6 Bh2 Nd7 Bc7 Nc5 Ke8 Nb3 Bb2 Nc5 Bh2 
                            Nd3 Ba1 Ne5 Bg1 Nf3 Bb6 Nd4 Bb2 Nf5 Kd8 Ne3 Ba5 Kf5 
27  +76.84   00:01   3189K  b8=B Nc5 Kd8 Ke6 Bh2 Nd7 Bc7 Nc5 Ke8 Nb3 Bb2 Nc5 Bh2 
                            Nd3 Ba1 Ne5 Bg3 Kf5 Ke7 Ke4 Kf6 Kf3 Be1 Kg2 Kf5 Nf3 
                            Ba5 Nd4+ Ke4 Nb3 
28  +76.82   00:01   3899K  b8=B Nc5 Kd8 Ke6 Bh2 Nd7 Bc7 Nc5 Ke8 Nb3 Bb2 Nc5 Bh2 
                            Nd7 Ba1 Nf6+ Kf8 Nd7+ Kg7 Ne5 Bg1 Kf5 Bgd4 Nf3 Bdc3 
                            Ne5 Bd4 
29     #11   00:25 290513K  b8=Q Nf4 Qe5+ Kg4 Qe4 Kg5 Bf6+ Kxf6 Qxf4+ Ke6 Kc7 
                            Kd5 Qe3 Kc4 Kc6 Kb4 Qd3 Ka4 Kc5 Ka5 Qb5# 
30     #11   00:29 318053K  b8=Q Nf4 Qe5+ Kg4 Qe4 Kg5 Bf6+ Kxf6 Qxf4+ Ke6 Kc7 
                            Kd5 Qe3 Kc4 Kc6 Kb4 Qd3 Ka4 Kc5 Ka5 Qb5# 
31     #11   00:33 352238K  b8=Q Nf4 Qe5+ Kg4 Qe4 Kg5 Bf6+ Kxf6 Qxf4+ Ke6 Kc7 
                            Kd5 Qe3 Kc4 Kc6 Kb4 Qd3 Ka4 Kc5 Ka5 Qb5# 
32     #11   00:39 395492K  b8=Q Nf4 Qe5+ Kg4 Qe4 Kg5 Bf6+ Kxf6 Qxf4+ Ke6 Kc7 
                            Kd5 Qe3 Kc4 Kc6 Kb4 Qd3 Ka4 Kc5 Ka5 Qb5# 
33     #11   00:43 419812K  b8=Q Nf4 Qe5+ Kg4 Qe4 Kg5 Bf6+ Kxf6 Qxf4+ Ke6 Kc7 
                            Kd5 Qe3 Kc4 Kc6 Kb4 Qd3 Ka4 Kc5 Ka5 Qb5# 
Nodes: 419812094
Nodes/second: 9733193
Best move: b8=Q
Ponder move: Nf4
This is with a proposed bugfix, one thread, 64 Mb hash. Stockfish now thinks it is mate in 12 after 93. b8Q but maybe it is better with more cores.

2K5/1P6/4n3/5k2/8/8/8/B7 w - -

Engine: Stockfish Master 20140109 with KBBKNbugfix (64 MB)
by Tord Romstad, Marco Costalba and Joona Kiiski

20/30 0:02 +M12 93.b8Q Nf4 94.Qe5+ Kg4 95.Qe4 Kg5
96.Bf6+ Kxf6 97.Qxf4+ Ke6 98.Qg5 Kd6
99.Qf5 Kc6 100.Qe5 Kb6 101.Kd7 Ka7
102.Qb5 Ka8 103.Kc7 Ka7 104.Qa4+ (5.141.143) 1746

21/30 0:03 +M12 93.b8Q Nf4 94.Qe5+ Kg4 95.Qe4 Kg5
96.Bf6+ Kxf6 97.Qxf4+ Ke6 98.Qg5 Kd6
99.Qf5 Kc6 100.Qe5 Kb6 101.Kd7 Ka7
102.Qb5 Ka8 103.Kc7 Ka7 104.Qa4+ (6.569.167) 1713

22/30 0:05 +M12 93.b8Q Nf4 94.Qe5+ Kg4 95.Qe4 Kg5
96.Bf6+ Kxf6 97.Qxf4+ Ke6 98.Qg5 Kd6
99.Qf5 Kc6 100.Qe5 Kb6 101.Kd7 Ka7
102.Qb5 Ka8 103.Kc7 Ka7 104.Qa4+ (9.464.921) 1694

23/30 0:07 +M12 93.b8Q Nf4 94.Qe5+ Kg4 95.Qe4 Kg5
96.Bf6+ Kxf6 97.Qxf4+ Ke6 98.Qg5 Kd6
99.Qf5 Kc6 100.Qe5 Kb6 101.Kd7 Ka7
102.Qb5 Ka8 103.Kc7 Ka7 104.Qa4+ (12.517.156) 1681

24/30 0:10 +M12 93.b8Q Nf4 94.Qe5+ Kg4 95.Qe4 Kg5
96.Bf6+ Kxf6 97.Qxf4+ Ke6 98.Qg5 Kd6
99.Qf5 Kc6 100.Qe5 Kb6 101.Kd7 Ka7
102.Qb5 Ka8 103.Kc7 Ka7 104.Qa4+ (17.646.684) 1655

25/30 0:16 +M12 93.b8Q Nf4 94.Qe5+ Kg4 95.Qe4 Kg5
96.Bf6+ Kxf6 97.Qxf4+ Ke6 98.Qg5 Kd6
99.Qf5 Kc6 100.Qe5 Kb6 101.Kd7 Ka7
102.Qb5 Ka8 103.Kc7 Ka7 104.Qa4+ (26.599.902) 1639

26/30 0:26 +M12 93.b8Q Nf4 94.Qe5+ Kg4 95.Qe4 Kg5
96.Bf6+ Kxf6 97.Qxf4+ Ke6 98.Qg5 Kd6
99.Qf5 Kc6 100.Qe5 Kb6 101.Kd7 Ka7
102.Qb5 Ka8 103.Kc7 Ka7 104.Qa4+ (43.124.351) 1618

27/30 0:47 +M12 93.b8Q Kg4 94.Qb4+ Kf5 95.Qb5+ Kg4
96.Qe2+ Kf5 97.Qe5+ Kg4 98.Qxe6+ Kf4
99.Qe2 Kg3 100.Be5+ Kh4 101.Bf4 Kh3
102.Qf3+ Kh4 103.Qg3+ Kh5 104.Qg5+ (76.005.285) 1610

28/30 1:01 +M12 93.b8Q Kg4 94.Qb4+ Kf5 95.Qb5+ Kg4
96.Qe2+ Kf5 97.Qe5+ Kg4 98.Qxe6+ Kf4
99.Qe2 Kg3 100.Be5+ Kh4 101.Bf4 Kh3
102.Qf3+ (98.759.265) 1613

29/30 1:30 +M12 93.b8Q Kg4 94.Qb4+ Kf5 95.Qb5+ Kg4
96.Qe2+ Kf5 97.Qe5+ Kg4 98.Qxe6+ Kf4
99.Qe2 Kf5 100.Qc4 Kg5 101.Kd7 Kh5
102.Ke6 Kg5 103.Qb4 (146.626.929) 1613

30/30 2:01 +M12 93.b8Q Kg4 94.Qb4+ Kf5 95.Qb5+ Kg4
96.Qe2+ Kf5 97.Qe5+ Kg4 98.Qxe6+ Kf4
99.Qe2 Kf5 100.Qc4 Kg5 101.Qe4 Kh5
102.Bf6 Kh6 103.Qg4 Kh7 104.Qg7+ (196.181.225) 1618

31/30 2:53 +M12 93.b8Q Kg4 94.Qb4+ Kf5 95.Qb5+ Kg4
96.Qe2+ Kf5 97.Qe5+ Kg4 98.Qxe6+ Kf4
99.Qe2 Kg3 100.Be5+ Kh4 101.Bf4 Kh3
102.Qf3+ Kh4 103.Qg3+ Kh5 104.Qg5+ (280.741.435) 1619

32/30 4:01 +M12 93.b8Q Kg4 94.Qb4+ Kf5 95.Qb5+ Kg4
96.Qe2+ Kf5 97.Qe5+ (392.625.229) 1623

33/30 6:53 +M12 93.b8Q Kg4 94.Qb4+ Kf5 95.Qb5+ Kg4
96.Qe2+ Kf5 97.Qe5+ Kg4 98.Qxe6+ Kf4
99.Qe5+ Kf3 100.Qd4 Ke2 101.Qe4+ Kd2
102.Bd4 Kd1 103.Qd3+ Kc1 104.Qe2 Kb1
105.Qb2+ (676.507.188) 1636

34/30 10:41 +M12 93.b8Q Kg4 94.Qb4+ Kf5 95.Qb5+ Kg4
96.Qe2+ Kf5 97.Qe5+ Kg4 98.Qxe6+ Kf3
99.Bd4 Kf4 100.Qe3+ Kg4 101.Qe4+ Kg5
102.Be5 Kh5 103.Qg2 Kh4 104.Bf6+ Kh5
105.Qg5+ (1.050.091.248) 1636

35/30 10:47 +M12 93.b8Q Kg4 94.Qb4+ Kf5 95.Qb5+ Kg4
96.Qe2+ Kf5 97.Qe5+ Kg4 98.Qxe6+ Kf3
99.Bd4 Kf4 100.Qe3+ Kg4 101.Qe4+ Kg5
102.Be5 Kh5 103.Qg2 Kh4 104.Bf6+ Kh5
105.Qg5+ (1.060.798.426) 1637

best move: b7-b8Q time: 10:47.794 min n/s: 1.637.866 nodes: 1.060.798.426

Actually there are only four moves that give away a half point, 93. b8B?, 93. b8N?, 93. Kb8? and 93. Kd7? because of the possible Knightfork Nc5+
Debugging is twice as hard as writing the code in the first
place. Therefore, if you write the code as cleverly as possible, you
are, by definition, not smart enough to debug it.
-- Brian W. Kernighan
Top
Vinvin
Posts: 5290
Joined: Thu Mar 09, 2006 9:40 am
Full name: Vincent Lejeune

Re: illogical eval from SF? BBKBPK +- 2.00 then BBKBK = 0.00

Post by Vinvin »

Eelco de Groot wrote:
zullil wrote:
Eelco de Groot wrote:Talking about wrong bishop-endgame evals, here is a real testgame played yesterday where one Stockfish on move 93 promotes to a bishop leading to a same coloured bishop pair. The eval is +76.79.
2K5/1P6/4n3/5k2/8/8/8/B7 w - - 3 93

This needs fixing. Stockfish doesn't switch to the obvious promotion to a queen until depth 29? :shock: (At least it eventually gets the correct evaluation of mate-in-11.)
...
This is with a proposed bugfix, one thread, 64 Mb hash. Stockfish now thinks it is mate in 12 after 93. b8Q but maybe it is better with more cores.
Is the bugfix covers the case N vs BB(same color) ?
Do SF sees h8=B is draw here ?
[d]kn6/7P/1K6/8/8/8/8/B7 w - -
Top
Joerg Oster
Posts: 974
Joined: Fri Mar 10, 2006 4:29 pm
Location: Germany
Full name: Jörg Oster

Re: illogical eval from SF? BBKBPK +- 2.00 then BBKBK = 0.00

Post by Joerg Oster »

Hi Eelco,

does it now play b8q immediately?
My bugfix version (with a slight addition to yours ;-) ) does!

Code: Select all

info depth 1 seldepth 1 score cp 1260 nodes 15 nps 7500 time 2 multipv 1 pv b7b8q
info depth 2 seldepth 2 score cp 1290 nodes 128 nps 42666 time 3 multipv 1 pv b7b8q f5g4 b8b4 g4f5
Best, Joerg.
Jörg Oster
Top
User avatar
Eelco de Groot
Posts: 4664
Joined: Sun Mar 12, 2006 2:40 am
Full name:   Eelco de Groot

Re: illogical eval from SF? BBKBPK +- 2.00 then BBKBK = 0.00

Post by Eelco de Groot »

Joerg Oster wrote:Hi Eelco,

does it now play b8q immediately?
My bugfix version (with a slight addition to yours ;-) ) does!

Code: Select all

info depth 1 seldepth 1 score cp 1260 nodes 15 nps 7500 time 2 multipv 1 pv b7b8q
info depth 2 seldepth 2 score cp 1290 nodes 128 nps 42666 time 3 multipv 1 pv b7b8q f5g4 b8b4 g4f5
Best, Joerg.
Hi Joerg,

Yes, Shredder only starts the output after 20 plies but b8B is now much lower in the list hopefully, Shredder put it around move 8 or 9. In console mode I get

Stockfish 110114 by Tord Romstad, Marco Costalba and Joona Kiiski
position fen 2K5/1P6/4n3/5k2/8/8/8/B7 w - - 3 93
go infinite
info depth 1 seldepth 1 score cp 1266 nodes 15 nps 5000 time 3 multipv 1 pv b7b8
q
info depth 2 seldepth 2 score cp 1302 nodes 130 nps 32500 time 4 multipv 1 pv b7
b8q f5g4 b8b4 g4f5
info depth 3 seldepth 4 score cp 1341 nodes 341 nps 68200 time 5 multipv 1 pv b7
b8q e6f4 b8e5 f5g4
info depth 4 seldepth 5 score cp 1380 nodes 1081 nps 154428 time 7 multipv 1 pv
b7b8q e6f4 b8e5 f5g4 a1d4
info depth 5 seldepth 6 score cp 1393 nodes 2954 nps 268545 time 11 multipv 1 pv
b7b8q e6f4 b8e5 f5g4 a1d4 g4f3

Eelco
Debugging is twice as hard as writing the code in the first
place. Therefore, if you write the code as cleverly as possible, you
are, by definition, not smart enough to debug it.
-- Brian W. Kernighan
Top
Joerg Oster
Posts: 974
Joined: Fri Mar 10, 2006 4:29 pm
Location: Germany
Full name: Jörg Oster

Re: illogical eval from SF? BBKBPK +- 2.00 then BBKBK = 0.00

Post by Joerg Oster »

No problem! :D

Code: Select all

info depth 1 seldepth 2 score cp 0 nodes 21 nps 10500 time 2 multipv 1 pv b8d7 b6a5
info depth 2 seldepth 3 score cp 0 nodes 67 nps 33500 time 2 multipv 1 pv b8d7 b6a5 d7c5
info depth 3 seldepth 4 score cp 0 nodes 158 nps 79000 time 2 multipv 1 pv b8d7 b6a5 d7c5 h8b2
Jörg Oster
Top
User avatar
Eelco de Groot
Posts: 4664
Joined: Sun Mar 12, 2006 2:40 am
Full name:   Eelco de Groot

Re: illogical eval from SF? BBKBPK +- 2.00 then BBKBK = 0.00

Post by Eelco de Groot »

Vinvin wrote:
Eelco de Groot wrote:
zullil wrote:
Eelco de Groot wrote:Talking about wrong bishop-endgame evals, here is a real testgame played yesterday where one Stockfish on move 93 promotes to a bishop leading to a same coloured bishop pair. The eval is +76.79.
2K5/1P6/4n3/5k2/8/8/8/B7 w - - 3 93

This needs fixing. Stockfish doesn't switch to the obvious promotion to a queen until depth 29? :shock: (At least it eventually gets the correct evaluation of mate-in-11.)
...
This is with a proposed bugfix, one thread, 64 Mb hash. Stockfish now thinks it is mate in 12 after 93. b8Q but maybe it is better with more cores.
Is the bugfix covers the case N vs BB(same color) ?
Do SF sees h8=B is draw here ?
[d]kn6/7P/1K6/8/8/8/8/B7 w - -
Hi Vincent,

This is maybe not such a good test, Stockfish sees mate in 5 at depth 15 :)

Stockfish 110114 by Tord Romstad, Marco Costalba and Joona Kiiski
position fen kn6/7P/1K6/8/8/8/8/B7 w - -
go depth 40
info depth 1 seldepth 1 score cp 1489 nodes 17 nps 4250 time 4 multipv 1 pv h7h8
q
info depth 2 seldepth 3 score cp 1360 nodes 227 nps 56750 time 4 multipv 1 pv b6
b5 b8c6 h7h8q a8b7
info depth 3 seldepth 4 score cp 1374 nodes 366 nps 73200 time 5 multipv 1 pv b6
b5 b8d7 h7h8q a8b7
info depth 4 seldepth 5 score cp 1447 nodes 517 nps 103400 time 5 multipv 1 pv b
6b5 b8d7 h7h8q a8b7 h8g7
info depth 5 seldepth 6 score cp 1480 nodes 1198 nps 171142 time 7 multipv 1 pv
b6b5 b8d7 h7h8q a8b7 h8d8 d7b8 d8b6 b7c8
info depth 6 seldepth 8 score cp 1530 nodes 2327 nps 258555 time 9 multipv 1 pv
b6b5 a8a7 h7h8q a7b7 h8g7 b7c8 a1e5
info depth 7 seldepth 8 score cp 8924 nodes 2935 nps 293500 time 10 multipv 1 pv
b6b5 a8a7 h7h8q a7b7 h8g7 b7c8 g7e7 b8d7 e7e8 c8c7 a1e5
info depth 8 seldepth 10 score cp 8924 nodes 3231 nps 293727 time 11 multipv 1 p
v b6b5 a8a7 h7h8q a7b7 h8g7 b7c8 g7e7 b8d7 e7e8 c8c7 a1e5 d7e5 e8e5 c7d7
info depth 9 seldepth 12 score cp 8924 nodes 4044 nps 337000 time 12 multipv 1 p
v b6b5 b8d7 h7h8q a8b7 h8e8 b7c7 a1e5 d7e5 e8e5 c7d7 e5e4
info depth 10 seldepth 16 score cp 8934 nodes 6023 nps 401533 time 15 multipv 1
pv b6b5 a8a7 h7h8q a7b7 h8g7 b7c8 g7e7 b8d7 e7e8 c8c7 a1e5 d7e5 e8e5 c7d7 b5b6
info depth 11 seldepth 16 score cp 8934 nodes 6561 nps 437400 time 15 multipv 1
pv b6b5 a8a7 h7h8q a7b7 h8g7 b7c8 g7e7 b8d7 e7e8 c8c7 a1e5 d7e5 e8e5 c7d7 b5b6
info depth 12 seldepth 16 score cp 9346 nodes 36118 nps 737102 time 49 multipv 1
pv b6b5 a8b7 h7h8q b7c7 h8e8 c7d6 e8b8 d6e6 b8f4 e6d7 a1d4 d7e6 f4f6 e6d7
info depth 13 seldepth 16 score cp 9357 nodes 57753 nps 837000 time 69 multipv 1
pv b6b5 a8b7 h7h8q b7c7 h8e8 c7d6 e8b8 d6e7 b8f4 e7e6 a1d4 e6e7 b5c6 e7e8
info depth 14 seldepth 20 score mate 5 nodes 187660 nps 1008924 time 186 multipv
1 pv a1c3 b8d7 b6c7 a8a7 h7h8q a7a6 h8a8 a6b5 a8c6
info depth 15 seldepth 20 score mate 5 nodes 191192 nps 1011597 time 189 multipv
1 pv a1c3 b8d7 b6c7 a8a7 h7h8q a7a6 h8a8 a6b5 a8c6
info depth 16 seldepth 20 score mate 5 nodes 196902 nps 1009753 time 195 multipv
1 pv a1c3 b8d7 b6c7 a8a7 h7h8q a7a6 h8a8 a6b5 a8c6
info depth 17 seldepth 20 score mate 5 nodes 204793 nps 1023965 time 200 multipv
1 pv a1c3 b8d7 b6c7 a8a7 h7h8q a7a6 h8a8 a6b5 a8c6
info depth 18 seldepth 20 score mate 5 nodes 215087 nps 1024223 time 210 multipv
1 pv a1c3 b8d7 b6c7 a8a7 h7h8q a7a6 h8a8 a6b5 a8c6
info depth 19 seldepth 20 score mate 5 nodes 227887 nps 1055032 time 216 multipv
1 pv a1c3 b8d7 b6c7 a8a7 h7h8q a7a6 h8a8 a6b5 a8c6
info depth 20 seldepth 20 score mate 5 nodes 243386 nps 1072185 time 227 multipv
1 pv a1c3 b8d7 b6c7 a8a7 h7h8q a7a6 h8a8 a6b5 a8c6
info depth 21 seldepth 20 score mate 5 nodes 261850 nps 1109533 time 236 multipv
1 pv a1c3 b8d7 b6c7 a8a7 h7h8q a7a6 h8a8 a6b5 a8c6
info depth 22 seldepth 20 score mate 5 nodes 282395 nps 1129580 time 250 multipv
1 pv a1c3 b8d7 b6c7 a8a7 h7h8q a7a6 h8a8 a6b5 a8c6
info depth 23 seldepth 20 score mate 5 nodes 305073 nps 1177888 time 259 multipv
1 pv a1c3 b8d7 b6c7 a8a7 h7h8q a7a6 h8a8 a6b5 a8c6
info depth 24 seldepth 20 score mate 5 nodes 329757 nps 1212341 time 272 multipv
1 pv a1c3 b8d7 b6c7 a8a7 h7h8q a7a6 h8a8 a6b5 a8c6
info depth 25 seldepth 20 score mate 5 nodes 354921 nps 1258585 time 282 multipv
1 pv a1c3 b8d7 b6c7 a8a7 h7h8q a7a6 h8a8 a6b5 a8c6
info depth 26 seldepth 20 score mate 5 nodes 380858 nps 1304308 time 292 multipv
1 pv a1c3 b8d7 b6c7 a8a7 h7h8q a7a6 h8a8 a6b5 a8c6
info depth 27 seldepth 20 score mate 5 nodes 407046 nps 1343386 time 303 multipv
1 pv a1c3 b8d7 b6c7 a8a7 h7h8q a7a6 h8a8 a6b5 a8c6
info depth 28 seldepth 20 score mate 5 nodes 432670 nps 1382332 time 313 multipv
1 pv a1c3 b8d7 b6c7 a8a7 h7h8q a7a6 h8a8 a6b5 a8c6
info depth 29 seldepth 20 score mate 5 nodes 458628 nps 1415518 time 324 multipv
1 pv a1c3 b8d7 b6c7 a8a7 h7h8q a7a6 h8a8 a6b5 a8c6
info depth 30 seldepth 20 score mate 5 nodes 485435 nps 1449059 time 335 multipv
1 pv a1c3 b8d7 b6c7 a8a7 h7h8q a7a6 h8a8 a6b5 a8c6
info depth 31 seldepth 20 score mate 5 nodes 511146 nps 1481582 time 345 multipv
1 pv a1c3 b8d7 b6c7 a8a7 h7h8q a7a6 h8a8 a6b5 a8c6
info depth 32 seldepth 20 score mate 5 nodes 538334 nps 1507938 time 357 multipv
1 pv a1c3 b8d7 b6c7 a8a7 h7h8q a7a6 h8a8 a6b5 a8c6
info depth 33 seldepth 20 score mate 5 nodes 564079 nps 1532823 time 368 multipv
1 pv a1c3 b8d7 b6c7 a8a7 h7h8q a7a6 h8a8 a6b5 a8c6
info depth 34 seldepth 20 score mate 5 nodes 593195 nps 1561039 time 380 multipv
1 pv a1c3 b8d7 b6c7 a8a7 h7h8q a7a6 h8a8 a6b5 a8c6
info depth 35 seldepth 20 score mate 5 nodes 621794 nps 1582173 time 393 multipv
1 pv a1c3 b8d7 b6c7 a8a7 h7h8q a7a6 h8a8 a6b5 a8c6
info depth 36 seldepth 20 score mate 5 nodes 654275 nps 1607555 time 407 multipv
1 pv a1c3 b8d7 b6c7 a8a7 h7h8q a7a6 h8a8 a6b5 a8c6
info depth 37 seldepth 20 score mate 5 nodes 694209 nps 1629598 time 426 multipv
1 pv a1c3 b8d7 b6c7 a8a7 h7h8q a7a6 h8a8 a6b5 a8c6
info depth 38 seldepth 20 score mate 5 nodes 729347 nps 1650106 time 442 multipv
1 pv a1c3 b8d7 b6c7 a8a7 h7h8q a7a6 h8a8 a6b5 a8c6
info depth 39 seldepth 20 score mate 5 nodes 765845 nps 1664880 time 460 multipv
1 pv a1c3 b8d7 b6c7 a8a7 h7h8q a7a6 h8a8 a6b5 a8c6
info depth 40 seldepth 20 score mate 5 nodes 804147 nps 1682315 time 478 multipv
1 pv a1c3 b8d7 b6c7 a8a7 h7h8q a7a6 h8a8 a6b5 a8c6
info nodes 804147 time 479
bestmove a1c3 ponder b8d7

Stockfish 110114 by Tord Romstad, Marco Costalba and Joona Kiiski
position fen [d]kn5B/8/1K6/8/8/8/8/B7 b - -
go depth 40
info depth 1 seldepth 2 score cp -136 nodes 36 nps 12000 time 3 multipv 1 pv b8d
7 b6c7
info depth 2 seldepth 3 score cp -118 nodes 77 nps 19250 time 4 multipv 1 pv b8d
7 b6c7 d7b8
info depth 3 seldepth 5 score cp -123 nodes 241 nps 60250 time 4 multipv 1 pv b8
d7 b6c6 d7b8 c6c7 a8a7
info depth 4 seldepth 7 score cp -118 nodes 463 nps 92600 time 5 multipv 1 pv b8
d7 b6c6 d7b8 c6c7 b8a6 c7b6 a6b8
info depth 5 seldepth 8 score cp -123 nodes 624 nps 124800 time 5 multipv 1 pv b
8d7 b6c6 d7b8 c6c7 b8a6 c7b6 a6b8 b6a5
info depth 6 seldepth 10 score cp -123 nodes 1147 nps 163857 time 7 multipv 1 pv
b8d7 b6c6 d7b8 c6c7 a8a7 h8b2 b8a6 c7c6 a6b8 c6b5
info depth 7 seldepth 11 score cp -123 nodes 1634 nps 204250 time 8 multipv 1 pv
b8d7 b6c6 d7b8 c6c7 a8a7 h8b2 b8a6 c7c8 a6b8 b2a3
info depth 8 seldepth 13 score cp -126 nodes 2204 nps 244888 time 9 multipv 1 pv
b8d7 b6c6 d7b8 c6c7 a8a7 a1d4 a7a8 d4g1 b8a6 c7c6 a6c7 h8b2
info depth 9 seldepth 15 score cp -123 nodes 3314 nps 331400 time 10 multipv 1 p
v b8d7 b6c6 d7b8 c6c7 b8a6 c7b6 a6b8 h8b2 b8d7 b6c6 d7b8 c6c7 a8a7 b2a3 b8a6 c7c
8
info depth 10 seldepth 18 score cp -136 nodes 8936 nps 446800 time 20 multipv 1
pv b8d7 b6c7 d7b8 a1c3 a8a7 h8d4 a7a8 c7b6 b8d7 b6c6 d7b8 c6c7 b8a6 c7b6 a6b8 c3
b2 b8d7 b6c7 d7c5
info depth 11 seldepth 18 score cp -136 nodes 10441 nps 453956 time 23 multipv 1
pv b8d7 b6c7 d7b8 a1c3 a8a7 h8d4 a7a8 c7b6 b8d7 b6c6 d7b8 c6c7 b8a6 c7b6 a6b8 c
3b2 b8d7 b6c7 d7c5
info depth 12 seldepth 22 score cp -126 nodes 14479 nps 499275 time 29 multipv 1
pv b8d7 b6c7 d7b8 a1c3 a8a7 h8d4 a7a8 c7b6 b8d7 b6c6 d7b8 c6c7 b8a6 c7b6 a6c7 c
3b2 c7d5 b6c6 d5e7 c6c7 e7c8
info depth 13 seldepth 22 score cp -126 nodes 15518 nps 470242 time 33 multipv 1
pv b8d7 b6c7 d7b8 a1c3 a8a7 h8d4 a7a8 c7b6 b8d7 b6c6 d7b8 c6c7 b8a6 c7b6 a6c7 c
3b2 c7d5 b6c6 d5e7 c6c7 e7c8
info depth 14 seldepth 22 score cp -126 nodes 20332 nps 508300 time 40 multipv 1
pv b8d7 b6c7 d7b8 a1c3 b8a6 c7b6 a6b8 h8d4 b8d7 b6c6 d7b8 c6c7 b8a6 c7b6 a6c7 c
3b2 c7d5 b6c6 d5e7 c6c7 e7c8 d4g1
info depth 15 seldepth 23 score cp -126 nodes 29762 nps 572346 time 52 multipv 1
pv b8d7 b6c7 d7b8 a1c3 b8a6 c7b6 a6b8 h8d4 b8d7 b6c6 d7b8 c6c7 b8a6 c7b6 a6c7 b
6c6 c7a6 d4e5 a8a7 c3d4 a7a8 d4b2 a8a7 b2d4
info depth 16 seldepth 30 score cp -126 nodes 56671 nps 691109 time 82 multipv 1
pv b8d7 b6c7 d7b8 a1c3 b8a6 c7b6 a6b8 h8d4 b8d7 b6c6 d7b8 c6c7 b8a6 c7b6 a6c7 b
6c6 c7a6 d4e5 a6b8 c6c7 b8a6 c7b6 a6b4 c3d2 b4d5 b6c6 d5e7 c6c7 e7d5 c7d8
info depth 17 seldepth 30 score cp -126 nodes 57288 nps 682000 time 84 multipv 1
pv b8d7 b6c7 d7b8 a1c3 b8a6 c7b6 a6b8 h8d4 b8d7 b6c6 d7b8 c6c7 b8a6 c7b6 a6c7 b
6c6 c7a6 d4e5 a6b8 c6c7 b8a6 c7b6 a6b4 c3d2 b4d5 b6c6 d5e7 c6c7 e7d5 c7d8
info depth 18 seldepth 30 score cp -128 nodes 86971 nps 776526 time 112 multipv
1 pv b8d7 b6c7 d7b8 a1c3 b8a6 c7b6 a6b8 h8d4 b8d7 b6c7 d7c5 d4g1 c5e6 c7b6 a8b8
c3e5 b8c8 e5b2 e6d8 g1h2 d8e6 b2f6 c8d7 f6e5 e6d8
info depth 19 seldepth 30 score cp -126 nodes 120563 nps 920328 time 131 multipv
1 pv b8d7 b6c7 d7b8 a1c3 b8a6 c7b6 a6b8 h8d4 b8d7 b6c7 d7b8 d4f2 b8a6 c7c6 a6b8
c6b6 b8d7 b6c7 d7b8 f2d4 b8a6 c7b6 a6c7 b6c6 c7a6 d4f2
info depth 20 seldepth 30 score cp -126 nodes 127461 nps 923630 time 138 multipv
1 pv b8d7 b6c7 d7b8 a1c3 b8a6 c7b6 a6b8 h8d4 b8d7 b6c7 d7b8 d4f2 b8a6 c7c6 a6b8
c6b6 b8d7 b6c7 d7b8 f2d4 b8a6 c7b6 a6c7 b6c6 c7a6 d4f2
info depth 21 seldepth 32 score cp -136 nodes 263580 nps 1237464 time 213 multip
v 1 pv b8d7 b6c7 d7b8 h8d4 b8a6 c7b6 a6b4 d4c5 b4d5 b6c6 d5c7 c5d6 c7a6 c6b6 a6b
8 d6g3 b8d7 b6c7 d7b8 g3e5 b8a6 c7b6 a6b4 e5d6 b4d5 b6c6 d5b6 d6g3 b6c8 c6c7 c8a
7 a1e5
info depth 22 seldepth 39 score cp -136 nodes 301533 nps 1283119 time 235 multip
v 1 pv b8d7 b6c7 d7b8 h8d4 b8a6 c7b6 a6b4 d4c5 b4d5 b6c6 d5c7 c5d6 c7a6 c6b6 a6b
8 d6a3 b8d7 b6c6 d7b8 c6c7 b8a6 c7b6 a6b8 a1d4 b8d7 b6c7 d7c5 d4e3 c5a6 c7c6 a6c
7 e3b6 c7a6 b6d4 a6c7 a3d6 c7a6 d6a3
info depth 23 seldepth 39 score cp -136 nodes 357994 nps 1335798 time 268 multip
v 1 pv b8d7 b6c7 d7b8 h8d4 b8a6 c7b6 a6b4 d4c5 b4d5 b6c6 d5c7 c5d6 c7a6 c6b6 a6b
8 d6a3 b8d7 b6c6 d7b8 c6c7 b8a6 c7b6 a6b8 a1d4 b8d7 b6c7 d7c5 d4b2 c5d3 b2d4 d3b
4 d4g7 b4a6 c7b6 a6b8 g7a1
info depth 24 seldepth 39 score cp -136 nodes 388662 nps 1363726 time 285 multip
v 1 pv b8d7 b6c7 d7b8 h8d4 b8a6 c7b6 a6b4 d4c5 b4d5 b6c6 d5c7 c5d6 c7a6 c6b6 a6b
8 d6a3 b8d7 b6c6 d7b8 c6c7 b8a6 c7b6 a6b8 a1d4 b8d7 b6c7 d7c5 d4b2 c5d3 b2d4 d3b
4 d4g7 b4a6 c7b6 a6b8 g7a1
info depth 25 seldepth 39 score cp -136 nodes 454388 nps 1415538 time 321 multip
v 1 pv b8d7 b6c7 d7b8 h8d4 b8a6 c7b6 a6b4 d4c5 b4d5 b6c6 d5c7 c5d6 c7a6 c6b6 a6b
8 d6a3 b8d7 b6c6 d7b8 c6c7 b8a6 c7b6 a6b8 a1d4 b8d7 b6c7 d7c5 d4b2 c5d3 b2a1 d3c
5 a1b2
info depth 26 seldepth 39 score cp -136 nodes 582676 nps 1486418 time 392 multip
v 1 pv b8d7 b6c7 d7b8 h8d4 b8a6 c7b6 a6b4 d4c5 b4d5 b6c6 d5c7 c5d6 c7a6 c6b6 a6b
8 d6a3 b8d7 b6c6 d7b8 c6c7 b8a6 c7b6 a6b8 a1d4 b8d7 b6c7 d7c5 c7c6 c5b7 c6d7 b7a
5 d7c7 a5b7 d4b6 b7c5 c7c6 c5a6 b6d4 a6c7 a3d6 c7a6 c6b6 a6b8 d6a3
info depth 27 seldepth 39 score cp -136 nodes 687366 nps 1530881 time 449 multip
v 1 pv b8d7 b6c7 d7b8 h8d4 b8a6 c7b6 a6b4 d4c5 b4d5 b6c6 d5c7 c5d6 c7a6 c6b6 a6b
8 d6a3 b8d7 b6c6 d7b8 c6c7 b8a6 c7b6 a6b8 a1d4 b8d7 b6c6 d7b8 c6c7 b8a6 c7b6 a6b
8 d4b2 b8d7 b6c7 d7c5 b2h8 c5b7 h8d4 b7c5 c7c6 c5b7 c6d7 b7a5 d7c7 a5b7 d4b6 b7c
5 c7c6 c5a6 b6d4 a6c7 a3d6 c7a6 c6b6 a6b8 d6a3
info depth 28 seldepth 39 score cp -136 nodes 888320 nps 1591971 time 558 multip
v 1 pv b8d7 b6c7 d7b8 h8d4 b8a6 c7b6 a6b4 d4c5 b4d5 b6c6 d5c7 c5d6 c7a6 c6b6 a6b
8 d6a3 b8d7 b6c6 d7b8 c6c7 b8a6 c7b6 a6b8 a1d4 b8d7 b6c6 d7b8 c6c7 b8a6 c7b6 a6b
8 d4b2 b8d7 b6c7 d7c5 b2h8 c5b7 h8d4 b7c5 c7c6 c5b7 c6d7 b7a5 d7c7 a5b7 d4b6 b7c
5 c7c6 c5a6 b6d4 a6c7 a3d6 c7a6 c6b6 a6b8 d6a3
info depth 29 seldepth 46 score cp -136 nodes 1023032 nps 1626441 time 629 multi
pv 1 pv b8d7 b6c7 d7b8 h8d4 b8a6 c7b6 a6b4 d4c5 b4d5 b6c6 d5c7 c5d6 c7a6 c6b6 a6
b8 d6a3 b8d7 b6c6 d7b8 c6c7 b8a6 c7b6 a6b8 a1d4 b8d7 b6c6 d7b8 c6c7 b8a6 c7b6 a6
b8 d4b2 b8d7 b6c6 d7b8 c6c7 a8a7 b2d4 a7a8 d4f2 b8a6 c7b6 a6b8 f2d4
info depth 30 seldepth 50 score cp -136 nodes 1252912 nps 1668324 time 751 multi
pv 1 pv b8d7 b6c7 d7b8 h8d4 b8a6 c7b6 a6b4 d4c5 b4d5 b6c6 d5c7 c5d6 c7a6 c6b6 a6
b8 d6a3 b8d7 b6c6 d7b8 c6c7 b8a6 c7b6 a6b8 a1d4 b8d7 b6c6 d7b8 c6c7 b8a6 c7b6 a6
b8 d4b2 b8d7 b6c6 d7b8 c6c7 a8a7 b2d4 a7a8 d4f2 b8a6 c7c6 a6b8 c6b6 b8d7 b6c7 d7
c5 c7c6 c5b7 a3b2 b7a5 c6c7
info depth 31 seldepth 50 score cp -136 nodes 1442988 nps 1697632 time 850 multi
pv 1 pv b8d7 b6c7 d7b8 h8d4 b8a6 c7b6 a6b4 d4c5 b4d5 b6c6 d5c7 c5d6 c7a6 c6b6 a6
b8 d6a3 b8d7 b6c6 d7b8 c6c7 b8a6 c7b6 a6b8 a1d4 b8d7 b6c6 d7b8 c6c7 b8a6 c7b6 a6
b8 d4b2 b8d7 b6c6 d7b8 c6c7 a8a7 b2d4 a7a8 d4f2 b8a6 c7c6 a6b8 c6b6 b8d7 b6c7 d7
c5 c7c6 c5b7 a3b2 b7a5 c6c7
info depth 32 seldepth 50 score cp -136 nodes 1682866 nps 1727788 time 974 multi
pv 1 pv b8d7 b6c7 d7b8 h8d4 b8a6 c7b6 a6b4 d4c5 b4d5 b6c6 d5c7 c5d6 c7a6 c6b6 a6
b8 d6a3 b8d7 b6c6 d7b8 c6c7 b8a6 c7b6 a6b8 a1d4 b8d7 b6c6 d7b8 c6c7 b8a6 c7b6 a6
b8 d4b2 b8d7 b6c6 d7b8 c6c7 a8a7 b2d4 a7a8 d4f2 b8a6 c7c6 a6b8 c6b6 b8d7 b6c7 d7
c5 c7c6
info depth 33 seldepth 50 score cp -136 nodes 2080178 nps 1762862 time 1180 mult
ipv 1 pv b8d7 b6c7 d7b8 h8d4 b8a6 c7b6 a6b4 d4c5 b4d5 b6c6 d5c7 c5d6 c7a6 c6b6 a
6b8 d6a3 b8d7 b6c6 d7b8 c6c7 b8a6 c7b6 a6b8 a1d4 b8d7 b6c6 d7b8 c6c7 b8a6 c7b6 a
6b8 d4b2 b8d7 b6c6 d7b8 c6c7 a8a7 b2d4 a7a8 d4f2 b8a6 c7c6 a6b8 c6c7
info depth 34 seldepth 55 score cp -136 nodes 2701795 nps 1791641 time 1508 mult
ipv 1 pv b8d7 b6c7 d7b8 h8d4 b8a6 c7b6 a6b4 d4c5 b4d5 b6c6 d5c7 c5d6 c7a6 c6b6 a
6b8 d6a3 b8d7 b6c6 d7b8 c6c7 b8a6 c7b6 a6b8 a1d4 b8d7 b6c6 d7b8 c6c7 b8a6 c7b6 a
6b8 d4b2 b8d7 b6c6 d7b8 c6c7 a8a7 b2d4 a7a8 d4f2 b8a6 c7c6 a6b8 c6c7
info depth 35 seldepth 55 score cp -136 nodes 3214587 nps 1813077 time 1773 mult
ipv 1 pv b8d7 b6c7 d7b8 h8d4 b8a6 c7b6 a6b4 d4c5 b4d5 b6c6 d5c7 c5d6 c7a6 c6b6 a
6b8 d6a3 b8d7 b6c6 d7b8 c6c7 b8a6 c7b6 a6b8 a1d4 b8d7 b6c6 d7b8 c6c7 b8a6 c7b6 a
6b8 d4b2 b8d7 b6c6 d7b8 c6c7 a8a7 b2d4 a7a8 c7b6
info depth 36 seldepth 56 score cp -136 nodes 4706280 nps 1856520 time 2535 mult
ipv 1 pv b8d7 b6c7 d7b8 h8d4 b8a6 c7b6 a6b4 d4c5 b4d5 b6c6 d5c7 c5d6 c7a6 c6b6 a
6b8 d6a3 b8d7 b6c6 d7b8 c6c7 b8a6 c7b6 a6b8 a1d4 b8d7 b6c6 d7b8 c6c7 b8a6 c7b6 a
6b8 d4b2 b8d7 b6c6 d7b8 c6c7 a8a7 b2d4 a7a8 c7b6
info depth 37 seldepth 56 score cp -136 nodes 5367303 nps 1863646 time 2880 mult
ipv 1 pv b8d7 b6c7 d7b8 h8d4 b8a6 c7b6 a6b4 d4c5 b4d5 b6c6 d5c7 c5d6 c7a6 c6b6 a
6b8 d6a3 b8d7 b6c6 d7b8 c6c7 b8a6 c7b6 a6b8 a1d4 b8d7 b6c6 d7b8 c6c7 b8a6 c7b6 a
6b8 d4b2 b8d7 b6c6 d7b8 c6c7 a8a7 b2d4 a7a8 c7b6
info depth 38 currmove b8c6 currmovenumber 2
info depth 38 currmove b8a6 currmovenumber 3
info depth 38 seldepth 60 score cp -136 nodes 6270692 nps 1866277 time 3360 mult
ipv 1 pv b8d7 b6c7 d7b8 h8d4 b8a6 c7b6 a6b4 d4c5 b4d5 b6c6 d5c7 c5d6 c7a6 c6b6 a
6b8 d6a3 b8d7 b6c6 d7b8 c6c7 b8a6 c7b6 a6b8 a1d4 b8d7 b6c6 d7b8 c6c7 b8a6 c7b6 a
6b8 d4b2 b8d7 b6c6 d7b8 c6c7 a8a7 b2d4 a7a8 d4c3 b8a6 c7b6 a6b8 c3e1 b8d7 b6c7 d
7c5 e1f2 c5a6 c7b6 a6b8 a3b2 b8d7 b6c7 d7b8 b2a3 b8a6
info depth 39 currmove b8d7 currmovenumber 1
info depth 39 currmove b8c6 currmovenumber 2
info depth 39 currmove b8a6 currmovenumber 3
info depth 39 seldepth 60 score cp -136 nodes 7274181 nps 1878177 time 3873 mult
ipv 1 pv b8d7 b6c7 d7b8 h8d4 b8a6 c7b6 a6b4 d4c5 b4d5 b6c6 d5c7 c5d6 c7a6 c6b6 a
6b8 d6a3 b8d7 b6c6 d7b8 c6c7 b8a6 c7b6 a6b8 a1d4 b8d7 b6c6 d7b8 c6c7 b8a6 c7b6 a
6b8 d4b2 b8d7 b6c6 d7b8 c6c7 a8a7 b2d4 a7a8 c7b6
info depth 40 currmove b8d7 currmovenumber 1
info depth 40 currmove b8c6 currmovenumber 2
info depth 40 currmove b8a6 currmovenumber 3
info depth 40 seldepth 60 score cp -136 nodes 8127115 nps 1887393 time 4306 mult
ipv 1 pv b8d7 b6c7 d7b8 h8d4 b8a6 c7b6 a6b4 d4c5 b4d5 b6c6 d5c7 c5d6 c7a6 c6b6 a
6b8 d6a3 b8d7 b6c6 d7b8 c6c7 b8a6 c7b6 a6b8 a1d4 b8d7 b6c6 d7b8 c6c7 b8a6 c7b6 a
6b8 d4b2 b8d7 b6c6 d7b8 c6c7 a8a7 b2d4 a7a8 c7b6
info nodes 8127115 time 4307
bestmove b8d7 ponder b6c7


Eelco
Last edited by Eelco de Groot on Sat Jan 11, 2014 11:24 am, edited 1 time in total.
Debugging is twice as hard as writing the code in the first
place. Therefore, if you write the code as cleverly as possible, you
are, by definition, not smart enough to debug it.
-- Brian W. Kernighan
Top
Vinvin
Posts: 5290
Joined: Thu Mar 09, 2006 9:40 am
Full name: Vincent Lejeune

Re: illogical eval from SF? BBKBPK +- 2.00 then BBKBK = 0.00

Post by Vinvin »

Eelco de Groot wrote:
Vinvin wrote:
Eelco de Groot wrote:
zullil wrote:
Eelco de Groot wrote:Talking about wrong bishop-endgame evals, here is a real testgame played yesterday where one Stockfish on move 93 promotes to a bishop leading to a same coloured bishop pair. The eval is +76.79.
2K5/1P6/4n3/5k2/8/8/8/B7 w - - 3 93

This needs fixing. Stockfish doesn't switch to the obvious promotion to a queen until depth 29? :shock: (At least it eventually gets the correct evaluation of mate-in-11.)
...
This is with a proposed bugfix, one thread, 64 Mb hash. Stockfish now thinks it is mate in 12 after 93. b8Q but maybe it is better with more cores.
Is the bugfix covers the case N vs BB(same color) ?
Do SF sees h8=B is draw here ?
[d]kn6/7P/1K6/8/8/8/8/B7 w - -
Hi Vincent,

This is maybe not such a good test, Stockfish sees mate in 5 at depth 15 :)
...
Eelco
Let's say in mutli PV, 6 best lines after 10 seconds ;-)
Top
Joerg Oster
Posts: 974
Joined: Fri Mar 10, 2006 4:29 pm
Location: Germany
Full name: Jörg Oster

Re: illogical eval from SF? BBKBPK +- 2.00 then BBKBK = 0.00

Post by Joerg Oster »

Eelco, you need to test the position with the bishop promotion.

[d]kn5B/8/1K6/8/8/8/8/B7 b - - 0 1

Code: Select all

position fen kn5B/8/1K6/8/8/8/8/B7 b - - 0 1
go depth 10
info depth 1 seldepth 2 score cp 0 nodes 21 nps 21000 time 1 multipv 1 pv b8d7 b6a5
info depth 2 seldepth 3 score cp 0 nodes 67 nps 67000 time 1 multipv 1 pv b8d7 b6a5 d7c5
info depth 3 seldepth 4 score cp 0 nodes 158 nps 79000 time 2 multipv 1 pv b8d7 b6a5 d7c5 h8b2
info depth 4 seldepth 5 score cp 0 nodes 299 nps 99666 time 3 multipv 1 pv b8d7 b6a5 d7c5 h8b2 c5d7
info depth 5 seldepth 6 score cp 0 nodes 471 nps 157000 time 3 multipv 1 pv b8d7 b6a5 d7c5 h8b2 c5d7 b2c1
info depth 6 seldepth 7 score cp 0 nodes 690 nps 172500 time 4 multipv 1 pv b8d7 b6a5 d7c5 h8b2 c5d7 b2c1 d7b8
info depth 7 seldepth 8 score cp 0 nodes 889 nps 177800 time 5 multipv 1 pv b8d7 b6a5 d7c5 h8b2 c5d7 b2c1 d7b8 c1b2
info depth 8 seldepth 9 score cp 0 nodes 1156 nps 192666 time 6 multipv 1 pv b8d7 b6a5 d7c5 h8b2 c5d7 b2c1 d7b8 c1b2 b8d7
info depth 9 seldepth 10 score cp 0 nodes 1526 nps 218000 time 7 multipv 1 pv b8d7 b6a5 d7c5 h8b2 c5d7 b2c1 d7b8 c1b2 b8d7
info depth 10 seldepth 10 score cp 0 nodes 2123 nps 265375 time 8 multipv 1 pv b8d7 b6a5 d7c5 h8b2 c5d7 b2c1 d7b8 c1b2 b8d7
info nodes 2123 time 8
bestmove b8d7 ponder b6a5
If you don't mind, I will do a pull request with my version ...

Edit: Oh, I just realize you did test it, but no draw score :(
Last edited by Joerg Oster on Sat Jan 11, 2014 11:32 am, edited 1 time in total.
Jörg Oster
Top

Return to “Computer Chess Club: General Topics”