Does Wb2UCI get confused by large evaluation scores?

[Alejandro Dubrovsky]

Geschrieben von:/Posted by: Alejandro Dubrovsky at 24 March 2004 16:25:23:

This is an excerpt of a wb2uci.log sent by Claude Dubois. Anyone else had similar experiences? (Wb2UCI gets told to quit, it tells engine to quite, and game is assigned to black, when its mate in two for white). I think it's Fritz interface (Claude?)
position is
1q3Q2/p1k3pp/1pB1p3/3pN1pn/3P4/4P3/1P3PPP/4K2R w K - 0 22

111.500: C> position startpos moves c2c4 c7c6 d2d4 d7d5 c4d5 c6d5 b1c3 g8f6 c1f4 c8f5 d1b3 b7b6 c3b5 b8c6 b5c7 e8d7 c7a8 d8a8 e2e3 e7e6 a2a3 f6h5 f4g5 f7f6 a1c1 f8a3 b3a3 f6g5 f1b5 h8c8 g1f3 d7e8 f3e5 f5c2 c1c2 a8b8 c2c6 c8c6 b5c6 e8d8 a3f8 d8c7
111.500: C> go movestogo 19 wtime 136000 btime 121000
111.500: C< info depth 1 score 1
111.500: S< time 13600
111.500: S< otim 12100
111.500: S< usermove d8c7
111.540: S> 2 9999955 2 2 Qe7+ Kc8 Qd7#
111.540: C< info depth 2 score cp 9999955 time 20 nodes 2 nps 100 pv f8e7 c7c8 e7d7
111.540: S> 3 9999955 2 87 Qe7+ Kc8 Qd7#
111.540: C< info depth 3 score cp 9999955 time 20 nodes 87 nps 4350 pv f8e7 c7c8 e7d7
111.540: S> 4 9999955 2 96 Qe7+ Kc8 Qd7#
111.540: C< info depth 4 score cp 9999955 time 20 nodes 96 nps 4800 pv f8e7 c7c8 e7d7
111.540: S> 5 9999955 3 176 Qe7+ Kc8 Qd7#
111.540: C< info depth 5 score cp 9999955 time 30 nodes 176 nps 5866 pv f8e7 c7c8 e7d7
111.540: S> move f8e7
111.540: C< bestmove f8e7 ponder c7c8
112.091: C> quit
112.091: S< quit



SmallPotato homepage

[Alejandro Dubrovsky]

Geschrieben von:/Posted by: Alejandro Dubrovsky at 24 March 2004 16:26:51:
Als Antwort auf:/In reply to: Does Wb2UCI get confused by large evaluation scores? geschrieben von:/posted by: Alejandro Dubrovsky at 24 March 2004 16:25:23:

Since wb2uci is obviously doing the right thing here

This is an excerpt of a wb2uci.log sent by Claude Dubois. Anyone else had similar experiences? (Wb2UCI gets told to quit, it tells engine to quite, and game is assigned to black, when its mate in two for white). I think it's Fritz interface (Claude?)
position is
1q3Q2/p1k3pp/1pB1p3/3pN1pn/3P4/4P3/1P3PPP/4K2R w K - 0 22

111.500: C> position startpos moves c2c4 c7c6 d2d4 d7d5 c4d5 c6d5 b1c3 g8f6 c1f4 c8f5 d1b3 b7b6 c3b5 b8c6 b5c7 e8d7 c7a8 d8a8 e2e3 e7e6 a2a3 f6h5 f4g5 f7f6 a1c1 f8a3 b3a3 f6g5 f1b5 h8c8 g1f3 d7e8 f3e5 f5c2 c1c2 a8b8 c2c6 c8c6 b5c6 e8d8 a3f8 d8c7
111.500: C> go movestogo 19 wtime 136000 btime 121000
111.500: C< info depth 1 score 1
111.500: S< time 13600
111.500: S< otim 12100
111.500: S< usermove d8c7
111.540: S> 2 9999955 2 2 Qe7+ Kc8 Qd7#
111.540: C< info depth 2 score cp 9999955 time 20 nodes 2 nps 100 pv f8e7 c7c8 e7d7
111.540: S> 3 9999955 2 87 Qe7+ Kc8 Qd7#
111.540: C< info depth 3 score cp 9999955 time 20 nodes 87 nps 4350 pv f8e7 c7c8 e7d7
111.540: S> 4 9999955 2 96 Qe7+ Kc8 Qd7#
111.540: C< info depth 4 score cp 9999955 time 20 nodes 96 nps 4800 pv f8e7 c7c8 e7d7
111.540: S> 5 9999955 3 176 Qe7+ Kc8 Qd7#
111.540: C< info depth 5 score cp 9999955 time 30 nodes 176 nps 5866 pv f8e7 c7c8 e7d7
111.540: S> move f8e7
111.540: C< bestmove f8e7 ponder c7c8
112.091: C> quit
112.091: S< quit

SmallPotato homepage

[Claude Dubois]

Geschrieben von:/Posted by: Claude Dubois at 24 March 2004 17:11:20:
Als Antwort auf:/In reply to: Question should be, do some interfaces get confused by large evaluation scores. geschrieben von:/posted by: Alejandro Dubrovsky at 24 March 2004 16:26:51:

Hello Alejandro,
You can see my answear to your post under le Système du Suisse : Round 4
I think that MatScore solve the problem.
Best wishes
Claude


Le Système du Suisse

[Peter Fendrich]

Geschrieben von:/Posted by: Peter Fendrich at 24 March 2004 17:12:32:
Als Antwort auf:/In reply to: Question should be, do some interfaces get confused by large evaluation scores. geschrieben von:/posted by: Alejandro Dubrovsky at 24 March 2004 16:26:51:

The uci protocol excpects score values scaled to centipawns. That's what cp stands for.
Your very high values is probably a problem for UCI GUI's
/Peter

Since wb2uci is obviously doing the right thing here

This is an excerpt of a wb2uci.log sent by Claude Dubois. Anyone else had similar experiences? (Wb2UCI gets told to quit, it tells engine to quite, and game is assigned to black, when its mate in two for white). I think it's Fritz interface (Claude?)
position is
1q3Q2/p1k3pp/1pB1p3/3pN1pn/3P4/4P3/1P3PPP/4K2R w K - 0 22

111.500: C> position startpos moves c2c4 c7c6 d2d4 d7d5 c4d5 c6d5 b1c3 g8f6 c1f4 c8f5 d1b3 b7b6 c3b5 b8c6 b5c7 e8d7 c7a8 d8a8 e2e3 e7e6 a2a3 f6h5 f4g5 f7f6 a1c1 f8a3 b3a3 f6g5 f1b5 h8c8 g1f3 d7e8 f3e5 f5c2 c1c2 a8b8 c2c6 c8c6 b5c6 e8d8 a3f8 d8c7
111.500: C> go movestogo 19 wtime 136000 btime 121000
111.500: C< info depth 1 score 1
111.500: S< time 13600
111.500: S< otim 12100
111.500: S< usermove d8c7
111.540: S> 2 9999955 2 2 Qe7+ Kc8 Qd7#
111.540: C< info depth 2 score cp 9999955 time 20 nodes 2 nps 100 pv f8e7 c7c8 e7d7
111.540: S> 3 9999955 2 87 Qe7+ Kc8 Qd7#
111.540: C< info depth 3 score cp 9999955 time 20 nodes 87 nps 4350 pv f8e7 c7c8 e7d7
111.540: S> 4 9999955 2 96 Qe7+ Kc8 Qd7#
111.540: C< info depth 4 score cp 9999955 time 20 nodes 96 nps 4800 pv f8e7 c7c8 e7d7
111.540: S> 5 9999955 3 176 Qe7+ Kc8 Qd7#
111.540: C< info depth 5 score cp 9999955 time 30 nodes 176 nps 5866 pv f8e7 c7c8 e7d7
111.540: S> move f8e7
111.540: C< bestmove f8e7 ponder c7c8
112.091: C> quit
112.091: S< quit

[Alejandro Dubrovsky]

Geschrieben von:/Posted by: Alejandro Dubrovsky at 24 March 2004 18:56:03:
Als Antwort auf:/In reply to: Re: Question should be, do some interfaces get confused by large evaluation scores. geschrieben von:/posted by: Peter Fendrich at 24 March 2004 17:12:32:

The uci protocol excpects score values scaled to centipawns. That's what cp stands for.
Your very high values is probably a problem for UCI GUI's
/Peter

yes, you are right. I just read the UCI protocol for the first time. It wants 'mate' for mate (makes sense i suppose
Thank for the heads up.



SmallPotato homepage