Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Discrepancy between score and PV

Author: Normand M. Blais

Date: 04:13:37 07/22/03

Go up one level in this thread


On July 22, 2003 at 04:15:58, Uri Blass wrote:

>On July 21, 2003 at 23:53:48, Normand M. Blais wrote:
>
>>The following position occured in a game between two chess programs (My program
>>had white).
>>
>>[d]r3r1k1/1p3ppp/2n1qn2/p1b1pNB1/3pP3/P2B4/1PP2PPP/R2Q1RK1 w - -
>>
>>After the game, I decided to let the program analyse the position again just for
>>fun. But then I notice that, after finding the correct move, the pv line at
>>depth 6 was ending in a checkmate while the score at that depth was 0.02. I
>>suspect that this is a bug. Any idea how this could happend?
>
>Did your program show the checkmate in the pv or another interface?
>
>In case that another interface gave the checkmate then it is possible that your
>program did not know that the position is a checkmate and evaluated it as 0.02
>
>It is impossible with movei because it use the number of moves to calculate it's
>evaluation so there is no chance that it is going to evaluate mate positions as
>0.02.
>
>Uri

During the game, the program identify a mate as the best line while pondering:

2796271 <first : hint: Qe6
2796271 <first : check opening database
2796281 <first : no pondering move from opening database
2796281 <first : 1 16 0 57 a4
2796301 <first : 2 6 2 1149 a4 Rac8
2796311 <first : 3 30 3 1443 Qf3 Nd7 Rfd1
2796351 <second: 1 -20 0 224 Qd8
2796351 <second: 2 -19 0 714 Qd8 Re1
2796482 <first : 4 20 20 7710 Qe2 Nd7 Bc4 Qg6
2796502 <second: 3 -20 15 13866 Qe6 Bxf6 Qxf6
2796752 <second: 4 -14 40 39583 Qd8 Bxf6 Qxf6 Re1
2797563 <first : 5 21 128 35143 Qf3 Nd7 a4 Rac8 Rad1
2800397 <second: 5 -14 405 370449 Qe6 b3 Be7 f4 exf4
2804593 <second: move Qe6
Interrupting first
2804593 >first : time 24309
otim 19913
2804593 >first : Qe6
2804593 <first : while pondering: Qe6
2804633 <first : 6 31 4 212833 Bc4 Qxc4 Bxf6 Bf8 Qg4 Qxc2 Nxg7 Bxg7 Qxg7#
2806326 <first : user played the predicted move: Qe6
2806326 <first : user                      move: Qe6
2806326 <first : actual position
2806326 <first :
2806326 <first : 8  r . . . r . k .
2806326 <first : 7  . p . . . p p p
2806326 <first : 6  . . n . q n . .
2806326 <first : 5  p . b . p N B .
2806326 <first : 4  . . . p P . . .
2806326 <first : 3  P . . B . . . .
2806326 <first : 2  . P P . . P P P
2806326 <first : 1  R . . Q . R K .
2806326 <first :
2806326 <first :    a b c d e f g h
2806326 <first :
2806326 <first : sending pondering move to winboard
2806326 <first : move Bc4

here is the whole game:

[Event "Computer chess game"]
[Site "NORMAND"]
[Date "2003.07.21"]
[Round "5"]
[White "Alex 1.40.20"]
[Black "CEFAP 0.72"]
[Result "1-0"]
[TimeControl "40/300"]

1. Nf3 d5 2. d4 Nf6 3. e3 e6 4. Bd3 c5 5. O-O Be7 6. Nc3 O-O 7. dxc5 Bxc5
8. Bd2 Bd7 9. Ne5 Nc6 10. Nxd7 Qxd7 11. e4 d4 12. Ne2 e5 13. a3 a5 14. Bg5
Qd6 15. Ng3 Rfe8 16. Nf5 Qe6 17. Bc4 Qxc4 18. Bxf6 Bf8 19. Nxg7 Rec8 20.
Qg4 h6 21. Ne6+ Kh7 22. Nxf8+ Rxf8 23. Qg7#
{White mates} 1-0

Normand




This page took 0 seconds to execute

Last modified: Thu, 15 Apr 21 08:11:13 -0700

Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.