Resign bug in DanChess 1.04? (debug EngineX vs. DanChess)

[Günther Simon]

Geschrieben von: / Posted by: Günther Simon at 17 January 2004 15:24:44:

I set up a test match between EngineX and DanChess (20 games).
After 2 games the score was 1:1, when DanChess resigned
as White in the second game, but after this it resigned all
other games in move 1?? (Winboard 4.27 of course)
In the posted example part of the debug one can see, that DanChess
somehow got out of sync and sent a resign and a move at the same
time in game 3 and all other games!.
What happened?
Günther
1674790 <first : move e3e2
Interrupting second
1674790 >second: time 24435
otim 16439
1674790 >second: e3e2
1674840 <second: 5 -927 0 6552 1. Nd3? Qe4+ 2. Kh2 Qxd3 3. Rh1
1674840 <second: 5 -955 5 15595 1. Nd3 Qe4+ 2. Kh2 Qxd3 3. Rc1
1674840 <second: 6 -992 5 17720 1. Nd3? Qe4+ 2. Kh2 Qxd3 3. Rf1
1674900 <second: 6 -1344 11 27249 1. Nd3 Qe4+ 2. Kh2 Qxd3 3. Re1 Rxh4+
1675280 <second: 6 -1315 49 114059 1. Ne6 e1=q 2. Rxe1 Qxe1+ 3. Kg2 Qxe6
1675340 <second: 7 -1350 55 133814 1. Ne6? Qxe6 2. a4 e1=q 3. Rxe1 Qxe1+ 4. Kh2
1675720 <second: 7 -1350 88 227574 1. Ne6! Qd5+ 2. Kh2 Ra8
1675890 <second: 7 -1349 104 269663 1. Ne6 e1=q 2. Rxe1 Qxe1+ 3. Kg2 Qxe6 4. a4
1676110 <second: 8 -1384 126 328014 1. Ne6? e1=q 2. Rxe1 Qxe1+ 3. Kg2 Rb8 4. a4 Qxe6
1684460 <second: 8 -9990 967 3309774 1. Ne6 e1=q 2. Nf4 Rxh4+ 3. gxh4 Qxh4+ 4. Kg2 Qe4+
1689890 <second: 8 -9988 1510 5207676 1. Nd7 Rxh4+ 2. gxh4 Qe4+ 3. Kh2 Qf4+ 4. Kh1 Qf3+
1691210 <second: 8 -1885 1642 5636809 1. Nd3 Qe3 2. Rc1 Qxc1+ 3. Kh2 Qd2 4. Nf2 e1=q
1692370 <second: 9 -1885 1758 5980001 1. Nd3 Qe3 2. Rc1 Qxc1+ 3. Kh2 Qd2 4. Nc1 e1=q+ 5. Ne2
1692370 <second: resign
GameEnds(32, White resigns, 1)
Interrupting first
1692370 >first : result 0-1 {White resigns}
Interrupting second
1692370 >second: result 0-1 {White resigns}
1692370 >first : force
1692370 >second: force
Reset(0, 1) from gameMode 8
recognized 'normal' (-1) as variant normal
GameEnds(0, (null), 2)
1702420 >first : new
random
1702420 >first : level 40 5 0
1702420 >first : post
1702420 >first : hard
1702420 >first : easy
1702420 >first : force
1702420 >second: new
random
1702420 >second: level 40 5 0
1702420 >second: post
1702420 >second: hard
1702420 >second: easy
1702420 >second: force
1702420 >first : computer
1702420 >second: computer
1702420 >first : time 30000
otim 30000
1702420 >first : go
1702420 <first : Error (unknown command): computer
1702420 <first :  1    -3      0         1 h2h3 
1702420 <first :  1     0      0         2 h2h4 
1702420 <first :  1     2      0         3 g2g3 
1702420 <first :  1     3      0         4 g2g4 
1702420 <first :  1     4      0         6 f2f4 
1702420 <first :  1    20      0         7 e2e3 
1702420 <first :  1    29      0         8 e2e3 
1702420 <first :  1    42      0        15 e2e4 
1702420 <first :  1    47      0        17 d2d4 
1702420 <first :  1    63      0        25 g1f3 
1702420 <first :  2     0      0        97 g1f3 g8f6 
1702420 <first :  3     0      1       982 g1f3 b8c6 b1c3 g8f6 
1702420 <first :  3    20      1      1078 e2e3 
1702420 <second: Error (unknown command): random
1702420 <second: Error (unknown command): computer
1702470 <first :  3    19      2      1266 e2e3 e7e6 
1702470 <first :  3    38      3      1676 e2e4 b8c6 g1f3 
1702470 <first :  3    47      3      1888 d2d4 g8f6 b1c3 
1702470 <first :  3    51      4      2280 b1c3 b8c6 d2d4 
1702470 <first :  4    35      6      3551 b1c3 b8c6 e2e4 g8f6 g1f3 
1702470 <first :  5    35      7      4105 b1c3 b8c6 e2e4 g8f6 
1702800 <first :  6    21     40     39635 b1c3 g8f6 g1f3 d7d5 d2d4 b8d7 
1703240 <first :  7    17     82     87386 b1c3 d7d5 e2e4 d5e4 c3e4 g8f6 f1d3 
1703570 <first :  7    27    114    117369 e2e3 b8c6 g1f3 g8f6 f1d3 d7d5 e1g1 
1707140 <first :  8     0    473    479039 e2e3 g8f6 g1f3 e7e6 f1d3 f8d6 e1g1 e8g8 
1709940 <first :  9    20    754    719892 e2e3 
1711700 <first :  9    29    928    892993 e2e3 g8f6 b1c3 e7e6 f1d3 f8d6 g1e2 e8g8 e1g1 
1716260 <first :  9    36   1382   1365847 e2e4 b8c6 g1f3 d7d5 e4d5 d8d5 f1e2 g8f6 e1g1 
1735920 <first : move e2e4
1735920 >second: time 30000
otim 26650
1735920 >second: e2e4
1735920 >second: black
1735920 >second: go
1736800 <second: resign
GameEnds(31, Black resigns, 1)
Interrupting first
1736800 >first : result 1-0 {Black resigns}
Interrupting second
1736800 >second: result 1-0 {Black resigns}
1736800 >first : force
1736800 >second: force
1736800 <second: move b8c6
Undoing extra move from second, gameMode 8
1736800 >second: undo
Reset(0, 1) from gameMode 8
recognized 'normal' (-1) as variant normal
GameEnds(0, (null), 2)
1746850 >first : new
random
1746850 >first : level 40 5 0
1746850 >first : post
1746850 >first : hard
1746850 >first : easy
1746850 >first : force
1746850 >second: new
random
1746850 >second: level 40 5 0
1746850 >second: post
1746850 >second: hard
1746850 >second: easy
1746850 >second: force
1746850 >first : computer
1746850 >second: computer
1746850 >second: black
1746850 >second: time 30000
otim 30000
1746850 >second: white
1746850 >second: go
1746850 <first : Error (unknown command): computer
1746850 <second: Error (unknown command): random
1746850 <second: Error (unknown command): computer
1746850 <second: resign
GameEnds(32, White resigns, 1)
1746850 >first : result 0-1 {White resigns}
Interrupting second
1746850 >second: result 0-1 {White resigns}
1746850 >first : force
1746850 >second: force
1746850 <second: move d2d4

[Volker Pittlik]

Geschrieben von: / Posted by: Volker Pittlik at 17 January 2004 15:37:58:
Als Antwort auf: / In reply to: Resign bug in DanChess 1.04? (debug EngineX vs. DanChess) geschrieben von: / posted by: Günther Simon at 17 January 2004 15:24:44:

I set up a test match between EngineX and DanChess (20 games).
After 2 games the score was 1:1, when DanChess resigned
as White in the second game, but after this it resigned all
other games in move 1?? (Winboard 4.27 of course)
In the posted example part of the debug one can see, that DanChess
somehow got out of sync and sent a resign and a move at the same
time in game 3 and all other games!.
What happened?

Did you play with /-reuse /reuse2?

[Günther Simon]

Geschrieben von: / Posted by: Günther Simon at 17 January 2004 15:41:29:
Als Antwort auf: / In reply to: Re: Resign bug in DanChess 1.04? (debug EngineX vs. DanChess) geschrieben von: / posted by: Volker Pittlik at 17 January 2004 15:37:58:

I set up a test match between EngineX and DanChess (20 games).
After 2 games the score was 1:1, when DanChess resigned
as White in the second game, but after this it resigned all
other games in move 1?? (Winboard 4.27 of course)
In the posted example part of the debug one can see, that DanChess
somehow got out of sync and sent a resign and a move at the same
time in game 3 and all other games!.
What happened?

Did you play with /-reuse /reuse2?

No Volker. Do you think DanChess needs this parameter?
I will give it a try.
Thanks,
Günther

[Günther Simon]

Geschrieben von: / Posted by: Günther Simon at 17 January 2004 16:19:47:
Als Antwort auf: / In reply to: Re: Resign bug in DanChess 1.04? (debug EngineX vs. DanChess) geschrieben von: / posted by: Günther Simon at 17 January 2004 15:41:29:

I set up a test match between EngineX and DanChess (20 games).
After 2 games the score was 1:1, when DanChess resigned
as White in the second game, but after this it resigned all
other games in move 1?? (Winboard 4.27 of course)
In the posted example part of the debug one can see, that DanChess
somehow got out of sync and sent a resign and a move at the same
time in game 3 and all other games!.
What happened?

Did you play with /-reuse /reuse2?

No Volker. Do you think DanChess needs this parameter?
I will give it a try.
Thanks,
Günther

Hi Volker,
I tried it now with 'xreuse' and that made it even worser
already after the first change of colour!
DanChess lost 16 seconds! on its clock for restarting now
in game 2
Günther
943840 >second: time 30000
otim 30000
943840 >second: white
943840 >second: go
943840 <first : Error (unknown command): computer
960590 <second: 
960590 <second: WHITE<1>: 
960590 <second: feature setboard=1 analyze=1 draw=0 done=1 myname="DANCHESS 1.04"
Interrupting second
960590 >second: accepted setboard
960590 >second: accepted analyze
960590 >second: accepted draw
960590 >second: accepted done
960590 >second: accepted myname
960590 <second: Error (unknown command): random
960590 <second: Error (unknown command): computer
960590 <second: move d2d4

[Daniel]

Geschrieben von: / Posted by: Daniel at 17 January 2004 17:42:32:
Als Antwort auf: / In reply to: New problem for DanChess with xreuse! (debug EngineX vs. DanChess) geschrieben von: / posted by: Günther Simon at 17 January 2004 16:19:47:

Hi Gunther
I must have made some very bad mistake in the resign.
I will fix it soon and send it to you as soon as possible.
But if it can't wait , i will send you a non-resigning version of
danchess tommorrow, with the resign value set to -INFINITY .Sorry for the inconvinience.Damn,all versions are buggy
By the way,i don't keep previous versions of my source code,so I can't send
you the same exact Danchess 1.04 with only the bug fixed. I have added some passed pawn eval and improved endgame knowledge. I hope it's okay.
regards
Daniel
PS: Thank god Leo's 6th division restarts the engine after each game.

[Günther Simon]

Geschrieben von: / Posted by: Günther Simon at 17 January 2004 18:48:44:
Als Antwort auf: / In reply to: Re: New problem for DanChess with xreuse! (debug EngineX vs. DanChess) geschrieben von: / posted by: Daniel at 17 January 2004 17:42:32:

Hi Gunther
I must have made some very bad mistake in the resign.
I will fix it soon and send it to you as soon as possible.
But if it can't wait , i will send you a non-resigning version of
danchess tommorrow, with the resign value set to -INFINITY .Sorry for the inconvinience.Damn,all versions are buggy
By the way,i don't keep previous versions of my source code,so I can't send
you the same exact Danchess 1.04 with only the bug fixed. I have added some passed pawn eval and improved endgame knowledge. I hope it's okay.
regards
Daniel
PS: Thank god Leo's 6th division restarts the engine after each game.

Hi Daniel,
It doesn't hurry and it were just test games. I won't have found
this bug at all, if I had not played without WBTM this time
(In Joris WBTM all games are started as a new game IIRC - haven't
done a single match there so far though)

Best wishes,
Günther