ChessGUI 0.136 is out

[Matthias Gemuh]

ChessGUI 0.136 is out

- if playing tournament with GUI book, check whether book is used in at least one game. If not used, manual settings update should help.

- if created on starting/resuming tournament, Excluded_Engines.txt and Included_Engines.txt in GUI_DATA folder may be interesting to look at later.

Matthias.

[pedrox]

Hi Matthias,

I have a version now supports ChessGUI and without bugs in the normal variant and FRC.

I've done the following:

1. Correcting the loss of time in the last move of time control. Now I can play fast as 40 moves in 15 seconds.

2. In the 0135 version of ChessGUI I see that DanaSah send Illegal a8a8 move.

In reality this is not an illegal move, this is by the "result" command xboard.

Code: Select all

=========================================================================================================================
Eng1 &#40;danasah416&#41; <- 4 120 0 2780 a2a8  b8d8  c3d4  d8e8  d4b6  d3b3
Eng1 &#40;danasah416&#41; <- 5 129 1 7248 a2a8  d3d8  a8b8  d8b8  a1b1  e4c5
Eng1 &#40;danasah416&#41; <- 6 119 3 17111 a2a8  b8a8  a1a8  f8f7  c3d4  g7h7  f1e1  d3b3  a8b7
Eng1 &#40;danasah416&#41; <- 7 106 3 24169 a2a8  b8a8  a1a8  f8f7  c3f6  e7f6  a8h8  d3b3
Eng1 &#40;danasah416&#41; <- 8 116 6 47308 a2a8  b8a8  a1a8  f8f7  a8c8  g7g8  c8f5  g8g5  f5h7  h8g7
Eng1 &#40;danasah416&#41; <- 9 136 26 185564 a2a8  d3d8  a8b8  d8b8  a1a4  e4d6  a4d4  b8c7  d4d5  e7e5
Eng1 &#40;danasah416&#41; <- Pipe01Time 0000000015143546
Exact move time for danasah416 
Eng1 &#40;danasah416&#41; <- move a2a8
Cancelling GameMoveTimerEvent&#40;) for danasah416 at 0000000015143546.
 ==========================================================================================================================

So far everything is correct. But now DanaSah inexplicably sends a result of stalemate when it is the turn of the another engine.

Code: Select all

 ==========================================================================================================================
Eng1 &#40;danasah416&#41; <- 1/2-1/2 &#123;Stalemate&#125;
Eng2 &#40;Hermann 2.4&#41; <- info depth 1 seldepth 2 time 0 nodes 8 pv d3d8 a8b8 d8b8 a1a4 e4c5 score cp -25 hashfull 0 nps 8000
Eng2 &#40;Hermann 2.4&#41; <- Pipe02Time 0000000015143875
Exact move time for Hermann 2.4 
Eng2 &#40;Hermann 2.4&#41; <- info depth 8 seldepth 25 time 281 nodes 114380 hashfull 13 nps 407046 score cp -28
Eng2 &#40;Hermann 2.4&#41; <- bestmove d3d8 ponder a8b8
Cancelling GameMoveTimerEvent&#40;) for Hermann 2.4 at 0000000015143875.
 ==========================================================================================================================

This is very strange because she was giving the correct PV, impossible to understand why the engine sends the command.

I know that ChessGUI ignores the command result, but when DanaSah sends this command stops playing.

Code: Select all

==========================================================================================================================
Eng1 &#40;danasah416&#41; <- Pipe01Time 0000000015143921
Exact move time for danasah416 
Eng1 &#40;danasah416&#41; <- Illegal move&#58; d3d8
Eng1 &#40;danasah416&#41; <- 1/2-1/2 &#123;Stalemate&#125;
Eng1 &#40;danasah416&#41; <- move a8a8
Cancelling GameMoveTimerEvent&#40;) for danasah416 at 0000000015143937.
 ========================================================================================================================== 

The solution is not send the command result when DanaSah is playing in ChessGUI and then everything is OK, but I think that maybe you can have an problem in ChessGUI (FRC), the engines with problems are always xboard/ winboard not UCI.

3. I also found because the program crash when playing in FRC ChessGUI, I have a problem with the key hash and castling in FRC.

Here is a difference between Arena and ChessGUI, if the problem occurs in the Arena GUI the engine simply does not crash and lost in time. If the problem occurs in ChessGUI the engine crash.

I have changed the way to calculate the key (a bit slower) and now the problem does not occur.

If you or Ray would like to try the program is:

http://danasah.googlepages.com/danasah397c.zip

Best,

Pedro

[Matthias Gemuh]

Hi Pedro,
thanks for the new DanaSah.
There is a stalemate problem somewhere, either in DanaSah or ChessGUI.

If ChessGUI had detected a stalemate, the GUI would have correctly ended the game without telling Hermann to continue playing.
Do you have pgn of the game ?

As for DanaSah sending result during opponent's time, that does not happen
ChessGUI output the result string during opponent's time because string came after move and is then just ignored as a meaningless comment that needs no immediate attention. Only a resign is process after move.

Best,
Matthias.

[Matthias Gemuh]

Here is a difference between Arena and ChessGUI, if the problem occurs in the Arena GUI the engine simply does not crash and lost in time. If the problem occurs in ChessGUI the engine crash.

Best,

Pedro

That is a tough one. I have no clue.
Maybe somebody in this strong forum can help ?

Best,
Matthias.

[Matthias Gemuh]

If ChessGUI had detected a stalemate, the GUI would have correctly ended the game without telling Hermann to continue playing.
Do you have pgn of the game ?

As for DanaSah sending result during opponent's time, that does not happen
ChessGUI output the result string during opponent's time because string came after move and is then just ignored as a meaningless comment that needs no immediate attention. Only a resign is process after move.

Best,
Matthias.

Maybe ChessGUI should handle wrong result claims differently.
Now engine is allowed to play on and lose on time.
Maybe it is better to penalize wrong claim with immediate loss instead.
I shall correct this in next ChessGUI.

Matthias.

[Matthias Gemuh]

2. In the 0135 version of ChessGUI I see that DanaSah send Illegal a8a8 move.

Best,

Pedro

a8a8, a1a1 usually means engine returned without searching, move = 0.
DanaSah did not search because it considered game had ended, I think.

Matthias.

[Matthias Gemuh]

ChessGUI 0.136 is out

- if playing tournament with GUI book, check whether book is used in at least one game. If not used, manual settings update should help.

- if created on starting/resuming tournament, Excluded_Engines.txt and Included_Engines.txt in GUI_DATA folder may be interesting to look at later.

Matthias.

ChessGUI 0.137 is out

A bug was found and fixed.
It affected only variants, not normal chess.
If you have not seen ChessGUI terminate a game by rejecting a legal castle, the bug has not triggered for you.

BTW, I have updated the castling routine for variants in my chess engines.
Because nothing has changed for normal chess, version numbers remain unchanged.

Matthias.

[Graham Banks]

Hi Matthias,

I can't get Waxman 2008 to use the GUI book when it has White. I have tried installing as both a WB engine and as a UCI engine (through wb2uci).
Could you please take a look for me?

Cheers, Graham.

[Matthias Gemuh]

Hi Matthias,

I can't get Waxman 2008 to use the GUI book when it has White. I have tried installing as both a WB engine and as a UCI engine (through wb2uci).
Could you please take a look for me?

Cheers, Graham.

Hi Graham,

I shall look, but do you know whether Waxman 2008 uses GUI book correctly in some other GUI ?
ChessGUI 0.137 and 0.138 have redefined many options.
I hope no negative side effects have crept in.

Best,
Matthias.

[Graham Banks]

Hi Matthias,

I can't get Waxman 2008 to use the GUI book when it has White. I have tried installing as both a WB engine and as a UCI engine (through wb2uci).
Could you please take a look for me?

Cheers, Graham.

Hi Graham,

I shall look, but do you know whether Waxman 2008 uses GUI book correctly in some other GUI ?
ChessGUI 0.137 and 0.138 have redefined many options.
I hope no negative side effects have crept in.

Best,
Matthias.

I had Waxman 2008 working okay under the Fritz 11 GUI.