TalkChess.com

ChessGUI awards a loss on time even if an engine lacks only 1 millisec to finish a session. Use only ChessGUI 0.133 to eventually see what I mean.

I hear that Arena is not so strict.
I don't know how strict Winboard or other GUIs are.

I think the ChessGUI approach is fine.
Or should a GUI tolerate a 1000 millisec failure to meet the TC ?

Matthias.

Matthias Gemuh wrote:ChessGUI awards a loss on time even if an engine lacks only 1 millisec to finish a session. Use only ChessGUI 0.133 to eventually see what I mean.

I hear that Arena is not so strict.
I don't know how strict Winboard or other GUIs are.

I think the ChessGUI approach is fine.
Or should a GUI tolerate a 1000 millisec failure to meet the TC ?

Matthias.

That's fine with me and WB does the same. A GUI should never allow overstepping the time, no matter how small the margin.
I have seen logs when Arena even allowed engines to start with negative time for the next move.
There are a few programs(e.g. Monarch UCI) which simply overstep their time by some centi sec(not even milli sec)
and they won't be punished by Arena and e.g. the Lokasoft UCI adapter for ChessTiger, which
sometimes uses 10-70 ms more than allowed, but at least it doesn't happen too frequentely with the latest ChessTiger
after some time management fix from Christophe.

Guenther

Matthias Gemuh wrote:ChessGUI awards a loss on time even if an engine lacks only 1 millisec to finish a session. Use only ChessGUI 0.133 to eventually see what I mean.

I hear that Arena is not so strict.
I don't know how strict Winboard or other GUIs are.

I think the ChessGUI approach is fine.
Or should a GUI tolerate a 1000 millisec failure to meet the TC ?

Matthias.

Yes,just like the Arena GUI,that will save a lot of trouble....

Dr.Wael Deeb wrote:

Matthias Gemuh wrote:ChessGUI awards a loss on time even if an engine lacks only 1 millisec to finish a session. Use only ChessGUI 0.133 to eventually see what I mean.

I hear that Arena is not so strict.
I don't know how strict Winboard or other GUIs are.

I think the ChessGUI approach is fine.
Or should a GUI tolerate a 1000 millisec failure to meet the TC ?

Matthias.

Yes,just like the Arena GUI,that will save a lot of trouble....

Doc,This is not "saving trouble" (What trouble anyway?) but ... cheating...

Best
Fonzy

F. Bluemers wrote:

Dr.Wael Deeb wrote:

Matthias Gemuh wrote:ChessGUI awards a loss on time even if an engine lacks only 1 millisec to finish a session. Use only ChessGUI 0.133 to eventually see what I mean.

I hear that Arena is not so strict.
I don't know how strict Winboard or other GUIs are.

I think the ChessGUI approach is fine.
Or should a GUI tolerate a 1000 millisec failure to meet the TC ?

Matthias.

Yes,just like the Arena GUI,that will save a lot of trouble....

Doc,This is not "saving trouble" (What trouble anyway?) but ... cheating...

Best
Fonzy

I mean to have less games lost on time in our tournaments....come on,it's only a few millseconds after all

Dr.Wael Deeb wrote:

F. Bluemers wrote:

Dr.Wael Deeb wrote:

Matthias Gemuh wrote:ChessGUI awards a loss on time even if an engine lacks only 1 millisec to finish a session. Use only ChessGUI 0.133 to eventually see what I mean.

I hear that Arena is not so strict.
I don't know how strict Winboard or other GUIs are.

I think the ChessGUI approach is fine.
Or should a GUI tolerate a 1000 millisec failure to meet the TC ?

Matthias.

Yes,just like the Arena GUI,that will save a lot of trouble....

Doc,This is not "saving trouble" (What trouble anyway?) but ... cheating...

Best
Fonzy

I mean to have less games lost on time in our tournaments....come on,it's only a few millseconds after all

Don't reward the lazy programmers ,give them hell

Guenther wrote:

Matthias Gemuh wrote:ChessGUI awards a loss on time even if an engine lacks only 1 millisec to finish a session. Use only ChessGUI 0.133 to eventually see what I mean.

I hear that Arena is not so strict.
I don't know how strict Winboard or other GUIs are.

I think the ChessGUI approach is fine.
Or should a GUI tolerate a 1000 millisec failure to meet the TC ?

Matthias.

That's fine with me and WB does the same. A GUI should never allow overstepping the time, no matter how small the margin.
I have seen logs when Arena even allowed engines to start with negative time for the next move.
There are a few programs(e.g. Monarch UCI) which simply overstep their time by some centi sec(not even milli sec)
and they won't be punished by Arena and e.g. the Lokasoft UCI adapter for ChessTiger, which
sometimes uses 10-70 ms more than allowed, but at least it doesn't happen too frequentely with the latest ChessTiger
after some time management fix from Christophe.

Guenther

I am of your opinion.
I saw an engine overstep by 200ms, playing at "40moves/60s repeated".
I will however let the user set this margin in next ChessGUI.

Best,
Matthias.

Dr.Wael Deeb wrote:

Matthias Gemuh wrote:ChessGUI awards a loss on time even if an engine lacks only 1 millisec to finish a session. Use only ChessGUI 0.133 to eventually see what I mean.

I hear that Arena is not so strict.
I don't know how strict Winboard or other GUIs are.

I think the ChessGUI approach is fine.
Or should a GUI tolerate a 1000 millisec failure to meet the TC ?

Matthias.

Yes,just like the Arena GUI,that will save a lot of trouble....

I will let the user set this margin in next ChessGUI.

Best,
Matthias.

Matthias Gemuh wrote:

Dr.Wael Deeb wrote:

Matthias Gemuh wrote:ChessGUI awards a loss on time even if an engine lacks only 1 millisec to finish a session. Use only ChessGUI 0.133 to eventually see what I mean.

I hear that Arena is not so strict.
I don't know how strict Winboard or other GUIs are.

I think the ChessGUI approach is fine.
Or should a GUI tolerate a 1000 millisec failure to meet the TC ?

Matthias.

Yes,just like the Arena GUI,that will save a lot of trouble....

I will let the user set this margin in next ChessGUI.

Best,
Matthias.

Its better if broken engines get fixed
Margins will make things worse

Best
Fonzy

Matthias Gemuh wrote:

Guenther wrote:

Matthias Gemuh wrote:ChessGUI awards a loss on time even if an engine lacks only 1 millisec to finish a session. Use only ChessGUI 0.133 to eventually see what I mean.

I hear that Arena is not so strict.
I don't know how strict Winboard or other GUIs are.

I think the ChessGUI approach is fine.
Or should a GUI tolerate a 1000 millisec failure to meet the TC ?

Matthias.

That's fine with me and WB does the same. A GUI should never allow overstepping the time, no matter how small the margin.
I have seen logs when Arena even allowed engines to start with negative time for the next move.
There are a few programs(e.g. Monarch UCI) which simply overstep their time by some centi sec(not even milli sec)
and they won't be punished by Arena and e.g. the Lokasoft UCI adapter for ChessTiger, which
sometimes uses 10-70 ms more than allowed, but at least it doesn't happen too frequentely with the latest ChessTiger
after some time management fix from Christophe.

Guenther

I am of your opinion.
I saw an engine overstep by 200ms, playing at "40moves/60s repeated".
I will however let the user set this margin in next ChessGUI.

Best,
Matthias.

Maybe, this happened to me.

I developed the program in Arena, I've played hundreds of games even in bullet as 1+1 or 40/1 and I have not had any loss in time.

If someone tells you that he had problems, the first thing I think is the user is doing something wrong (problems with your computer...), rather it is mime problem.

But I see now that the engine has problems of lost in time in chessGUI (and perhaps in winboard) for games bullet.

I would have preferred that Arena not to be tolerant with time, because at the time to develop the engine I could have tuned.

The next version will be tested with chessgui or winboard.

Pedro