Re: WBEC Ridderkerk new results.

[Leo Dijksman]

Geschrieben von:/Posted by: Leo Dijksman at 28 May 2004 17:21:07:

Qualify matches:
Fruit = 2nd Division
Amateur 2.80 1½01 2.5/4
Fruit X05/21 0½10 1.5/4
------------------------------------
Fruit X05/21 011½ 2.5/4
The Baron 1.3.0b3 100½ 1.5/4
------------------------------------
Fruit X05/21 11 2.0/2
Arasan 7.4 00 0.0/2 = match stopped because the neccesary 6 points are scored already.
------------------------------------
Qualified!
Playing 1st Division matches now, scored 4,5/10 sofar, another 1,5/2 needed (vs Pepito).

Qualify matches:
Spike = 3rd Division
Spike 0.3b 101½ 2.5/4
Butcher 1.42c 010½ 1.5/4
------------------------------------
NullMover 0.24d 1½10 2.5/4
Spike 0.3b 0½01 1.5/4
------------------------------------
Spike 0.3b ½½11 3.0/4
Averno 0.70a ½½00 1.0/4
------------------------------------
Qualified!
Playing 2nd Division matches now, scored 2/4 sofar, another 4/8 needed.
Also Naum will start today!

Crosstables and pgn on the WBEC homepage.
Enginelist, latest updates/new engines:
28/05/2004:
Updated: none found
Best wishes,
Leo.


WBEC Ridderkerk homepage.

[Fabien Letouzey]

Geschrieben von:/Posted by: Fabien Letouzey at 28 May 2004 17:27:39:
Als Antwort auf:/In reply to: Re: WBEC Ridderkerk new results. geschrieben von:/posted by: Leo Dijksman at 28 May 2004 17:21:07:

Playing 1st Division matches now, scored 4,5/10 sofar, another 1,5/2 needed (vs Pepito).

Oh well
Not as bad as I expected anyway.
Fabien.

[Volker Böhm]

Geschrieben von:/Posted by: Volker Böhm at 28. May 2004 18:36:
Als Antwort auf:/In reply to: Re: WBEC Ridderkerk new results. geschrieben von:/posted by: Fabien Letouzey at 28 May 2004 17:27:39:

Playing 1st Division matches now, scored 4,5/10 sofar, another 1,5/2 needed (vs Pepito).

Oh well
Not as bad as I expected anyway.
Fabien.

Gratulation! You have built an excellent program.
Greetings Volker

[Norm Pollock]

Geschrieben von:/Posted by: Norm Pollock at 28 May 2004 18:49:18:
Als Antwort auf:/In reply to: Re: WBEC Ridderkerk new results. geschrieben von:/posted by: Fabien Letouzey at 28 May 2004 17:27:39:

Why is Fruit playing Pepito? That seems to be wrong.
I thought the new engine plays 7-8-9 of the Division in order to qualify. Pepito finished #10 in Div 1. The 7-8-9 in the recent Div 1 are Abrok, SOS and Kaissa. When Fruit played for Div 2, it did indeed play 7-8-9, which are Amateur, TheBaron and Arasan.
- Norm

[Fabien Letouzey]

Geschrieben von:/Posted by: Fabien Letouzey at 28 May 2004 18:55:36:
Als Antwort auf:/In reply to: Re: WBEC Ridderkerk new results. geschrieben von:/posted by: Norm Pollock at 28 May 2004 18:49:18:

Why is Fruit playing Pepito? That seems to be wrong.
I thought the new engine plays 7-8-9 of the Division in order to qualify. Pepito finished #10 in Div 1. The 7-8-9 in the recent Div 1 are Abrok, SOS and Kaissa. When Fruit played for Div 2, it did indeed play 7-8-9, which are Amateur, TheBaron and Arasan.
- Norm

Why do you ask me?
Note that I would prefer to play Abrok, as I have Pepito on Linux.
Fabien.

[Norm Pollock]

Geschrieben von:/Posted by: Norm Pollock at 28 May 2004 19:00:04:
Als Antwort auf:/In reply to: Re: WBEC Ridderkerk new results. geschrieben von:/posted by: Fabien Letouzey at 28 May 2004 18:55:36:

Why do you ask me?
Note that I would prefer to play Abrok, as I have Pepito on Linux.
Fabien.

I'm asking Leo.
-Norm

[Leo Dijksman]

Geschrieben von:/Posted by: Leo Dijksman at 28 May 2004 19:19:16:
Als Antwort auf:/In reply to: Re: WBEC Ridderkerk new results. geschrieben von:/posted by: Fabien Letouzey at 28 May 2004 18:55:36:

Why is Fruit playing Pepito? That seems to be wrong.
I thought the new engine plays 7-8-9 of the Division in order to qualify. Pepito finished #10 in Div 1. The 7-8-9 in the recent Div 1 are Abrok, SOS and Kaissa. When Fruit played for Div 2, it did indeed play 7-8-9, which are Amateur, TheBaron and Arasan.
- Norm

Why do you ask me?
Note that I would prefer to play Abrok, as I have Pepito on Linux.
Fabien.

Fruit leads with 3-0 vs Pepito, still a 1/2 point needed from the last game
The "problem" is the 0,5-3,5 lose vs Kaissa 1.7!
Best wishes,
Leo.
P.S.Are you still online?



WBEC Ridderkerk homepage.

[Leo Dijksman]

Geschrieben von:/Posted by: Leo Dijksman at 28 May 2004 19:25:05:
Als Antwort auf:/In reply to: Re: WBEC Ridderkerk new results. geschrieben von:/posted by: Norm Pollock at 28 May 2004 19:00:04:

Why do you ask me?
Note that I would prefer to play Abrok, as I have Pepito on Linux.
Fabien.

I'm asking Leo.
-Norm

Why me? )
7,8 and 9 of division 2 = 25,26 and 27 of the total ranking list which are 8,9 and 10 of the 1st because Rebel was taken out of the Premier!
That means also that if no new engine qualify's for the Premier Ktulu is promoted and WildCat takes his place at the promotion tourney.
Best wishes,
Leo.



WBEC Ridderkerk homepage.

[Fabien Letouzey]

Geschrieben von:/Posted by: Fabien Letouzey at 28 May 2004 19:27:20:
Als Antwort auf:/In reply to: Re: WBEC Ridderkerk new results. geschrieben von:/posted by: Leo Dijksman at 28 May 2004 19:19:16:

Fruit leads with 3-0 vs Pepito, still a 1/2 point needed from the last game
The "problem" is the 0,5-3,5 lose vs Kaissa 1.7!
Best wishes,
Leo.
P.S.Are you still online?

Yes, unfortunately
I don't worry for the losses, I hope the games will be interesting for the development.
Fabien.

[Norm Pollock]

Geschrieben von:/Posted by: Norm Pollock at 28 May 2004 19:40:08:
Als Antwort auf:/In reply to: Re: WBEC Ridderkerk new results. geschrieben von:/posted by: Leo Dijksman at 28 May 2004 19:25:05:

7,8 and 9 of division 2 = 25,26 and 27 of the total ranking list which are 8,9 and 10 of the 1st because Rebel was taken out of the Premier!
That means also that if no new engine qualify's for the Premier Ktulu is promoted and WildCat takes his place at the promotion tourney.
Best wishes,
Leo.

-Norm

Leo,
Then shouldn't that logic continue to 8-9-10 (instead of 7-8-9) in Div 2, Div 3, etc?
-Norm

[Leo Dijksman]

Geschrieben von:/Posted by: Leo Dijksman at 28 May 2004 19:45:25:
Als Antwort auf:/In reply to: Re: WBEC Ridderkerk new results. geschrieben von:/posted by: Fabien Letouzey at 28 May 2004 19:27:20:

Fruit leads with 3-0 vs Pepito, still a 1/2 point needed from the last game
The "problem" is the 0,5-3,5 lose vs Kaissa 1.7!
Best wishes,
Leo.
P.S.Are you still online?

Yes, unfortunately
I don't worry for the losses, I hope the games will be interesting for the development.
Fabien.

Are you working?
Something about PolyGlot and the "hangs" with Shredder (7.04), I have looked very carefull at them after I found that they always seems to happen vs the same opponents, it seems to me that it only happen when the opponent sends a move first followed by a result claim like resign/draw!?
At all the 12 "hangs" I checked here that was the case, it never happend if the opponent send a result when its his move!
Here is an example where it was "hang":
202750 first : a7d7
204828 first : result 1-0 {Black resigns}
204828 >second: result 1-0 {Black resigns}
204828 >first : force
204828 >second: force
204828 >second: ping 2
Exiting: Match Shredder_704 vs. Tao 5.7b03uci: final score 1-0-0
GameEnds(0, (null), 2)
204828 >first : quit
204828 >second: quit
Do you think its possible to "fix" this when I am right?
Best wishes,
Leo.



WBEC Ridderkerk homepage.

[Leo Dijksman]

Geschrieben von:/Posted by: Leo Dijksman at 28 May 2004 19:47:26:
Als Antwort auf:/In reply to: Re: WBEC Ridderkerk new results. geschrieben von:/posted by: Norm Pollock at 28 May 2004 19:40:08:

7,8 and 9 of division 2 = 25,26 and 27 of the total ranking list which are 8,9 and 10 of the 1st because Rebel was taken out of the Premier!
That means also that if no new engine qualify's for the Premier Ktulu is promoted and WildCat takes his place at the promotion tourney.
Best wishes,
Leo.

-Norm

Leo,
Then shouldn't that logic continue to 8-9-10 (instead of 7-8-9) in Div 2, Div 3, etc?
-Norm

No, each time a new engine qualifys all lower ranked engines go 1 rank down!
Starting with AnMon at division 1.
Best wishes,
Leo.



WBEC Ridderkerk homepage.

[Fabien Letouzey]

Geschrieben von:/Posted by: Fabien Letouzey at 28 May 2004 19:53:42:
Als Antwort auf:/In reply to: Re: WBEC Ridderkerk new results. = PolyGlot/Shredder geschrieben von:/posted by: Leo Dijksman at 28 May 2004 19:45:25:

Are you working?
Something about PolyGlot and the "hangs" with Shredder (7.04), I have looked very carefull at them after I found that they always seems to happen vs the same opponents, it seems to me that it only happen when the opponent sends a move first followed by a result claim like resign/draw!?
At all the 12 "hangs" I checked here that was the case, it never happend if the opponent send a result when its his move!
Here is an example where it was "hang":
202750 204828 Interrupting first
204828 >first : time 8609
otim 7414
204828 >first : a7d7
204828 GameEnds(31, Black resigns, 1)
Interrupting first
204828 >first : result 1-0 {Black resigns}
204828 >second: result 1-0 {Black resigns}
204828 >first : force
204828 >second: force
204828 >second: ping 2
Exiting: Match Shredder_704 vs. Tao 5.7b03uci: final score 1-0-0
GameEnds(0, (null), 2)
204828 >first : quit
204828 >second: quit
Do you think its possible to "fix" this when I am right?

No, for some reason I am staying tuned
Send me zipped WinBoard+PolyGlot log files for the same game. I have little hope, but your finding is interesting. With log files, I will think carefully about it.
With AutoQuit activated, that means PolyGlot could not even receive the game result (to be checked in log files)!
I don't want to give false hope. I think the problem is between the WBTM and WinBoard, and there is little PolyGlot can do at that level. We are fighting at millisecond level here.
Fabien.

[Fabien Letouzey]

Geschrieben von:/Posted by: Fabien Letouzey at 28 May 2004 19:56:12:
Als Antwort auf:/In reply to: Re: WBEC Ridderkerk new results. geschrieben von:/posted by: Leo Dijksman at 28 May 2004 19:47:26:

Leo,
Then shouldn't that logic continue to 8-9-10 (instead of 7-8-9) in Div 2, Div 3, etc?
-Norm

No, each time a new engine qualifys all lower ranked engines go 1 rank down!
Starting with AnMon at division 1.
Best wishes,
Leo.

Which is the whole point of getting a good rank in the new-engine qualify ...
Fabien.

[Leo Dijksman]

Geschrieben von:/Posted by: Leo Dijksman at 28 May 2004 19:59:44:
Als Antwort auf:/In reply to: Re: WBEC Ridderkerk new results. = PolyGlot/Shredder geschrieben von:/posted by: Fabien Letouzey at 28 May 2004 19:53:42:

Are you working?
Something about PolyGlot and the "hangs" with Shredder (7.04), I have looked very carefull at them after I found that they always seems to happen vs the same opponents, it seems to me that it only happen when the opponent sends a move first followed by a result claim like resign/draw!?
At all the 12 "hangs" I checked here that was the case, it never happend if the opponent send a result when its his move!
Here is an example where it was "hang":
202750 >204828 >Interrupting first
204828 >first : time 8609
otim 7414
204828 >first : a7d7
204828 >GameEnds(31, Black resigns, 1)
Interrupting first
204828 >first : result 1-0 {Black resigns}
204828 >second: result 1-0 {Black resigns}
204828 >first : force
204828 >second: force
204828 >second: ping 2
Exiting: Match Shredder_704 vs. Tao 5.7b03uci: final score 1-0-0
GameEnds(0, (null), 2)
204828 >first : quit
204828 >second: quit
Do you think its possible to "fix" this when I am right?

No, for some reason I am staying tuned
Send me zipped WinBoard+PolyGlot log files for the same game. I have little hope, but your finding is interesting. With log files, I will think carefully about it.
With AutoQuit activated, that means PolyGlot could not even receive the game result (to be checked in log files)!
ENGINE ponderhit
XBOARD 11 +599 194 234676 Rc7 h4 Rc2 Kf2 Ke6 b4 Kd5 b5 Rb2 b6 Kc6
I don't want to give false hope. I think the problem is between the WBTM and WinBoard, and there is little PolyGlot can do at that level. We are fighting at millisecond level here.
Fabien.

Autoquit is activated!
Here is the last part of the PolyGlot log from the same as above:
< ENGINE info nodes 198293
< ENGINE info currmovenumber 1 currmove e7c7
< ENGINE info depth 11 seldepth 11
< ENGINE info depth 11 seldepth 18 time 1937 score cp 599 nodes 234676 nps 121154 hashfull 20 pv e7c7 h2h4 c7c2 e1f2 f7e6 b2b4 e6d5 b4b5 c2b2 b5b6 d5c6
< ENGINE info currmovenumber 2 currmove f7e6
< XBOARD time 11459
< XBOARD otim 8704
< XBOARD f2e1
POLYGLOT MOVE Kxe1
POLYGLOT PONDERING -> THINKING (*** HIT ***)

any idee?
Leo.



WBEC Ridderkerk homepage.

[Norm Pollock]

Geschrieben von:/Posted by: Norm Pollock at 28 May 2004 20:04:18:
Als Antwort auf:/In reply to: Re: WBEC Ridderkerk new results. geschrieben von:/posted by: Fabien Letouzey at 28 May 2004 19:56:12:

Leo,
Then shouldn't that logic continue to 8-9-10 (instead of 7-8-9) in Div 2, Div 3, etc?
-Norm

No, each time a new engine qualifys all lower ranked engines go 1 rank down!
Starting with AnMon at division 1.
Best wishes,
Leo.

Which is the whole point of getting a good rank in the new-engine qualify ...
Fabien.

If Anmon is a "new" engine, then how come it is not treated as a "new" engine" It did not play for qualification to Div 1, and if that was waived because of its past history, then why doesn't Anmon play for qualification to the Premier Division?
-Norm

[Fabien Letouzey]

Geschrieben von:/Posted by: Fabien Letouzey at 28 May 2004 20:08:09:
Als Antwort auf:/In reply to: Re: WBEC Ridderkerk new results. = PolyGlot/Shredder geschrieben von:/posted by: Leo Dijksman at 28 May 2004 19:59:44:

With AutoQuit activated, that means PolyGlot could not even receive the game result (to be checked in log files)!
ENGINE ponderhit
XBOARD 11 +599 194 234676 Rc7 h4 Rc2 Kf2 Ke6 b4 Kd5 b5 Rb2 b6 Kc6

Autoquit is activated!
Here is the last part of the PolyGlot log from the same as above:
< ENGINE info nodes 198293
< ENGINE info currmovenumber 1 currmove e7c7
< ENGINE info depth 11 seldepth 11
< ENGINE info depth 11 seldepth 18 time 1937 score cp 599 nodes 234676 nps 121154 hashfull 20 pv e7c7 h2h4 c7c2 e1f2 f7e6 b2b4 e6d5 b4b5 c2b2 b5b6 d5c6
< ENGINE info currmovenumber 2 currmove f7e6
< XBOARD time 11459
< XBOARD otim 8704
< XBOARD f2e1
POLYGLOT MOVE Kxe1
POLYGLOT PONDERING -> THINKING (*** HIT ***)
any idee?

Nothing special; as expected PolyGlot did not have the time to receive the game result. It's possible that pondering slows down by just enough milliseconds to prevent PolyGlot from reading input quick enough.
I will try to write a special version that reads input from WinBoard as top priority. At the moment it tries to read as much as possible from both sides in one go (more efficient I/O).
But that won't be before Tuesday evening! Also as I said it won't fix the problem, it might just hide it a little better.
Send me full log files anyway, I might find a clue.
I'll think about your finding as well, see how it could make a difference.
Fabien.

[Fabien Letouzey]

Geschrieben von:/Posted by: Fabien Letouzey at 28 May 2004 20:09:51:
Als Antwort auf:/In reply to: Re: WBEC Ridderkerk new results. geschrieben von:/posted by: Norm Pollock at 28 May 2004 20:04:18:

Which is the whole point of getting a good rank in the new-engine qualify ...
Fabien.

If Anmon is a "new" engine, then how come it is not treated as a "new" engine" It did not play for qualification to Div 1, and if that was waived because of its past history, then why doesn't Anmon play for qualification to the Premier Division?

Why do you ask me??? ))
Fabien.

[Leo Dijksman]

Geschrieben von:/Posted by: Leo Dijksman at 28 May 2004 20:20:55:
Als Antwort auf:/In reply to: Re: WBEC Ridderkerk new results. = PolyGlot/Shredder geschrieben von:/posted by: Fabien Letouzey at 28 May 2004 20:08:09:

With AutoQuit activated, that means PolyGlot could not even receive the game result (to be checked in log files)!
ENGINE ponderhit
XBOARD 11 +599 194 234676 Rc7 h4 Rc2 Kf2 Ke6 b4 Kd5 b5 Rb2 b6 Kc6

Autoquit is activated!
Here is the last part of the PolyGlot log from the same as above:
< ENGINE info nodes 198293
< ENGINE info currmovenumber 1 currmove e7c7
< ENGINE info depth 11 seldepth 11
< ENGINE info depth 11 seldepth 18 time 1937 score cp 599 nodes 234676 nps 121154 hashfull 20 pv e7c7 h2h4 c7c2 e1f2 f7e6 b2b4 e6d5 b4b5 c2b2 b5b6 d5c6
< ENGINE info currmovenumber 2 currmove f7e6
< XBOARD time 11459
< XBOARD otim 8704
< XBOARD f2e1
POLYGLOT MOVE Kxe1
POLYGLOT PONDERING -> THINKING (*** HIT ***)
any idee?

Nothing special; as expected PolyGlot did not have the time to receive the game result. It's possible that pondering slows down by just enough milliseconds to prevent PolyGlot from reading input quick enough.
I will try to write a special version that reads input from WinBoard as top priority. At the moment it tries to read as much as possible from both sides in one go (more efficient I/O).
But that won't be before Tuesday evening! Also as I said it won't fix the problem, it might just hide it a little better.
Send me full log files anyway, I might find a clue.
I'll think about your finding as well, see how it could make a difference.
Fabien.

I will send them.
I am not so sure its a PolyGlot or WBTM problem, I think its a Shredder problem because I dont see it with the other engines.
The above debug/log is from Shredder vs Tao-uci both playing with PolyGlot, I get the same with Shredder/PolyGlot vs Maestro-uci/PolyGlot and Shredder vs "real" Winboard engines, only one thing is always the same (as far I have seen) and thats the result claim is send after the move is send.
Best wishes,
Leo.



WBEC Ridderkerk homepage.

[Fabien Letouzey]

Geschrieben von:/Posted by: Fabien Letouzey at 28 May 2004 20:25:50:
Als Antwort auf:/In reply to: Re: WBEC Ridderkerk new results. = PolyGlot/Shredder geschrieben von:/posted by: Leo Dijksman at 28 May 2004 20:20:55:

I will send them.
I am not so sure its a PolyGlot or WBTM problem, I think its a Shredder problem because I dont see it with the other engines.
The above debug/log is from Shredder vs Tao-uci both playing with PolyGlot, I get the same with Shredder/PolyGlot vs Maestro-uci/PolyGlot and Shredder vs "real" Winboard engines, only one thing is always the same (as far I have seen) and thats the result claim is send after the move is send.

Yes, the problem can only happen if the engine *also* has an I/O problem. Unfortunately a lot of engines have it. But the source of the "hanging" is not the engine.
In other words, there are two problems in different software that combine badly.
OK.
Fabien.