AEGT: rook/bishop vs Queen 2.44

[Will Singleton]

Geschrieben von:/Posted by: Will Singleton at 12 August 2004 22:35:16:


Queen 2.44 vs Rook class, 40/25, fx-51 2.2ghz
                             Score       1    2 
-------------------------------------------------
01: Queen 2.44             12.5 / 24   10w- 10b+
02: Comet B.68              2.0 /  2    1b+  1w+
03: Arasan 7.4              1.5 /  2    1b+  1w=
04: Naum 1.2                1.0 /  2    1b=  1w=
05: Postmodernist 1010a     1.0 /  2    1b+  1w-
06: The Baron 1.4.0 b 2     1.0 /  2    1b+  1w-
07: Amateur 2.80            1.0 /  2    1b-  1w+
08: Pharaon 2.62            1.0 /  2    1b+  1w-
09: Pepito 1.59             1.0 /  2    1b-  1w+
10: Terra 3.3 beta11        1.0 /  2    1b+  1w-
11: Knightdreamer 3.3       0.5 /  2    1b=  1w-
12: Frenzee 159             0.0 /  2    1b-  1w-
13: The Crazy Bishop 0052   0.0 /  2    1b-  1w-
-------------------------------------------------
24 games: +8 =3 -13

Queen 2.44 vs Bishop class, 40/25, fx-51 2.2ghz
                          Score       1    2    3 
---------------------------------------------------
01: Queen 2.44          11.5 / 24   10w+ 10b-  7w-
02: Spike 0.6            2.0 /  2    1b+  1w+ ....
03: Djinn 0.870          2.0 /  2    1b+  1w+ ....
04: Booot 4.03           1.5 /  2    1b+  1w= ....
05: Hermann 1.1.8        1.5 /  2    1b+  1w= ....
06: Cerebro 1.27         1.0 /  2    1b+  1w- ....
07: Trace 1.29           1.0 /  2    1b- ....  1b+
08: DanChess 1.06 DC     1.0 /  2    1b=  1w= ....
09: Bruja 1.7            1.0 /  2    1b=  1w= ....
10: Snitch 0.5.5         1.0 /  2    1b-  1w+ ....
11: BigLion 2.23i        0.5 /  2    1b=  1w- ....
12: Alarm 0.93.1         0.0 /  2    1b-  1w- ....
13: BlackBishop 0.9.7i   0.0 /  2    1b-  1w- ....
---------------------------------------------------
24 games: +7 =7 -10

Hard to tell from a small sample, but it's possible the rook and bishop classes are fairly close in strength. We'll know more after a few more gauntlets. (which are taking a long time, btw, due to Averno and others not resigning, grumble. This is my first shot at testing other progs, and now I understand how pissed off one can get spending *days* of cpu-time in lost positions.)

[Roger Brown]

Geschrieben von:/Posted by: Roger Brown at 12 August 2004 23:15:44:
Als Antwort auf:/In reply to: AEGT: rook/bishop vs Queen 2.44 geschrieben von:/posted by: Will Singleton at 12 August 2004 22:35:16:

We'll know more after a few more gauntlets. (which are taking a long time, btw, due to Averno and others not resigning, grumble. This is my first shot at testing other progs, and now I understand how pissed off one can get spending *days* of cpu-time in lost positions.)

Hello Will,
Leo, Gunther and other TD's will love you for that comment!!
Later.

[Ulrich Tuerke]

Geschrieben von:/Posted by: Ulrich Tuerke at 12 August 2004 23:47:59:
Als Antwort auf:/In reply to: AEGT: rook/bishop vs Queen 2.44 geschrieben von:/posted by: Will Singleton at 12 August 2004 22:35:16:

Hard to tell from a small sample, but it's possible the rook and bishop classes are fairly close in strength. We'll know more after a few more gauntlets. (which are taking a long time, btw, due to Averno and others not resigning, grumble. This is my first shot at testing other progs, and now I understand how pissed off one can get spending *days* of cpu-time in lost positions.)

>Queen 2.44 vs Rook class, 40/25, fx-51 2.2ghz
>                             Score       1    2 
>-------------------------------------------------
>01: Queen 2.44             12.5 / 24   10w- 10b+
>02: Comet B.68              2.0 /  2    1b+  1w+
>03: Arasan 7.4              1.5 /  2    1b+  1w=
>04: Naum 1.2                1.0 /  2    1b=  1w=
>05: Postmodernist 1010a     1.0 /  2    1b+  1w-
>06: The Baron 1.4.0 b 2     1.0 /  2    1b+  1w-
>07: Amateur 2.80            1.0 /  2    1b-  1w+
>08: Pharaon 2.62            1.0 /  2    1b+  1w-
>09: Pepito 1.59             1.0 /  2    1b-  1w+
>10: Terra 3.3 beta11        1.0 /  2    1b+  1w-
>11: Knightdreamer 3.3       0.5 /  2    1b=  1w-
>12: Frenzee 159             0.0 /  2    1b-  1w-
>13: The Crazy Bishop 0052   0.0 /  2    1b-  1w-
>-------------------------------------------------

Some minor comment: summing up the points gained by the opponents, I obtain the number 11.
Then Queen should have obtained 13 points instead of 12.5
Or have I got something wrong, Will ?
Thanks,
Uli
>24 games: +8 =3 -13
>
>Queen 2.44 vs Bishop class, 40/25, fx-51 2.2ghz
>                          Score       1    2    3 
>---------------------------------------------------
>01: Queen 2.44          11.5 / 24   10w+ 10b-  7w-
>02: Spike 0.6            2.0 /  2    1b+  1w+ ....
>03: Djinn 0.870          2.0 /  2    1b+  1w+ ....
>04: Booot 4.03           1.5 /  2    1b+  1w= ....
>05: Hermann 1.1.8        1.5 /  2    1b+  1w= ....
>06: Cerebro 1.27         1.0 /  2    1b+  1w- ....
>07: Trace 1.29           1.0 /  2    1b- ....  1b+
>08: DanChess 1.06 DC     1.0 /  2    1b=  1w= ....
>09: Bruja 1.7            1.0 /  2    1b=  1w= ....
>10: Snitch 0.5.5         1.0 /  2    1b-  1w+ ....
>11: BigLion 2.23i        0.5 /  2    1b=  1w- ....
>12: Alarm 0.93.1         0.0 /  2    1b-  1w- ....
>13: BlackBishop 0.9.7i   0.0 /  2    1b-  1w- ....
>---------------------------------------------------
>24 games: +7 =7 -10
>

[Will Singleton]

Geschrieben von:/Posted by: Will Singleton at 13 August 2004 00:05:35:
Als Antwort auf:/In reply to: Re: AEGT: rook/bishop vs Queen 2.44 geschrieben von:/posted by: Ulrich Tuerke at 12 August 2004 23:47:59:

Queen 2.44 vs Rook class, 40/25, fx-51 2.2ghz
                             Score       1    2
-------------------------------------------------
01: Queen 2.44             12.5 / 24   10w- 10b+
02: Comet B.68              2.0 /  2    1b+  1w+
03: Arasan 7.4              1.5 /  2    1b+  1w=
04: Naum 1.2                1.0 /  2    1b=  1w=
05: Postmodernist 1010a     1.0 /  2    1b+  1w-
06: The Baron 1.4.0 b 2     1.0 /  2    1b+  1w-
07: Amateur 2.80            1.0 /  2    1b-  1w+
08: Pharaon 2.62            1.0 /  2    1b+  1w-
09: Pepito 1.59             1.0 /  2    1b-  1w+
10: Terra 3.3 beta11        1.0 /  2    1b+  1w-
11: Knightdreamer 3.3       0.5 /  2    1b=  1w-
12: Frenzee 159             0.0 /  2    1b-  1w-
13: The Crazy Bishop 0052   0.0 /  2    1b-  1w-
-------------------------------------------------

Some minor comment: summing up the points gained by the opponents, I obtain the number 11.
Then Queen should have obtained 13 points instead of 12.5
Or have I got something wrong, Will ?
Thanks,
Uli

Right you are, Uli. I had to replay the Naum games because I used v1.0 instead of 1.2, and instead of running a new chart, I just changed the Naum result, forgetting to change Queen's total. Queen gets 13/24.

[Heinz van Kempen]

Geschrieben von:/Posted by: Heinz van Kempen at 13 August 2004 09:10:23:
Als Antwort auf:/In reply to: AEGT: rook/bishop vs Queen 2.44 geschrieben von:/posted by: Will Singleton at 12 August 2004 22:35:16:

Hard to tell from a small sample, but it's possible the rook and bishop classes are fairly close in strength. We'll know more after a few more gauntlets. (which are taking a long time, btw, due to Averno and others not resigning, grumble. This is my first shot at testing other progs, and now I understand how pissed off one can get spending *days* of cpu-time in lost positions.)

>Queen 2.44 vs Rook class, 40/25, fx-51 2.2ghz
>                             Score       1    2 
>-------------------------------------------------
>01: Queen 2.44             12.5 / 24   10w- 10b+
>02: Comet B.68              2.0 /  2    1b+  1w+
>03: Arasan 7.4              1.5 /  2    1b+  1w=
>04: Naum 1.2                1.0 /  2    1b=  1w=
>05: Postmodernist 1010a     1.0 /  2    1b+  1w-
>06: The Baron 1.4.0 b 2     1.0 /  2    1b+  1w-
>07: Amateur 2.80            1.0 /  2    1b-  1w+
>08: Pharaon 2.62            1.0 /  2    1b+  1w-
>09: Pepito 1.59             1.0 /  2    1b-  1w+
>10: Terra 3.3 beta11        1.0 /  2    1b+  1w-
>11: Knightdreamer 3.3       0.5 /  2    1b=  1w-
>12: Frenzee 159             0.0 /  2    1b-  1w-
>13: The Crazy Bishop 0052   0.0 /  2    1b-  1w-
>-------------------------------------------------
>24 games: +8 =3 -13
>
>Queen 2.44 vs Bishop class, 40/25, fx-51 2.2ghz
>                          Score       1    2    3 
>---------------------------------------------------
>01: Queen 2.44          11.5 / 24   10w+ 10b-  7w-
>02: Spike 0.6            2.0 /  2    1b+  1w+ ....
>03: Djinn 0.870          2.0 /  2    1b+  1w+ ....
>04: Booot 4.03           1.5 /  2    1b+  1w= ....
>05: Hermann 1.1.8        1.5 /  2    1b+  1w= ....
>06: Cerebro 1.27         1.0 /  2    1b+  1w- ....
>07: Trace 1.29           1.0 /  2    1b- ....  1b+
>08: DanChess 1.06 DC     1.0 /  2    1b=  1w= ....
>09: Bruja 1.7            1.0 /  2    1b=  1w= ....
>10: Snitch 0.5.5         1.0 /  2    1b-  1w+ ....
>11: BigLion 2.23i        0.5 /  2    1b=  1w- ....
>12: Alarm 0.93.1         0.0 /  2    1b-  1w- ....
>13: BlackBishop 0.9.7i   0.0 /  2    1b-  1w- ....
>---------------------------------------------------
>24 games: +7 =7 -10
>
Hi all ,
for such a big tournament with long timecontrol it really sums up to days lost for each tester, where more games could have been played. Not only those totally won positions cost time. There are still a lot of top engines not supporting tablebases and we voted that only those who have implemented the access should have them granted.
Yesterday alone I had several games where this caused 50 moves additionally and because we also voted that neither the GUI nor the tester is allowed to adjudicate games I had some "fun" to watch a 50 more moves K+R against K+N in a game Movei vs. KoK, a total of 179 moves in a game Fruit vs. Wild Cat and Fruit the latter not "knowing" that it can´t win with bishop and pawn against the lone king when the square the pawn should promote is not the same colour the diagonal has where the bishop is on, etc..
I hope we can take some decisions for AEGT 2 avoiding such superfluent time losses.
Best Regards
Heinz

[Rémi Coulom]

Geschrieben von:/Posted by: Rémi Coulom at 13 August 2004 09:31:01:
Als Antwort auf:/In reply to: AEGT: rook/bishop vs Queen 2.44 geschrieben von:/posted by: Will Singleton at 12 August 2004 22:35:16:

TCB can be told to resign when eval reaches some threshold. I do not use it by default because I am not 100% certain it is safe. Also, there is still the possibility to win because of opponent's bug. Anyway, if you'd like to use this option, in tcbxinit.cui:

special
 &#59; Opening book
 &#59; ~~~~~~~~~~~~
 pgnbook <book.pgn

 &#59; Hash table (size in bytes)
 &#59; ~~~~~~~~~~
 size 10000000

 &#59; Resign options
 &#59; ~~~~~~~~~~~~~~
 resigneval 500
x

I guess Arena can also do something similar automatically, anyway.
Rémi

[Volker Pittlik]

Geschrieben von:/Posted by: Volker Pittlik at 13 August 2004 13:53:22:
Als Antwort auf:/In reply to: Re: AEGT: rook/bishop vs Queen 2.44 geschrieben von:/posted by: Heinz van Kempen at 13 August 2004 09:10:23:

...

Yesterday alone I had several games where this caused 50 moves additionally and because we also voted that neither the GUI nor the tester is allowed to adjudicate games I had some "fun" to watch a 50 more moves K+R against K+N in a game Movei vs. KoK, a total of 179 moves in a game Fruit vs. Wild Cat and Fruit the latter not "knowing" that it can´t win with bishop and pawn against the lone king when the square the pawn should promote is not the same colour the diagonal has where the bishop is on, etc..
I hope we can take some decisions for AEGT 2 avoiding such superfluent time

losses.
Yes but sometimes (seldom) such games come to a normal end. I've seen one in round 1 of AEGT:


Adjudicated positions should be checked carefully especially if they have been adjudicated be me.
Volker

[Will Singleton]

Geschrieben von:/Posted by: Will Singleton at 13 August 2004 17:53:31:
Als Antwort auf:/In reply to: TCB's resigneval option geschrieben von:/posted by: Rémi Coulom at 13 August 2004 09:31:01:

TCB can be told to resign when eval reaches some threshold. I do not use it by default because I am not 100% certain it is safe. Also, there is still the possibility to win because of opponent's bug. Anyway, if you'd like to use this option, in tcbxinit.cui:

I guess Arena can also do something similar automatically, anyway.
Rémi

>special
> &#59; Opening book
> &#59; ~~~~~~~~~~~~
> pgnbook <book.pgn
>
> &#59; Hash table (size in bytes)
> &#59; ~~~~~~~~~~
> size 10000000
>
> &#59; Resign options
> &#59; ~~~~~~~~~~~~~~
> resigneval 500
>x
>
Thanks, I'll try it.

[Dan Honeycutt]

Geschrieben von:/Posted by: Dan Honeycutt at 13 August 2004 23:08:48:
Als Antwort auf:/In reply to: Re: AEGT: rook/bishop vs Queen 2.44 geschrieben von:/posted by: Heinz van Kempen at 13 August 2004 09:10:23:

Yesterday alone I had several games where this caused 50 moves additionally and because we also voted that neither the GUI nor the tester is allowed to adjudicate games
I hope we can take some decisions for AEGT 2 avoiding such superfluent time losses.
Best Regards
Heinz

I think you should vote again. This maybe made sense when AEGT team was in the formative stages. Now the team is together, all testers are experienced and knowledgable and should be allowed to adjudicate games.
Great tournament with close fights in all classes. As usual I was wrong with my prediction that Pepito would coast in Rook class. Only Green Light has what looks like a comfortable lead, but there are other engines in Queen class capable of reeling off a string of wins. We shall see.
Dan H.

I had some "fun" to watch a 50 more moves K+R against K+N in a game Movei vs. KoK, a total of 179 moves in a game Fruit vs. Wild Cat and Fruit the latter not "knowing" that it can´t win with bishop and pawn against the lone king when the square the pawn should promote is not the same colour the diagonal has where the bishop is on, etc..