The final qualifier for the Stars Only 4

[Igor Gorelikov]

Geschrieben von: / Posted by: Igor Gorelikov at 20 February 2004 13:44:00:

The final qualifier for the Stars Only 4
========================================
The final qualifier for the Stars Only event is concluded.
To be qualified, a new version needs to take upper place than the best
previous version.
To be qualified, a new engine needs to take upper place than the fourth star
(that is Chinito 3.25).

List of participants
--------------------
Crafty 16.19 (the best Crafty version)
Crafty 16.16 DC (the best old version)
Crafty 19.10 SE Euwe (the best new version)
Aristarch 4.41
Delfi 4.4
SmarThink 0.17a
Thinker 4.5a New Engine (New to SO)
Thinker 4.5b
Chinito 3.25 examiner for new engines
Ruffian 1.0.1 for reference

Time control: 30'+3"
Celeron 2.6GHZ 256MB; hash settings is up to 38MB in sum for each engine;
resign is set to 6; few 5-men Nalimov's TBs.
Arena is used to arrange the event.
Stars Only 4, Qual-3
Celeron 2.6GHZ 256MB, 2004.02.16 - 2004.02.20
1 2 3 4 5 6 7 8 9 0 S R B
-------------------------------------------------------------------------
1 Delfi 4.4 XX =1 =0 1= 1= == 1= =1 11 11 13 1 107.00
2 Ruffian 1.0.1 =0 XX 0= 1= 01 11 == 1= == 11 11 2 90.75
3 Aristarch 4.41 =1 1= XX 11 0= 00 =0 1= 11 01 10= 3 93.00
4 Chinito 3.25 0= 0= 00 XX 1= 01 1= 10 =1 11 9= 4 76.50
5 Thinker 4.5b 0= 10 1= 0= XX =1 1= == == 0= 9 5-6 80.50
6 Thinker 4.5a == 00 11 10 =0 XX 00 11 == =1 9 5-6 77.25
7 SmarThink 0.17a 0= == =1 0= 0= 11 XX 10 01 0= 8= 7 76.75
8 Crafty 16.19 =0 0= 0= 01 == 00 01 XX =1 =0 6= 8-10 57.25
9 Crafty 19.10 SE Euwe 00 == 00 =0 == == 10 =0 XX =1 6= 8-10 55.25
0 Crafty 16.16 DC 00 00 10 00 1= =0 1= =1 =0 XX 6= 8-10 54.25
-------------------------------------------------------------------------
Games 90/90, +25 -30 =35

Promoted to SO
==============
Delfi 4.4 and Aristarch 4.41 is promoted to the Stars Only event.
Failed promotion
================
Thinker 4.5b but it was very close to promotion so it will be included
into the qualifiers for the next cycle.
Thinker 4.5a
Crafty 19.10 SE Euwe but it was very close to promotion so it will be
included into the qualifiers for the next cycle.
Crafty 16.16 DC but it was very close to promotion so it will be
included into the qualifiers for the next cycle.
SmarThink 0.17a


So there are 4 new entries in total:
1. Delfi 4.4
2. SOS 4 for Arena
3. AnMon 5.30
4. Aristarch 4.41
They all will play 10-game matches vs 7 stars at tournament (7-hour) time
control. The first matches are arranged for Delfi 4.4.
Best regards,
Igor Gorelikov
The games of the qualifier can be downloaded from here:
http://www.digichess.gr/infiniteloop/games/SO4_q3.zip
=================================================
For more info on the Stars Only event apply to
http://www.digichess.gr/infiniteloop/home.php

[Uri Blass]

Geschrieben von: / Posted by: Uri Blass at 20 February 2004 14:26:13:
Als Antwort auf: / In reply to: The final qualifier for the Stars Only 4 geschrieben von: / posted by: Igor Gorelikov at 20 February 2004 13:44:00:

The final qualifier for the Stars Only 4
========================================
The final qualifier for the Stars Only event is concluded.
To be qualified, a new version needs to take upper place than the best
previous version.
To be qualified, a new engine needs to take upper place than the fourth star
(that is Chinito 3.25).

List of participants
--------------------
Crafty 16.19 (the best Crafty version)
Crafty 16.16 DC (the best old version)
Crafty 19.10 SE Euwe (the best new version)
Aristarch 4.41
Delfi 4.4
SmarThink 0.17a
Thinker 4.5a New Engine (New to SO)
Thinker 4.5b
Chinito 3.25 examiner for new engines
Ruffian 1.0.1 for reference

Time control: 30'+3"
Celeron 2.6GHZ 256MB; hash settings is up to 38MB in sum for each engine;>Arena is used to arrange the event.
Stars Only 4, Qual-3
Celeron 2.6GHZ 256MB, 2004.02.16 - 2004.02.20
1 2 3 4 5 6 7 8 9 0 S R B
-------------------------------------------------------------------------
1 Delfi 4.4 XX =1 =0 1= 1= == 1= =1 11 11 13 1 107.00
2 Ruffian 1.0.1 =0 XX 0= 1= 01 11 == 1= == 11 11 2 90.75
3 Aristarch 4.41 =1 1= XX 11 0= 00 =0 1= 11 01 10= 3 93.00
4 Chinito 3.25 0= 0= 00 XX 1= 01 1= 10 =1 11 9= 4 76.50
5 Thinker 4.5b 0= 10 1= 0= XX =1 1= == == 0= 9 5-6 80.50
6 Thinker 4.5a == 00 11 10 =0 XX 00 11 == =1 9 5-6 77.25
7 SmarThink 0.17a 0= == =1 0= 0= 11 XX 10 01 0= 8= 7 76.75
8 Crafty 16.19 =0 0= 0= 01 == 00 01 XX =1 =0 6= 8-10 57.25
9 Crafty 19.10 SE Euwe 00 == 00 =0 == == 10 =0 XX =1 6= 8-10 55.25
0 Crafty 16.16 DC 00 00 10 00 1= =0 1= =1 =0 XX 6= 8-10 54.25
-------------------------------------------------------------------------
Games 90/90, +25 -30 =35

Promoted to SO
==============
Delfi 4.4 and Aristarch 4.41 is promoted to the Stars Only event.
Failed promotion
================
resign is set to 6; few 5-men Nalimov's TBs.
Thinker 4.5b but it was very close to promotion so it will be included
into the qualifiers for the next cycle.
Thinker 4.5a
Crafty 19.10 SE Euwe but it was very close to promotion so it will be
included into the qualifiers for the next cycle.
Crafty 16.16 DC but it was very close to promotion so it will be
included into the qualifiers for the next cycle.
SmarThink 0.17a


So there are 4 new entries in total:
1. Delfi 4.4
2. SOS 4 for Arena
3. AnMon 5.30
4. Aristarch 4.41
They all will play 10-game matches vs 7 stars at tournament (7-hour) time
control. The first matches are arranged for Delfi 4.4.
Best regards,
Igor Gorelikov

I believe that all these stupid GM only reduce Crafty chances to promote.
I guess that it is better that in the future you do not test some personality of human players for Crafty because it is logical to guess that they are slightly weaker than the default Crafty but not weaker enough not to have chances to be lucky in one tournament.
The result is that in a tournament for crafties there is more than 50% chances that the wrong crafty promotes if you test a lot of personalities.
My opinion is that time is limited and it is better to use it to test only a single new version of crafty when first qualification tournament should be between different crafties but between one new crafty and other engines.
Uri

[Igor Gorelikov]

Geschrieben von: / Posted by: Igor Gorelikov at 20 February 2004 14:45:56:
Als Antwort auf: / In reply to: Re: The final qualifier for the Stars Only 4 geschrieben von: / posted by: Uri Blass at 20 February 2004 14:26:13:

I believe that all these stupid GM only reduce Crafty chances to promote.
I guess that it is better that in the future you do not test some personality of human players for Crafty because it is logical to guess that they are slightly weaker than the default Crafty but not weaker enough not to have chances to be lucky in one tournament.
The result is that in a tournament for crafties there is more than 50% chances that the wrong crafty promotes if you test a lot of personalities.
My opinion is that time is limited and it is better to use it to test only a single new version of crafty when first qualification tournament should be between different crafties but between one new crafty and other engines.
Uri

"they are slightly weaker than the default Crafty"
My tests shows quite the reverse.
And I'm not sure that personalities are a bad way to try different settings.
But just from curiosity I'm ready to run a gauntlet for the "clear" version of Crafty 19.10 vs all participants of the final qualifier.
My opinion is that there is no "wrong crafty" and all of them are very close by strenght.
regards,
Igor

[Uri Blass]

Geschrieben von: / Posted by: Uri Blass at 20 February 2004 15:37:17:
Als Antwort auf: / In reply to: Re: The final qualifier for the Stars Only 4 geschrieben von: / posted by: Igor Gorelikov at 20 February 2004 14:45:56:

I believe that all these stupid GM only reduce Crafty chances to promote.
I guess that it is better that in the future you do not test some personality of human players for Crafty because it is logical to guess that they are slightly weaker than the default Crafty but not weaker enough not to have chances to be lucky in one tournament.
The result is that in a tournament for crafties there is more than 50% chances that the wrong crafty promotes if you test a lot of personalities.

"they are slightly weaker than the default Crafty"
My tests shows quite the reverse.
And I'm not sure that personalities are a bad way to try different settings.
But just from curiosity I'm ready to run a gauntlet for the "clear" version of Crafty 19.10 vs all participants of the final qualifier.
My opinion is that there is no "wrong crafty" and all of them are very close by strenght.

I do not claim that testing personalities is a bad idea but if you play with them only a short number of games then even if one of them is slightly stronger then you still have a bigger probability to use weaker personality then the default.
The difference is probably too small to get conclusions without playing hundreds of games and if you do not use your time mainly for testing
Crafty you cannot know.
probably most of the crafty personalities are slightly weaker than default crafty and even if there is one that is slightly stronger most of the chances are that a weaker version is going to qualify.
I agree that the difference between different crafties is small.
Uri