Gothmog 0.4.7 sel.6 gauntlet completed---live games

[Heinz van Kempen]

Geschrieben von:/Posted by: Heinz van Kempen at 27 March 2004 17:45:32:

Hi ,
now all results for Gothmog 0.4.7 sel.6 in the Nunn gauntlet over all positions with 4 min. + 2 sec. on Athlon 2600+.


Gothmog 0.4.7sel6   - Patzer 3.61           23.5 - 16.5
Gothmog 0.4.7sel6   - Smar Think 0.17a      15.0 - 25.0
Gothmog 0.4.7sel6   - AnMon5.30             20.5 - 19.5
Gothmog 0.4.7sel6   - Ktulu 4.2             16.5 - 23.5
Gothmog 0.4.7sel6   - Dragon 4.5 CF         20.5 - 19.5
Gothmog 0.4.7sel6   - Slow Chess 2.89b      22.0 - 18.0
Gothmog 0.4.7sel6   - Leila 0.53h           25.5 - 14.5
Gothmog 0.4.7sel6   - Movei 00_8_178        21.0 - 19.0
Gothmog 0.4.7sel6   - Tao 5.6               20.0 - 20.0
Gothmog 0.4.7sel6   - Amyan 1.593b          18.0 - 22.0
Gothmog 0.4.7sel6   - Anaconda 1.6.2        16.5 - 23.5
Gothmog 0.4.7sel6   - Pepito v1.59          16.0 - 24.0



This gives 51 ELO points more than Gothmog 0.4.6 in my updated Blitz rating list with Nunn positions at this time. Depending on further results for opponents in this gauntlet and on more results to come for Gothmog 0.4.6 in Nunn Top tournament this difference might increase or decrease.
I repeat again that more games are needed (at least 480) and you can already see that Comet B68 is going down like I expected. Same will happen for Ikarus, maybe only after 680 games.
http://www.husvankempen.de/nunn/
Live games from the Pyramid can be watched here today in this order here:
http://www.husvankempen.de/nunn/tour/index.htm
Snitch 0.3.47 vs. Leila 0.53h
Gothmog 0.4.7 sel 6 vs. Amy 0.8.7
Cerebro 1.23 vs. Abrok 5.0
Slow Chess 2.89b vs. Smar Think 0.17a
Games are with 20 min. + 10 sec. each move.
Currently in progress is Snitch vs. Leila
Best Regards
Heinz

[Robert Allgeuer]

Geschrieben von:/Posted by: Robert Allgeuer at 27 March 2004 19:30:48:
Als Antwort auf:/In reply to: Gothmog 0.4.7 sel.6 gauntlet completed---live games geschrieben von:/posted by: Heinz van Kempen at 27 March 2004 17:45:32:

Hi ,
now all results for Gothmog 0.4.7 sel.6 in the Nunn gauntlet over all positions with 4 min. + 2 sec. on Athlon 2600+.



This gives 51 ELO points more than Gothmog 0.4.6 in my updated Blitz rating list with Nunn positions at this time. Depending on further results for opponents in this gauntlet and on more results to come for Gothmog 0.4.6 in Nunn Top tournament this difference might increase or decrease.
I repeat again that more games are needed (at least 480) and you can already see that Comet B68 is going down like I expected. Same will happen for Ikarus, maybe only after 680 games.
http://www.husvankempen.de/nunn/
Live games from the Pyramid can be watched here today in this order here:
http://www.husvankempen.de/nunn/tour/index.htm
Snitch 0.3.47 vs. Leila 0.53h
Gothmog 0.4.7 sel 6 vs. Amy 0.8.7
Cerebro 1.23 vs. Abrok 5.0
Slow Chess 2.89b vs. Smar Think 0.17a
Games are with 20 min. + 10 sec. each move.
Currently in progress is Snitch vs. Leila
Best Regards
Heinz

>Gothmog 0.4.7sel6   - Patzer 3.61           23.5 - 16.5
>Gothmog 0.4.7sel6   - Smar Think 0.17a      15.0 - 25.0
>Gothmog 0.4.7sel6   - AnMon5.30             20.5 - 19.5
>Gothmog 0.4.7sel6   - Ktulu 4.2             16.5 - 23.5
>Gothmog 0.4.7sel6   - Dragon 4.5 CF         20.5 - 19.5
>Gothmog 0.4.7sel6   - Slow Chess 2.89b      22.0 - 18.0
>Gothmog 0.4.7sel6   - Leila 0.53h           25.5 - 14.5
>Gothmog 0.4.7sel6   - Movei 00_8_178        21.0 - 19.0
>Gothmog 0.4.7sel6   - Tao 5.6               20.0 - 20.0
>Gothmog 0.4.7sel6   - Amyan 1.593b          18.0 - 22.0
>Gothmog 0.4.7sel6   - Anaconda 1.6.2        16.5 - 23.5
>Gothmog 0.4.7sel6   - Pepito v1.59          16.0 - 24.0
>
>
even if more games are needed, still very interesting. These results give a clear hint where Gothmog fits in. Have you run a similar test with sel 4, as the comparison between these two settings would be interesting?
Robert

[Heinz van Kempen]

Geschrieben von:/Posted by: Heinz van Kempen at 27 March 2004 20:03:52:
Als Antwort auf:/In reply to: Re: Gothmog 0.4.7 sel.6 gauntlet completed---live games geschrieben von:/posted by: Robert Allgeuer at 27 March 2004 19:30:48:

Hi ,
now all results for Gothmog 0.4.7 sel.6 in the Nunn gauntlet over all positions with 4 min. + 2 sec. on Athlon 2600+.



This gives 51 ELO points more than Gothmog 0.4.6 in my updated Blitz rating list with Nunn positions at this time. Depending on further results for opponents in this gauntlet and on more results to come for Gothmog 0.4.6 in Nunn Top tournament this difference might increase or decrease.
I repeat again that more games are needed (at least 480) and you can already see that Comet B68 is going down like I expected. Same will happen for Ikarus, maybe only after 680 games.
http://www.husvankempen.de/nunn/
Live games from the Pyramid can be watched here today in this order here:
http://www.husvankempen.de/nunn/tour/index.htm
Snitch 0.3.47 vs. Leila 0.53h
Gothmog 0.4.7 sel 6 vs. Amy 0.8.7
Cerebro 1.23 vs. Abrok 5.0
Slow Chess 2.89b vs. Smar Think 0.17a
Games are with 20 min. + 10 sec. each move.
Currently in progress is Snitch vs. Leila
Best Regards
Heinz

even if more games are needed, still very interesting. These results give a clear hint where Gothmog fits in. Have you run a similar test with sel 4, as the comparison between these two settings would be interesting?
Robert

>>Gothmog 0.4.7sel6   - Patzer 3.61           23.5 - 16.5
>>Gothmog 0.4.7sel6   - Smar Think 0.17a      15.0 - 25.0
>>Gothmog 0.4.7sel6   - AnMon5.30             20.5 - 19.5
>>Gothmog 0.4.7sel6   - Ktulu 4.2             16.5 - 23.5
>>Gothmog 0.4.7sel6   - Dragon 4.5 CF         20.5 - 19.5
>>Gothmog 0.4.7sel6   - Slow Chess 2.89b      22.0 - 18.0
>>Gothmog 0.4.7sel6   - Leila 0.53h           25.5 - 14.5
>>Gothmog 0.4.7sel6   - Movei 00_8_178        21.0 - 19.0
>>Gothmog 0.4.7sel6   - Tao 5.6               20.0 - 20.0
>>Gothmog 0.4.7sel6   - Amyan 1.593b          18.0 - 22.0
>>Gothmog 0.4.7sel6   - Anaconda 1.6.2        16.5 - 23.5
>>Gothmog 0.4.7sel6   - Pepito v1.59          16.0 - 24.0
>>
>>
Hello Robert,
no, so far I did not a test with sel.4 (sel.7 would also be interesting). Regrettably I cannot do all at once, as my other tournaments should also proceed and I also promised other gauntlets. But I know one guy who also runs a lot of game and could give a comparison. )
Best Regards
Heinz
P.S.: Still astonished how Amy could lose that live Pyramid game against Gothmog in the end.

[Robert Allgeuer]

Geschrieben von:/Posted by: Robert Allgeuer at 27 March 2004 23:22:41:
Als Antwort auf:/In reply to: Re: Gothmog 0.4.7 sel.6 gauntlet completed---live games geschrieben von:/posted by: Heinz van Kempen at 27 March 2004 20:03:52:

no, so far I did not a test with sel.4 (sel.7 would also be interesting). Regrettably I cannot do all at once, as my other tournaments should also proceed and I also promised other gauntlets. But I know one guy who also runs a lot of game and could give a comparison. )
Best Regards
Heinz
P.S.: Still astonished how Amy could lose that live Pyramid game against Gothmog in the end.

I know, just a try to leverage on other results ...
A gut feeling of mine is that sel 6 could be more effective at longer time controls than at shorter ones ...
Robert

[Tord Romstad]

Geschrieben von:/Posted by: Tord Romstad at 28 March 2004 08:46:55:
Als Antwort auf:/In reply to: Re: Gothmog 0.4.7 sel.6 gauntlet completed---live games geschrieben von:/posted by: Robert Allgeuer at 27 March 2004 23:22:41:

no, so far I did not a test with sel.4 (sel.7 would also be interesting). Regrettably I cannot do all at once, as my other tournaments should also proceed and I also promised other gauntlets. But I know one guy who also runs a lot of game and could give a comparison. )

I know, just a try to leverage on other results ...
A gut feeling of mine is that sel 6 could be more effective at longer time controls than at shorter ones ...

Heinz, you have helped me much more than I deserve already, and it is definitely time for
you to test some other engines now. Thank you very much for the 480 games played with
0.4.7! I think your results give a rather accurate picture of the current strength of my
engine.
But if somebody else would like to experiment with various selectivity settings, I am
currently most interested in the range 3-6. I think 7 is probably a bit too high (though
I am not sure).
Interesting. What is this gut feeling based on? I still don't have any idea (not even a
gut feeling) what the best setting is myself.
Tord

[Robert Allgeuer]

Geschrieben von:/Posted by: Robert Allgeuer at 28 March 2004 12:47:05:
Als Antwort auf:/In reply to: Re: Gothmog 0.4.7 sel.6 gauntlet completed---live games geschrieben von:/posted by: Tord Romstad at 28 March 2004 08:46:55:

no, so far I did not a test with sel.4 (sel.7 would also be interesting). Regrettably I cannot do all at once, as my other tournaments should also proceed and I also promised other gauntlets. But I know one guy who also runs a lot of game and could give a comparison. )

I know, just a try to leverage on other results ...
A gut feeling of mine is that sel 6 could be more effective at longer time controls than at shorter ones ...

Heinz, you have helped me much more than I deserve already, and it is definitely time for
you to test some other engines now. Thank you very much for the 480 games played with
0.4.7! I think your results give a rather accurate picture of the current strength of my
engine.
But if somebody else would like to experiment with various selectivity settings, I am
currently most interested in the range 3-6. I think 7 is probably a bit too high (though
I am not sure).
Interesting. What is this gut feeling based on? I still don't have any idea (not even a
gut feeling) what the best setting is myself.
Tord

essentially from
1) results posted in CCC and here seem to suggest this, but of course only small pieces of the big picture only, so far.
2) also intuitively: higher selectivity weeds out the search tree and thus results in a smaller branch factor. Normally the smaller the branch factor the better the algorithm scales with longer thinking time.
I wanted to look at sel4 and sel6 a bit under my conditions here, failed however due to the "move rejection" problem that I have posted a few messages further up.
Robert