Think Again!

[Ovyron]

I made this tourney to test the great engine Thinker 5.1C, I'm still awed by its extreme aggressiveness, it makes sound sacrifices that the opponents have trouble understanding, and they are surprised later on to discover that they have a lost game. I also added the Active personality for comparison (Ironically, the Active personality seems more passive.)

I know that the Active one is meant to be only used against humans, but I wanted to see how it fared.

Settings
Athlon 2.1Ghz 32bit
ChessPartner GUI
1'+ 7" with ponder off
64MB hash each engine
No tablebases (Adjudicating games with online tablebases was taking too long)
No books
2 Rounds Gauntlet each Thinker and one match between Thinkers, 60 games

Code: Select all

Think Again (Midway results)
2008.03.17 - 2008.03.22
                             Score     Th AT Gl Fr Sp To Ry Na br
------------------------------------------------------------------
 1: Thinker                10.5 / 16   XX 11 01 01 10 =0 10 11 11
 2: ActiveThinker           4.0 / 16   00 XX 00 00 00 10 0= =0 11
 3: Glaurung 2.0.1          3.0 /  4   10 11 XX .. .. .. .. .. ..
 4: Fruit 2.3.1             3.0 /  4   10 11 .. XX .. .. .. .. ..
 5: Spike 1.2 Turin         3.0 /  4   01 11 .. .. XX .. .. .. ..
 6: Toga II 3.1.2SE Chekov  2.5 /  4   =1 01 .. .. .. XX .. .. ..
 7: Rybka 1.0 Beta 32-bit   2.5 /  4   01 1= .. .. .. .. XX .. ..
 8: Naum 2.0                1.5 /  4   00 =1 .. .. .. .. .. XX ..
 9: bright-0.2c             0.0 /  4   00 00 .. .. .. .. .. .. XX
------------------------------------------------------------------
30 games: +15 =3 -12

Ratings of participants so far:

Code: Select all

Rank Name                     Elo    +    - games score oppo. draws 
   1 Thinker                 2865  146  132    16   66%  2752    6% 
   2 Rybka 1.0 Beta 32-bit   2835   96   90    32   66%  2739   25% 
   3 Toga II 3.1.2SE Chekov  2826   94   90    32   63%  2740   25% 
   4 Spike 1.2 Turin         2807   94   91    32   59%  2742   25% 
   5 Glaurung 2.0.1          2793   95   93    32   56%  2744   19% 
   6 bright-0.2c             2752   95   95    32   50%  2749   13% 
   7 Fruit 2.3.1             2704   92   94    32   42%  2755   22% 
   8 Naum 2.0                2694   91   95    32   41%  2756   19% 
   9 ThinkerA                2609  131  155    16   25%  2784   13% 

PGN at the end of the tourney.

[ArmyBridge]

I wonder if you optimized Thinker to be agressive, some special setting? incredibly that the version that assumes would have to play passive played more aggressive than Active thinker... maybe it was because the opening, time ago I tought that Rybka was weak in tactics, suddenly Perl clarify this wrong tought with its extraordinary gambit's test, Rybka play all sort of gambits very very well ,... I was very Impressive Rybka and Naum smashed everyone in great fashion, I'll follow you test with great interest.
Regards

[Ovyron]

the version that assumes would have to play passive played more aggressive than Active thinker... maybe it was because the opening

That's how Thinker plays from the opening position, and after just changing one move from the opening, so I deem the playing style accurate.

It's as if Thinker had the "Active" and "Passive" names switched around , but the active Thinker assumes that it's playing a human and will play "blunders" that you wouldn't be able to take advantage of unless you were an engine.

I'll follow you test with great interest.
Regards

Thank you for your interest!