Naum results after one cycle + 4 matches

Tony Thomas · Post by **Tony Thomas** » Tue Feb 12, 2008 6:39 pm

I have a new leader in my premier edition tournament aka Manipulator Prince. The new manipulator Prince for now is naum. It scored 24.5/32 with black against 32 of the strongest non chessbase engines (minues loop) and it won the 4 games after that as well with white. With a start rating of 2942, which is 200 pints better than the previous version (take it with a grain of salt), you should also note that previous Naum didnt like my time control so it always performed below par..It is currently rated 110 points higher than free Rybka beta after 36 games. I have got another 92 more matches to go so we shall see how it turns out.

Code: Select all

1 Naum 3                    : 2942   36 (+ 24,=  9,-  3), 79.2 %

Rybka v1.0 Beta.w32           :   2 (+  1,=  1,-  0), 75.0 %
WildCat 7.0                   :   1 (+  1,=  0,-  0), 100.0 %
Spike 1.2 Turin               :   2 (+  2,=  0,-  0), 100.0 %
Smarthink 1.00                :   1 (+  1,=  0,-  0), 100.0 %
Prodeo 1.2                    :   1 (+  1,=  0,-  0), 100.0 %
Trace 1.37a                   :   1 (+  1,=  0,-  0), 100.0 %
Gandalf 6.01                  :   1 (+  1,=  0,-  0), 100.0 %
Ktulu 8.0                     :   2 (+  1,=  1,-  0), 75.0 %
Thinker 4.7a                  :   1 (+  0,=  1,-  0), 50.0 %
Pharaon 3.5.1                 :   1 (+  1,=  0,-  0), 100.0 %
SOS 5.1                       :   1 (+  1,=  0,-  0), 100.0 %
Ruffian 1.0.5                 :   1 (+  0,=  1,-  0), 50.0 %
SlowChess Blitz WV 2.1        :   1 (+  1,=  0,-  0), 100.0 %
Aristarch 4.50                :   1 (+  0,=  0,-  1),  0.0 %
CM10th D2Alos                 :   1 (+  1,=  0,-  0), 100.0 %
ChessTiger2007.1 UCI          :   2 (+  2,=  0,-  0), 100.0 %
Fruit 2.3                     :   1 (+  0,=  1,-  0), 50.0 %
Naum 2.2                      :   1 (+  0,=  0,-  1),  0.0 %
DeepSjeng27                   :   1 (+  1,=  0,-  0), 100.0 %
Delfi 5.2                     :   1 (+  1,=  0,-  0), 100.0 %
Movei00_8_438                 :   1 (+  1,=  0,-  0), 100.0 %
BugChess2_V1_5_2              :   1 (+  1,=  0,-  0), 100.0 %
Shredder11UCI                 :   1 (+  0,=  0,-  1),  0.0 %
Crafty 21.6 JA                :   1 (+  0,=  1,-  0), 50.0 %
Scorpio 2.0                   :   1 (+  1,=  0,-  0), 100.0 %
AlaricWB707                   :   1 (+  0,=  1,-  0), 50.0 %
Glaurung 2.0.1 JA             :   1 (+  1,=  0,-  0), 100.0 %
Hiarcs11.2SPUCI               :   1 (+  1,=  0,-  0), 100.0 %
Bright-0.2c                   :   1 (+  0,=  1,-  0), 50.0 %
Frenzee Dec 07                :   1 (+  1,=  0,-  0), 100.0 %
Zappa Mexico II               :   1 (+  1,=  0,-  0), 100.0 %
TogaII 1.4 beta 5c            :   1 (+  0,=  1,-  0), 50.0 %

Tony Thomas · Post by **Tony Thomas** » Tue Feb 12, 2008 8:14 pm

Rating dropped by 26 points after another 14 games due to one loss and way too many draws. Naum should have lost another game, but SOS ran out of time, and I was not close enough to the computer to adjudicate the match.

Code: Select all

1 Naum 3                    : 2916   50 (+ 32,= 14,-  4), 78.0 %

Rybka v1.0 Beta.w32           :   2 (+  1,=  1,-  0), 75.0 %
WildCat 7.0                   :   2 (+  2,=  0,-  0), 100.0 %
Spike 1.2 Turin               :   2 (+  2,=  0,-  0), 100.0 %
Smarthink 1.00                :   2 (+  1,=  1,-  0), 75.0 %
Prodeo 1.2                    :   2 (+  1,=  1,-  0), 75.0 %
Trace 1.37a                   :   2 (+  1,=  1,-  0), 75.0 %
Gandalf 6.01                  :   2 (+  1,=  1,-  0), 75.0 %
Ktulu 8.0                     :   2 (+  1,=  1,-  0), 75.0 %
Thinker 4.7a                  :   2 (+  1,=  1,-  0), 75.0 %
Pharaon 3.5.1                 :   2 (+  2,=  0,-  0), 100.0 %
SOS 5.1                       :   2 (+  2,=  0,-  0), 100.0 %
Ruffian 1.0.5                 :   2 (+  0,=  2,-  0), 50.0 %
SlowChess Blitz WV 2.1        :   2 (+  2,=  0,-  0), 100.0 %
Aristarch 4.50                :   2 (+  1,=  0,-  1), 50.0 %
CM10th D2Alos                 :   2 (+  2,=  0,-  0), 100.0 %
ChessTiger2007.1 UCI          :   2 (+  2,=  0,-  0), 100.0 %
Fruit 2.3                     :   2 (+  0,=  1,-  1), 25.0 %
Naum 2.2                      :   2 (+  1,=  0,-  1), 50.0 %
DeepSjeng27                   :   1 (+  1,=  0,-  0), 100.0 %
Delfi 5.2                     :   1 (+  1,=  0,-  0), 100.0 %
Movei00_8_438                 :   1 (+  1,=  0,-  0), 100.0 %
BugChess2_V1_5_2              :   1 (+  1,=  0,-  0), 100.0 %
Shredder11UCI                 :   1 (+  0,=  0,-  1),  0.0 %
Crafty 21.6 JA                :   1 (+  0,=  1,-  0), 50.0 %
Scorpio 2.0                   :   1 (+  1,=  0,-  0), 100.0 %
AlaricWB707                   :   1 (+  0,=  1,-  0), 50.0 %
Glaurung 2.0.1 JA             :   1 (+  1,=  0,-  0), 100.0 %
Hiarcs11.2SPUCI               :   1 (+  1,=  0,-  0), 100.0 %
Bright-0.2c                   :   1 (+  0,=  1,-  0), 50.0 %
Frenzee Dec 07                :   1 (+  1,=  0,-  0), 100.0 %
Zappa Mexico II               :   1 (+  1,=  0,-  0), 100.0 %
TogaII 1.4 beta 5c            :   1 (+  0,=  1,-  0), 50.0 %

Uri Blass · Post by **Uri Blass** » Wed Feb 13, 2008 9:26 am

Note that you did not test the best free version of rybka

http://computerchess.org.uk/ccrl/404/cg ... _length=30

Strelka 2.0 B 64-bit seems to be the best free version of rybka in the rybka family and I guess that for you 32 bits version of strelka is also better than 32 bits version of rybka

Uri

Tony Thomas · Post by **Tony Thomas** » Wed Feb 13, 2008 9:45 am

Strelka is strelka, and I am not interested in wasting my time testing another similar engine. The difference between Rybka 1.0 64bit and Strelka 2.0B 64bit is 22 points. If we consider that the difference between 32bit versions are similar then it is almost pointless. I would buy real rybka sooner or later.

ozziejoe · Post by **ozziejoe** » Wed Feb 13, 2008 9:48 am

Note that strelka is a stolen version of rybka. If tony were to test it, would he not be reinforcing the stealing of engine code? I think it shows alot of character and principle to.....ignore strelka

Tony Thomas · Post by **Tony Thomas** » Wed Feb 13, 2008 9:50 am

I was only able to get another 35 games, but the rating of Naum went up again. Here is the current standing..I am not sure why but Naum seems to be doing better with black. May be it has something to do with the opening book (tiny Naum book).

Code: Select all

  1 Naum 3                         : 2943   73  70    85    80.6 %   2696   27.1 %
  2 Hiarcs11.2SPUCI                : 2896   60  58   145    75.2 %   2704   15.2 %
  3 Shredder11UCI                  : 2851   50  49   169    70.1 %   2703   21.9 %
  4 HiarcsX54UCI                   : 2849   47  46   208    70.9 %   2694   17.8 %
  5 TogaII 1.4 beta 5c             : 2834   62  61   107    67.3 %   2708   22.4 %
  6 TogaII 1.2 beta 2a KS/EHP      : 2832   44  43   216    69.0 %   2693   22.2 %
  7 Fruit 2.3                      : 2831   46  45   186    67.7 %   2702   24.7 %
  8 Rybka v1.0 Beta.w32            : 2830   40  39   242    68.0 %   2699   25.2 %

Code: Select all

1 Naum 3                    : 2943   85 (+ 57,= 23,-  5), 80.6 %

Rybka v1.0 Beta.w32           :   3 (+  2,=  1,-  0), 83.3 %
WildCat 7.0                   :   3 (+  3,=  0,-  0), 100.0 %
Spike 1.2 Turin               :   3 (+  2,=  1,-  0), 83.3 %
Smarthink 1.00                :   3 (+  2,=  1,-  0), 83.3 %
Prodeo 1.2                    :   3 (+  2,=  1,-  0), 83.3 %
Trace 1.37a                   :   3 (+  2,=  1,-  0), 83.3 %
Gandalf 6.01                  :   3 (+  2,=  1,-  0), 83.3 %
Ktulu 8.0                     :   3 (+  1,=  2,-  0), 66.7 %
Thinker 4.7a                  :   3 (+  1,=  2,-  0), 66.7 %
Pharaon 3.5.1                 :   3 (+  2,=  1,-  0), 83.3 %
SOS 5.1                       :   3 (+  3,=  0,-  0), 100.0 %
Ruffian 1.0.5                 :   3 (+  0,=  3,-  0), 50.0 %
SlowChess Blitz WV 2.1        :   3 (+  3,=  0,-  0), 100.0 %
Aristarch 4.50                :   3 (+  2,=  0,-  1), 66.7 %
CM10th D2Alos                 :   3 (+  3,=  0,-  0), 100.0 %
ChessTiger2007.1 UCI          :   3 (+  3,=  0,-  0), 100.0 %
Fruit 2.3                     :   3 (+  0,=  2,-  1), 33.3 %
Naum 2.2                      :   3 (+  2,=  0,-  1), 66.7 %
DeepSjeng27                   :   3 (+  3,=  0,-  0), 100.0 %
Delfi 5.2                     :   3 (+  2,=  1,-  0), 83.3 %
Movei00_8_438                 :   3 (+  3,=  0,-  0), 100.0 %
BugChess2_V1_5_2              :   2 (+  2,=  0,-  0), 100.0 %
Shredder11UCI                 :   2 (+  1,=  0,-  1), 50.0 %
Crafty 21.6 JA                :   2 (+  1,=  1,-  0), 75.0 %
Scorpio 2.0                   :   2 (+  2,=  0,-  0), 100.0 %
AlaricWB707                   :   2 (+  1,=  1,-  0), 75.0 %
Glaurung 2.0.1 JA             :   2 (+  1,=  1,-  0), 75.0 %
Hiarcs11.2SPUCI               :   2 (+  1,=  1,-  0), 75.0 %
Bright-0.2c                   :   2 (+  1,=  1,-  0), 75.0 %
Frenzee Dec 07                :   2 (+  2,=  0,-  0), 100.0 %
Zappa Mexico II               :   2 (+  1,=  0,-  1), 50.0 %
TogaII 1.4 beta 5c            :   2 (+  1,=  1,-  0), 75.0 %

Tony Thomas · Post by **Tony Thomas** » Wed Feb 13, 2008 9:52 am

ozziejoe wrote:Note that strelka is a stolen version of rybka. If tony were to test it, would he not be reinforcing the stealing of engine code? I think it shows alot of character and principle to.....ignore strelka

Similar views here as well..I am a poacher turned game protector, so I cant comment on the stealing part.

Ovyron · Post by **Ovyron** » Wed Feb 13, 2008 11:10 am

ozziejoe wrote:Note that strelka is a stolen version of rybka. If tony were to test it, would he not be reinforcing the stealing of engine code? I think it shows alot of character and principle to.....ignore strelka

Vas claimed Strelka's code as his own, so basically Strelka is still his propriety and nothing is stolen.

Tony Thomas · Post by **Tony Thomas** » Wed Feb 13, 2008 1:27 pm

Ovyron wrote:
ozziejoe wrote:Note that strelka is a stolen version of rybka. If tony were to test it, would he not be reinforcing the stealing of engine code? I think it shows alot of character and principle to.....ignore strelka
Vas claimed Strelka's code as his own, so basically Strelka is still his propriety and nothing is stolen.

Vas was on the right side until he claimed the code as his own.

Tony Thomas · Post by **Tony Thomas** » Thu Feb 14, 2008 4:22 am

My current testing of Naum 3 is finished. It became the first engine to score 100 points in 124 games under my conditions. I know that currently only engine that can top that is probably Rybka. Then again Rybka does lot of impossible things, so it doesnt count. I integrated the engine to my measly 32580 game database and here is the elostat output. I know that my error margin is rather large but as of now it is showing an improvement of 203 points, something unheard of among top engines.

Code: Select all

2/13/2008 10:16:12 PM :

    Program                          Elo    +   -   Games   Score   Av.Op.  Draws

  1 Naum 3                         : 2987   61  59   127    80.7 %   2738   24.4 %
  2 Hiarcs11.2SPUCI                : 2938   59  57   147    75.2 %   2745   15.6 %
  3 Shredder11UCI                  : 2886   50  49   171    69.3 %   2744   21.6 %
  4 HiarcsX54UCI                   : 2884   47  46   208    70.9 %   2729   17.8 %
  5 Fruit 2.3                      : 2868   46  45   187    67.6 %   2740   25.1 %
  6 Fruit 2.3.1                      : 2867   44  43   216    69.0 %   2729   22.2 %
  7 TogaII 1.4 beta 5c             : 2865   60  59   110    66.4 %   2747   23.6 %
  8 Rybka v1.0 Beta.w32            : 2864   40  39   243    67.7 %   2736   25.1 %
  9 Glaurung 2.0.1 JA              : 2856   50  50   151    65.2 %   2747   24.5 %
 10 TogaII 1.2 beta 2a KS/EHP      : 2855   40  40   234    67.1 %   2731   26.5 %
 11 Shredder10UCI Balmung          : 2821   47  46   179    63.7 %   2723   21.2 %

Code: Select all

1 Naum 3                    : 2987  127 (+ 87,= 31,-  9), 80.7 %

BugChess2_V1_5_2              :   4 (+  3,=  1,-  0), 87.5 %
Crafty 21.6 JA                :   4 (+  2,=  2,-  0), 75.0 %
Rybka v1.0 Beta.w32           :   4 (+  3,=  1,-  0), 87.5 %
WildCat 7.0                   :   4 (+  4,=  0,-  0), 100.0 %
Spike 1.2 Turin               :   4 (+  2,=  2,-  0), 75.0 %
Smarthink 1.00                :   4 (+  3,=  1,-  0), 87.5 %
Prodeo 1.2                    :   4 (+  3,=  1,-  0), 87.5 %
Trace 1.37a                   :   4 (+  3,=  1,-  0), 87.5 %
Gandalf 6.01                  :   4 (+  3,=  1,-  0), 87.5 %
Ktulu 8.0                     :   4 (+  2,=  2,-  0), 75.0 %
Thinker 4.7a                  :   4 (+  2,=  2,-  0), 75.0 %
Pharaon 3.5.1                 :   4 (+  3,=  1,-  0), 87.5 %
SOS 5.1                       :   4 (+  4,=  0,-  0), 100.0 %
Ruffian 1.0.5                 :   4 (+  1,=  3,-  0), 62.5 %
SlowChess Blitz WV 2.1        :   4 (+  3,=  1,-  0), 87.5 %
Aristarch 4.50                :   4 (+  3,=  0,-  1), 75.0 %
CM10th D2Alos                 :   4 (+  3,=  0,-  1), 75.0 %
ChessTiger2007.1 UCI          :   4 (+  4,=  0,-  0), 100.0 %
Fruit 2.3                     :   4 (+  0,=  3,-  1), 37.5 %
Naum 2.2                      :   4 (+  3,=  0,-  1), 75.0 %
DeepSjeng27                   :   4 (+  4,=  0,-  0), 100.0 %
Delfi 5.2                     :   4 (+  3,=  1,-  0), 87.5 %
Movei00_8_438                 :   4 (+  4,=  0,-  0), 100.0 %
Shredder11UCI                 :   4 (+  3,=  0,-  1), 75.0 %
Scorpio 2.0                   :   4 (+  4,=  0,-  0), 100.0 %
AlaricWB707                   :   4 (+  2,=  1,-  1), 62.5 %
Glaurung 2.0.1 JA             :   4 (+  1,=  3,-  0), 62.5 %
Hiarcs11.2SPUCI               :   4 (+  1,=  2,-  1), 50.0 %
Bright-0.2c                   :   4 (+  3,=  1,-  0), 87.5 %
Frenzee Dec 07                :   4 (+  4,=  0,-  0), 100.0 %
Zappa Mexico II               :   4 (+  2,=  0,-  2), 50.0 %
TogaII 1.4 beta 5c            :   3 (+  2,=  1,-  0), 83.3 %

Naum results after one cycle + 4 matches

Naum results after one cycle + 4 matches

Rating after 50 games

Re: Rating after 50 games

Re: Rating after 50 games

Re: Rating after 50 games

Re: Rating after 50 games

Re: Rating after 50 games

Re: Rating after 50 games

Re: Rating after 50 games

Re: Rating after 50 games