SPCC: Testrun of Rebel Extreme 1.1 finished

[pohl4711]

My UHO-Top15 Ratinglist is the world's first engine-ratinglist, using UHO-openings, and the world's first ratinglist offering additionally Gamepair-statistics.

Rebel Extreme 1.1 was released by Ed Schroeder. Ed's own testings show very impressive EAS-scores of this new release. Testrun for my Full UHO Ratinglist finished (of course, Rebel Extreme 1.1 is way too weak for the UHO-Top15 Ratinglist).

Rebel Extreme 1.1 lost -86 Celo, compared to Rebel Extreme 1.0 in the testrun, but gained a lot of EAS-points (+68%) and is now (very close) behind Patricia 3.1 on rank #2 in my full EAS Ratinglist.

https://www.sp-cc.de/files/uho_full_list.txt

Code: Select all

     Program                     Celo    +    - Games    Score   Av.Op. Draws
 135 Patricia 5.0 avx512       : 3466    4    4 14000    42.4%   3520   44.9%
 136 Slow Chess 2.9 avx2       : 3449    3    3 27000    43.6%   3499   42.8%
 137 Cerberus 21124081r81      : 3447    4    4 14000    39.9%   3520   37.4%
 138 Patricia 250510 a512      : 3441    4    4 14000    39.0%   3520   43.0%
 139 Rebel Extreme avx2        : 3402    4    4 14000    34.0%   3520   39.3% (OLD)
 140 Patricia 4.0 avx2         : 3381    4    4 14000    31.4%   3520   32.2%
 141 Revenge 1.0 avx2          : 3356    6    6 15000    18.3%   3628   30.3%
 142 CSTal 2.1 EAS             : 3343    5    5 14000    26.9%   3520   29.6%
 143 Monty 241220 a512         : 3321    5    5 14000    24.6%   3520   35.4%
 144 Rebel Extreme 1.1         : 3316    5    5 14000    24.0%   3520   25.4% (NEW)
 145 Patricia 3.1 avx2         : 3210    5    5 14000    14.7%   3520   21.1%
 

Here the Top of my Full EAS Ratinglist:

Code: Select all

                                 bad  avg.win 
Rank  EAS-Score  sacs   shorts  draws  moves  Engine/player 
-------------------------------------------------------------------
   1    429481  51.61%  38.03%  05.49%   66   Patricia 3.1 avx2  
   2    419291  52.40%  44.25%  08.28%   65   Rebel Extreme 1.1 (NEW) 
   3    378256  50.51%  36.25%  06.08%   67   Patricia 250510 a512  
   4    376711  50.62%  36.92%  07.20%   68   CSTal 2.1 EAS  
   5    371208  50.79%  36.98%  05.85%   67   Patricia 5.0 avx512  
   6    347375  48.39%  32.33%  05.82%   67   Cerberus 21125081r9b  
   7    337432  46.96%  29.39%  03.46%   70   Patricia 4.0 avx2  
   8    312346  43.56%  27.63%  04.21%   68   Cerberus 21124081r81  
   9    250682  37.88%  23.88%  08.79%   75   Rebel Extreme avx2 (OLD) 

https://www.sp-cc.de

Also take a look at the EAS-Ratinglist, the world's first engine-ratinglist not measuring strength of engines but engines's style of play:
https://www.sp-cc.de/eas-ratinglist.htm

As usual, the most spectacular sac-wins of the latest testrun (won by the tested engine) can be replayed directly on my website. The pgn-viewer (by ChessBase) needs Javascript on and Adblockers off:
https://www.sp-cc.de/view-games-with-sacs.htm

(Perhaps you have to clear your browsercache with <strg>+<shift>+<delete> to reload the graphics/diagrams on my website)

[Rebel]

My UHO-Top15 Ratinglist is the world's first engine-ratinglist, using UHO-openings, and the world's first ratinglist offering additionally Gamepair-statistics.

Rebel Extreme 1.1 was released by Ed Schroeder. Ed's own testings show very impressive EAS-scores of this new release. Testrun for my Full UHO Ratinglist finished (of course, Rebel Extreme 1.1 is way too weak for the UHO-Top15 Ratinglist).

Rebel Extreme 1.1 lost -86 Celo, compared to Rebel Extreme 1.0 in the testrun, but gained a lot of EAS-points (+68%) and is now (very close) behind Patricia 3.1 on rank #2 in my full EAS Ratinglist.

Thanks for testing, but I still think you have some work to do.

Here the Top of my Full EAS Ratinglist:

Code: Select all

                                 bad  avg.win 
Rank  EAS-Score  sacs   shorts  draws  moves  Engine/player 
-------------------------------------------------------------------
   1    429481  51.61%  38.03%  05.49%   66   Patricia 3.1 avx2  
   2    419291  52.40%  44.25%  08.28%   65   Rebel Extreme 1.1 (NEW) 

This is what I get :

Code: Select all

1    411481  51.61%  38.03%  05.49%   66   Patricia 3.1 avx2
1    413291  52.40%  44.25%  08.28%   65   Rebel Extreme 1.1  

Percentages are exactly the same, so far so good.

But then the final EAS calculations are different.

Increase shortwins=60 even more ?

[pohl4711]

Percentages are exactly the same, so far so good.

But then the final EAS calculations are different.

Increase shortwins=60 even more ?

No, all fine here, the hardcoded shortwins value work as expected. The hardcoded shortwins value was made to get exactly this, an comparable EAS-scoring, when using the EAS-tool on a full ratinglist-gamebase and when using the tool on just a Gauntlet-testrun of one engine.
Small differences in the scoring are unavoidable for 2 reasons:
1) When the EAS-Tool calculates the shortwin limit, this limit is set to the next 5 or 10: 60,55,50,45,40 ,but never 62, 57, 52, 47, 42...
2) There is a bonus-point system for the overall length of won games for each engine:
"Additionally, if the average win game length of the engine is shorter than the average win game length of all games in the source.pgn, the engine gets 3000 EAS-points for each move, their won games are shorter in average. If the average win game length of the engine is higher than the average win game length of all games in the source.pgn, 1000 EAS-points are substracted for each move, their won games are shorter in average. But these substraction of points is done only on the EAS-points, the engine has received for their short wins (see above). The other EAS-points (for sacrifices and bad draws (see 1) and 3)) stay always in the calculation!"
This can also lead to some thousand EAS-points more or less, because the overall length of all won games in a gamebase can be different in a full ratinglist gamebase comapred to a Gauntlet-gamebase... (But I am thinking over it, to perhaps remove this idea from the new V6.0 of the EAS-Tool).