FCP: Final results from my 10.500 games FCT1-2015 Basic-T.

[Frank Quisinsky]

Hi there,

here are the final results!

10.500 45-minutes games on i7 4.3Ghz are clear for take off on runway 2.00.s2, wind is good.

6 engine updates and one new engine tested in the same tourney with a new Basic-Tourney I need for my test work in the year 2015. An all in one start point for myself.

Code: Select all

     Programs                      Score       1-0  =  0-1         EloS  Baye  Ordo  +   -  Draws  MoveØ won<50 lost<50
 01. Chiron 2.0 x64                693.0/1000  499 388 113  69.3%  2884  2877  2887  17 17  38.8%   87     13       1
 02. Hannibal 1.4b x64             651.5/1000  441 421 138  65.1%  2853  2849  2854  17 17  42.1%   91      6       1
 03. Texel 1.04 x64                637.5/1000  425 425 150  63.7%  2843  2840  2844  17 16  42.5%   88      9       3
 04. Nirvanachess 2.0a x64         626.5/1000  407 439 154  62.6%  2835  2833  2835  16 16  43.9%   86     15       0
 05. Hiarcs 14 WCSC w32            612.5/1000  402 421 177  61.3%  2825  2825  2825  17 16  42.1%   87     18       7
 06. Senpai 1.0 SSE42 x64          611.0/1000  393 436 171  61.1%  2824  2823  2824  16 16  43.6%   91      5      15
 07. Junior 13.3.00 x64            533.0/1000  327 412 261  53.3%  2772  2777  2768  17 17  41.2%   84     22       6
 08. Deuterium 14.3.34.130 POP x64 513.5/1000  292 443 265  51.3%  2759  2764  2754  16 16  44.3%   85      7      13
 09. iCE 2.0 v2240 POP x64         500.5/1000  271 459 270  50.0%  2750  2754  2745  16 16  45.9%   89     10       4
 10. Spark 1.0 x64                 498.5/1000  275 447 278  49.9%  2749  2754  2744  16 16  44.7%   89     18       1
 11. Fizbo 1.3.1 x64               484.5/1000  271 427 302  48.4%  2739  2747  2734  16 16  42.7%   82      7       7
 12. SmarThink 1.70 SSE3 x64       481.0/1000  274 414 312  48.1%  2737  2743  2731  16 17  41.4%   93      3       9
 13. Vajolet2 1.45 POP x64         470.0/1000  256 428 316  47.0%  2730  2737  2724  16 16  42.8%   88      1       8
 14. Gaviota 1.0 AVX x64           458.5/1000  258 401 341  45.9%  2722  2729  2715  17 17  40.1%   88      3      13
 15. Andscacs 0.70 POP x64         425.5/1000  215 421 364  42.5%  2700  2709  2692  16 16  42.1%   86      3      15
 16. Cheng4 0.36c x64              410.5/1000  195 431 374  41.0%  2690  2699  2681  16 16  43.1%   91      9       8
 17. Arasan 17.4 POP x64           396.0/1000  186 420 394  39.6%  2680  2691  2671  16 17  42.0%   91      5       9
 18. EXchess 7.51b x64             391.5/1000  193 397 410  39.1%  2677  2687  2667  17 17  39.7%   88     11       6
 19. DisasterArea 1.54 x64         390.0/1000  176 428 396  39.0%  2676  2687  2666  16 16  42.8%   93      2       7
 20. Rodent 1.6 Build 6 POP x64    374.0/1000  179 390 431  37.4%  2664  2676  2655  17 17  39.0%   88      2      19
 21. Crafty 24.1 SSE42 x64         341.0/1000  153 376 471  34.1%  2641  2654  2630  17 17  37.6%   86      7      24

1.
"Download games by Players" selection updated.

2.
FCP database updated (complete database with additional files).

3.
FCT1 2015 detail page updated
http://www.amateurschach.de/main/_fct1-2015.htm

4.
ToDo Calendar updated
Have a look in my News 134 too.

In the next hours I will added results for middelgame, early endgame and endgame. Need a bit more time, I will added it later ...

Have fun!

Best
Frank

[Frank Quisinsky]

Latest update to middlegame, endgame and so on will be available tomorrow. To late and I am sitting to Long on my PC.

Later, tomorrow, sorry for my last message!

Best
Frank

[carldaman]

Thanks for your tests Frank!

Is there a way to just download these completed 10,500 games?

Thanks,
CL

Hi there,

here are the final results!

10.500 45-minutes games on i7 4.3Ghz are clear for take off on runway 2.00.s2, wind is good.

6 engine updates and one new engine tested in the same tourney with a new Basic-Tourney I need for my test work in the year 2015. An all in one start point for myself.

Code: Select all

     Programs                      Score       1-0  =  0-1         EloS  Baye  Ordo  +   -  Draws  MoveØ won<50 lost<50
 01. Chiron 2.0 x64                693.0/1000  499 388 113  69.3%  2884  2877  2887  17 17  38.8%   87     13       1
 02. Hannibal 1.4b x64             651.5/1000  441 421 138  65.1%  2853  2849  2854  17 17  42.1%   91      6       1
 03. Texel 1.04 x64                637.5/1000  425 425 150  63.7%  2843  2840  2844  17 16  42.5%   88      9       3
 04. Nirvanachess 2.0a x64         626.5/1000  407 439 154  62.6%  2835  2833  2835  16 16  43.9%   86     15       0
 05. Hiarcs 14 WCSC w32            612.5/1000  402 421 177  61.3%  2825  2825  2825  17 16  42.1%   87     18       7
 06. Senpai 1.0 SSE42 x64          611.0/1000  393 436 171  61.1%  2824  2823  2824  16 16  43.6%   91      5      15
 07. Junior 13.3.00 x64            533.0/1000  327 412 261  53.3%  2772  2777  2768  17 17  41.2%   84     22       6
 08. Deuterium 14.3.34.130 POP x64 513.5/1000  292 443 265  51.3%  2759  2764  2754  16 16  44.3%   85      7      13
 09. iCE 2.0 v2240 POP x64         500.5/1000  271 459 270  50.0%  2750  2754  2745  16 16  45.9%   89     10       4
 10. Spark 1.0 x64                 498.5/1000  275 447 278  49.9%  2749  2754  2744  16 16  44.7%   89     18       1
 11. Fizbo 1.3.1 x64               484.5/1000  271 427 302  48.4%  2739  2747  2734  16 16  42.7%   82      7       7
 12. SmarThink 1.70 SSE3 x64       481.0/1000  274 414 312  48.1%  2737  2743  2731  16 17  41.4%   93      3       9
 13. Vajolet2 1.45 POP x64         470.0/1000  256 428 316  47.0%  2730  2737  2724  16 16  42.8%   88      1       8
 14. Gaviota 1.0 AVX x64           458.5/1000  258 401 341  45.9%  2722  2729  2715  17 17  40.1%   88      3      13
 15. Andscacs 0.70 POP x64         425.5/1000  215 421 364  42.5%  2700  2709  2692  16 16  42.1%   86      3      15
 16. Cheng4 0.36c x64              410.5/1000  195 431 374  41.0%  2690  2699  2681  16 16  43.1%   91      9       8
 17. Arasan 17.4 POP x64           396.0/1000  186 420 394  39.6%  2680  2691  2671  16 17  42.0%   91      5       9
 18. EXchess 7.51b x64             391.5/1000  193 397 410  39.1%  2677  2687  2667  17 17  39.7%   88     11       6
 19. DisasterArea 1.54 x64         390.0/1000  176 428 396  39.0%  2676  2687  2666  16 16  42.8%   93      2       7
 20. Rodent 1.6 Build 6 POP x64    374.0/1000  179 390 431  37.4%  2664  2676  2655  17 17  39.0%   88      2      19
 21. Crafty 24.1 SSE42 x64         341.0/1000  153 376 471  34.1%  2641  2654  2630  17 17  37.6%   86      7      24

1.
"Download games by Players" selection updated.

2.
FCP database updated (complete database with additional files).

3.
FCT1 2015 detail page updated
http://www.amateurschach.de/main/_fct1-2015.htm

4.
ToDo Calendar updated
Have a look in my News 134 too.

In the next hours I will added results for middelgame, early endgame and endgame. Need a bit more time, I will added it later ...

Have fun!

Best
Frank

[Frank Quisinsky]

Hi Carl,

I added the games again in the detail pages (under "Final results").

1. 10.500 games from Qualification-2014
http://www.amateurschach.de/main/_quali ... n-2014.htm

2. 10.500 games from FCT-2015
http://www.amateurschach.de/main/_fct1-2015.htm

Enough to this "Horror tourneys"!
Now "round-robin test-runs" will be follow!

Best
Frank

[Ferdy]

Thanks Frank. It is nice to review my games

[Frank Quisinsky]

Hi Ferdinand,

no problem!
Thanks to have your scripts and engine!

Working on the Detail pages to the stats at the moment ...
Very interesting what I can see ...

Later ...

Best
Frank

[Frank Quisinsky]

Code: Select all

 *****************************************************************************************
 
 FCT1 (Frank's Continuous Tourney 1), games with move number 020-059
 Elostat 1.3,   EloØ = 2.750, Category = 20

     Programs                         Elo   +   - Games Score Av.Op. Draws   Won Draw Lost
 01. Nirvanachess 2.0a x64          : 2852  29 25  172  64.8%  2746  69.2%    52  119    1
 02. Chiron 2.0 x64                 : 2842  35 33  146  63.4%  2747  62.3%    47   91    8
 03. Texel 1.04 x64                 : 2829  37 36  149  62.1%  2744  57.0%    50   85   14
 04. Junior 13.3.00 x64             : 2829  37 37  163  62.0%  2744  51.5%    59   84   20
 05. Spark 1.0 x64                  : 2802  30 28  159  58.5%  2743  70.4%    37  112   10
 06. Hiarcs 14 WCSC w32             : 2799  34 33  161  57.5%  2746  60.2%    44   97   20
 07. iCE 2.0 v2240 POP x64          : 2782  35 34  140  53.6%  2757  64.3%    30   90   20
 08. Fizbo 1.3.1 x64                : 2772  31 30  160  54.4%  2741  67.5%    33  108   19
 09. EXchess 7.51b x64              : 2763  36 36  141  52.1%  2748  61.7%    30   87   24
 10. Senpai 1.0 SSE42 x64           : 2754  33 33  135  48.9%  2762  68.1%    20   92   23
 11. Hannibal 1.4b x64              : 2748  36 36  110  50.0%  2748  69.1%    17   76   17
 12. SmarThink 1.70 SSE3 x64        : 2747  42 42  118  49.6%  2750  55.1%    26   65   27
 13. Deuterium 14.3.34.130 POP x64  : 2746  32 32  164  50.0%  2746  64.6%    29  106   29
 14. Cheng4 0.36c x64               : 2725  38 38  125  46.8%  2747  61.6%    20   77   28
 15. Vajolet2 1.45 POP x64          : 2722  35 35  144  44.8%  2758  61.8%    20   89   35
 16. Arasan 17.4 POP x64            : 2717  39 39  130  45.4%  2749  56.9%    22   74   34
 17. Andscacs 0.70 POP x64          : 2685  35 36  158  38.0%  2770  57.0%    15   90   53
 18. Gaviota 1.0 AVX x64            : 2678  35 36  145  40.3%  2746  60.0%    15   87   43
 19. DisasterArea 1.54 x64          : 2672  36 38  125  39.2%  2748  62.4%    10   78   37
 20. Rodent 1.6 Build 6 POP x64     : 2634  44 45  129  32.6%  2761  46.5%    12   60   57
 21. Crafty 24.1 SSE42 x64          : 2614  38 39  178  30.6%  2757  44.4%    15   79   84

 Games         : 1526
 White Perf.   : 54.8%
 Black Perf.   : 45.2%
 White Wins    :  374 Games (24.5%)
 Black Wins    :  229 Games (15.0%)
 Draws         :  923 Games (60.5%)

 ECO A00-A99   :  289 Games (18.9%)
 ECO B00-B99   :  348 Games (22.8%)
 ECO C00-C99   :  302 Games (19.8%)
 ECO D00-D99   :  324 Games (21.2%)
 ECO E00-E99   :  263 Games (17.2%)
 
 ***************************************************************************************** 
 
 FCT1 (Frank's Continuous Tourney 1), games with move number  060-099
 Elostat 1.3,   EloØ = 2.750, Category = 20

     Programs                         Elo   +   - Games Score Av.Op. Draws   Won Draw Lost
 01. Chiron 2.0 x64                 : 2928  27 26  601  75.0%  2737  24.3%   378  146   77
 02. Hannibal 1.4b x64              : 2880  25 25  611  69.3%  2739  25.4%   346  155  110
 03. Texel 1.04 x64                 : 2873  25 25  583  68.0%  2742  25.9%   321  151  111
 04. Nirvanachess 2.0a x64          : 2854  25 25  569  66.0%  2739  26.2%   301  149  119
 05. Senpai 1.0 SSE42 x64           : 2853  25 25  577  66.1%  2737  26.9%   304  155  118
 06. Hiarcs 14 WCSC w32             : 2851  26 25  580  65.0%  2744  23.4%   309  136  135
 07. Junior 13.3.00 x64             : 2755  23 23  610  50.7%  2750  31.8%   212  194  204
 08. Deuterium 14.3.34.130 POP x64  : 2754  23 23  596  51.0%  2747  33.2%   205  198  193
 09. iCE 2.0 v2240 POP x64          : 2753  24 24  574  50.7%  2748  30.7%   203  176  195
 10. Fizbo 1.3.1 x64                : 2745  22 22  643  49.5%  2749  31.7%   216  204  223
 11. Spark 1.0 x64                  : 2737  24 24  561  47.7%  2753  28.0%   189  157  215
 12. Vajolet2 1.45 POP x64          : 2733  24 24  572  47.6%  2749  27.1%   195  155  222
 13. Gaviota 1.0 AVX x64            : 2729  25 25  585  46.2%  2755  24.4%   199  143  243
 14. SmarThink 1.70 SSE3 x64        : 2720  25 25  570  45.9%  2749  25.8%   188  147  235
 15. Andscacs 0.70 POP x64          : 2696  23 23  607  41.9%  2753  29.5%   165  179  263
 16. Cheng4 0.36c x64               : 2667  25 25  567  37.2%  2758  25.4%   139  144  284
 17. Rodent 1.6 Build 6 POP x64     : 2661  24 24  600  36.9%  2754  26.5%   142  159  299
 18. DisasterArea 1.54 x64          : 2643  25 25  572  34.4%  2756  26.4%   121  151  300
 19. EXchess 7.51b x64              : 2638  26 26  582  33.7%  2756  22.3%   131  130  321
 20. Arasan 17.4 POP x64            : 2636  26 26  562  32.8%  2760  23.7%   118  133  311
 21. Crafty 24.1 SSE42 x64          : 2629  26 26  572  32.2%  2758  23.4%   117  134  321

 Games         : 6147
 White Perf.   : 56.5%
 Black Perf.   : 43.5%
 White Wins    : 2652 Games (43.1%)
 Black Wins    : 1847 Games (30.0%)
 Draws         : 1648 Games (26.8%)

 ECO A00-A99   : 1121 Games (18.2%)
 ECO B00-B99   : 1377 Games (22.4%)
 ECO C00-C99   : 1315 Games (21.4%)
 ECO D00-D99   : 1215 Games (19.8%)
 ECO E00-E99   : 1119 Games (18.2%)

 *****************************************************************************************
 
 FCT1 (Frank's Continuous Tourney 1), games with move number 100-299
 Elostat 1.3,   EloØ = 2.750, Category = 20

 Programs                         Elo   +   - Games Score Av.Op. Draws   Won Draw Lost
 01. Hannibal 1.4b x64              : 2832  23 22  279  62.0%  2747  68.1%    78  190   11
 02. Chiron 2.0 x64                 : 2815  27 27  253  59.1%  2751  59.7%    74  151   28
 03. Senpai 1.0 SSE42 x64           : 2796  24 23  288  56.8%  2749  65.6%    69  189   30
 04. Texel 1.04 x64                 : 2787  23 22  268  55.4%  2749  70.5%    54  189   25
 05. Hiarcs 14 WCSC w32             : 2786  22 22  259  55.2%  2750  72.6%    49  188   22
 06. Nirvanachess 2.0a x64          : 2779  25 25  259  53.9%  2753  66.0%    54  171   34
 07. Junior 13.3.00 x64             : 2778  29 29  227  54.2%  2749  59.0%    56  134   37
 08. Deuterium 14.3.34.130 POP x64  : 2773  29 28  240  53.1%  2752  57.9%    58  139   43
 09. SmarThink 1.70 SSE3 x64        : 2760  23 23  312  51.6%  2749  64.7%    60  202   50
 10. Arasan 17.4 POP x64            : 2745  22 22  308  49.5%  2749  69.2%    46  213   49
 11. Spark 1.0 x64                  : 2743  25 25  280  49.3%  2748  63.6%    49  178   53
 12. DisasterArea 1.54 x64          : 2734  23 23  303  47.7%  2750  65.7%    45  199   59
 12. Gaviota 1.0 AVX x64            : 2734  25 25  270  48.0%  2748  63.3%    44  171   55
 14. Andscacs 0.70 POP x64          : 2730  26 27  235  47.2%  2749  64.7%    35  152   48
 15. Vajolet2 1.45 POP x64          : 2729  24 24  284  46.8%  2751  64.8%    41  184   59
 16. iCE 2.0 v2240 POP x64          : 2728  23 23  286  47.0%  2749  67.5%    38  193   55
 17. Cheng4 0.36c x64               : 2719  22 22  308  45.8%  2748  68.2%    36  210   62
 18. EXchess 7.51b x64              : 2712  24 24  277  44.0%  2753  65.0%    32  180   65
 19. Rodent 1.6 Build 6 POP x64     : 2688  25 25  271  40.8%  2753  63.1%    25  171   75
 19. Fizbo 1.3.1 x64                : 2688  31 31  197  40.4%  2755  58.4%    22  115   60
 21. Crafty 24.1 SSE42 x64          : 2687  25 25  250  41.0%  2750  65.2%    21  163   66

 Games         : 2827
 White Perf.   : 51.6%
 Black Perf.   : 48.4%
 White Wins    :  537 Games (19.0%)
 Black Wins    :  449 Games (15.9%)
 Draws         : 1841 Games (65.1%)

 ECO A00-A99   :  517 Games (18.3%)
 ECO B00-B99   :  598 Games (21.2%)
 ECO C00-C99   :  593 Games (21.0%)
 ECO D00-D99   :  589 Games (20.8%)
 ECO E00-E99   :  530 Games (18.7%)
 
 *****************************************************************************************

[Frank Quisinsky]

Hi there,

if we are looking in the stats we can't see bigger differents to tested engines.

Elo Move-Number 020-059
http://www.amateurschach.de/main/060/_e ... _v2.00.htm

Elo Move-Number 060-099
http://www.amateurschach.de/main/100/_e ... _v2.00.htm

Elo Move-Number 100-299
http://www.amateurschach.de/main/300/_e ... _v2.00.htm

Interesting are the differents of the latest 6 engine updates to the predecessors.

1. Deuterium 14.3.34.130 POP x64
Clearly improved in king safty without to lost the endgame strength.

2. Nirvanachess 2.0b x64
Sensational middlegame stats. Lost only 1 game up to move number 59 and won a lot. The actual number 1 (middlegame) in FCT1 and here stronger as Chiron / Texel! Very interesting for analyzes after opening book moves.

3. EXchess 7.51b x64
Very nice style in the beginning of the game. Won here a lot of fast games too and is clearly stronger as with lesser pieces on the board.

4. Arasan 17.4 POP x64 & DisasterArea 1.54 x64
Very solid in all phases.

5. Fizbo 1.3.1 x64
Seems to be clearly improved in transposition endgame. I think Fizbo 1.3.1. lost a bit the attacking style, version 1.2 have. Not to 100% sure here.

6. Andscacs 0.70 POP x64
Great improvements and also better in kings safty as the last tested Version 0.64. Start of next week I will test the latest version of Andscacs.

Stats problems:
Cheng4 0.36c x64 played FCT-2015 clearly stronger as the Qualification-2014 tournament. Maybe Cheng is stronger vs. stronger opponents. After all I saw here possible, but for such an opinion the combination of games / opponents between both tourneys are not enough. Very difficult.

Long-shot guess:
Cheng could be better vs. stronger opponents!
Same for Vajolet2 I believe! Here I am sure because Vajolet is very positional and snappy, not important how strong are the opponents.

All in all:
Nirvanachess is great for analyzes after opening book moves. Same for iCE from this group of engines and for sure Junior and Spark. Could be good tips for correspondence chess players. EXchess a bit speculative and produced much surprises in the beginning of the games.

Eval of iCE after opening book moves is very interesting, wrote this before. Niravanchess in Version 2.0b is a secret TIP for correspondence players!

Have fun with my site.
All selection after my Horror FCT-2015 and Horror Qualification-2014 tourneys are updated now. Now with full power I have to test 7 engine updates.

Best
Frank

Have a look in the stats ... engines 2.650 - 2.900 Elo are in the endgame 150 Elo wide apart. But in the Transposition endgame around 300. Here the most engine-engine games comes with the decision. More blatant such stats with engines stronger as 2.900 Elo.

If we are speaking from a Rating List of engines and Elo we are speaking most of the time from this playing phase. Stats from humans are completly different. Here the middlegame / Transposition endgame and Elo differents betweens this phases are around the same.

[Frank Quisinsky]

Means with my latest comments that engines, stronger in middlegame could be have a clearly higher Elo (humens view). Most humans have here the most problems and engines with such strengths could be clearly stronger as we can see in a computer chess rating list.

Example:
NirvanaChess 2.0b in a eng-eng Rating list with ~ 2.850 Elo could be 2.950 Elo vs. humans.

The bad Point in each Rating List for Computer chess engines.
So many games with a decision in the late endgame never we can find in GM databases. So, most improvements for Computer Chess engines in the latest years are to see in the early endgame / Transposition endgame. And for such reason is Human Elo not to compare with Computer chess Elo. For me one reason more to give a bit more Information to my list because such things are nice to see if we are looking more in detail. Elo alone is very boring in Computer chess.

Best
Frank