FCP: Komodo 10.2 x64 C15 is still running vs. 49 opp.

Discussion of computer chess matches and engine tournaments.

Moderator: Ras

Frank Quisinsky
Posts: 7520
Joined: Wed Nov 18, 2009 7:16 pm
Location: Gutweiler, Germany
Full name: Frank Quisinsky

Re: Downloads ...

Post by Frank Quisinsky »

Hi Ray,

the different is a bit higher as 8 Elo if we compare 10.2 to 10.1.
I don't know why but Version 10.1 is around 10 Elo weaker as Version 10.1 here. Mayby more optimated for faster time controls.

Thanks for your comments to that work.

Downloads can be found here:
http://www.amateurschach.de/main/_download.htm

or here ...
Download games by players ...
http://www.amateurschach.de/main/_sgbp.htm

or the cross table by Ed here ...
http://www.amateurschach.de/main/cross-tab/v440.htm

But please wait with your download ...
I am working, the new database version isn't ready so far.

Note:
I have a very big problem here ...
Since v4.39 the database with game comments is higher as 2Gb. So in the database download (around 475Mb) I must create two *.pgn files. I will not deleate the game comments (eval, ply, time) because the game comments are important for opening book creators.

In the database download can be found all stats since I started the work, all fast draw games undo move 20 I replayed and all games over 300 moves I replayed and much more ...

The new database version is ~ 2 hours available (v4.40).

Best
Frank
Frank Quisinsky
Posts: 7520
Joined: Wed Nov 18, 2009 7:16 pm
Location: Gutweiler, Germany
Full name: Frank Quisinsky

Re: Comparsion ... Komodo in endgame Nr.1

Post by Frank Quisinsky »

Hi there,

here the comparsion ...
Komodo hold place 1 in endgames!!
OK, the question is now what ASM Fish 8 BMI2 x64 C15 can do.
Later ...

In endgames Komodo is again stronger ... nice to see on which gaming phases the programmers of Komodo are most working since the last version.

I can see it Mark & Larry ...
Please give no other comments or the others will never looking on my move stats ... OK?!

:-)

Also nice to see that Contempt = 15 is a bit stronger as Contempt = 10. Can be see with quantity of draw games between move number 20-59. This one comes not allone from the stronger Komodo version.

Larry are right with Contempt = 15 and an user here wrote to my Contempt = 10 experiment and tested out in own games with Contempt = 15.

The complete move stats can be found in menue system of my site.
Click on ... Move Nr. 020-059, 060-079, 080-099, 100-299

Best
Frank


C O M P A R S I O N

Code: Select all

Final results
-------------

   # Player                           :      Elo  Games  Score%   won  draw  lost  Points  Draw%  Error   OppAvg   OppE   OppD
  ----------------------------------------------------------------------------------------------------------------------------
   2 Komodo 10.2 x64 C15              :  3204.46   2450    86.6  1818   607    25  2121.5   24.8  14.26  2845.96  10.48   49.0
   2 Komodo 10 x64                    :  3195.98   2450    86.1  1802   614    34  2109.0   25.1  14.47  2845.45  10.49   49.0
     

	 
		 
Move number 020-059
-------------------
   2 Komodo 10.2 x64 C15              :  3143.14    369    85.1   260   108     1   314.0   29.3  31.75  2824.21  20.46   42.9
   2 Komodo 10 x64                    :  3121.69    398    83.5   271   123     4   332.5   30.9  29.59  2823.82  20.73   44.0

   
   
Move number 060-079
-------------------
   2 Komodo 10.2 x64 C15              :  3325.49   1038    92.3   887   143     8   958.5   13.8  31.13  2818.13  22.97   47.6
   2 Komodo 10 x64                    :  3314.17   1059    92.2   906   141    12   976.5   13.3  30.68  2817.39  22.92   47.9
   
   

Move number 080-099
-------------------
   2 Komodo 10.2 x64 C15              :  3297.53    596    90.4   486   105     5   538.5   17.6  38.46  2845.13  25.16   47.2
   2 Komodo 10 x64                    :  3295.81    580    89.3   467   102    11   518.0   17.6  39.21  2857.02  25.40   46.8
   
   
   
Move number 100-299
-------------------
   1 Komodo 10.2 x64 C15              :  3019.77    447    69.5   185   251    11   310.5   56.2  20.32  2870.45  14.95   43.2
   1 Komodo 10 x64                    :  3010.56    413    68.3   158   248     7   282.0   60.0  21.51  2870.27  14.94   43.5
Best
Frank
Modern Times
Posts: 3902
Joined: Thu Jun 07, 2012 11:02 pm

Re: Comparsion ... Komodo in endgame Nr.1

Post by Modern Times »

Well Frank do you see CEGT results:

Komodo 10.2 x64 1CPU: 3258 (+-18) - 1161 games (-13 to v. 10.1)
Frank Quisinsky
Posts: 7520
Joined: Wed Nov 18, 2009 7:16 pm
Location: Gutweiler, Germany
Full name: Frank Quisinsky

Re: Comparsion ... Komodo in endgame Nr.1

Post by Frank Quisinsky »

Ray,

looking a longer time not on CEGT 40 in 20 ...

Just wait ...

Hm ...
Not enough opponents, not enough games, and time controls is 3 times slower as in my list.

For engines so strong Komodo is you need more games and much more opponents. Have a look what I have with 49 opp. for Komodo 10.2 results after 26% and after 64% of games.

Results with 49 opp. after 2.450 games are absolutely clear and the stats too I created.

I believe that Komodo 10.1 is a version more for faster time controls, CEGT 40 in 20 used. Sure that 10.1 rating goes down with more opponents. The problem CEGT 40 / 20 list have ... not enough opponents.

Much more interesting is CEGT 5 +3.

Best
Frank

CEGT will have very fast results.
More exactly is more interesting as fast an inaccurate!
Very hard but fact.

Biggest problem here ...
Users can't wait!
Frank Quisinsky
Posts: 7520
Joined: Wed Nov 18, 2009 7:16 pm
Location: Gutweiler, Germany
Full name: Frank Quisinsky

Re: Ready ...

Post by Frank Quisinsky »

Hi Ray,

now online ...
If you like you can start directly the download here, so have not to jump to my site:

Download:
http://www.amateurschach.de/download/_f ... -files.zip (Note: It is: 478.09Mb).

Have fun!

Best
Frank