Last Update Engine:
Houdini 2.0z x64 6c (a new modified version by Salvador)
Its too early for any conclusions,but however the current Houdini 2.0z's performance is just GREAT
Well Done to Salvador/Robert Houdart !
>thanks for your test Sedat
Not at all...its my pleasure dear Salva
Really Bravo...i never noticed before a such performance by other engine !!
im tunning more the engine,since none of my opponents at WBCCC 2 had moved...
the first tuned version,that i lost in a hard drive failure,was about + 30 elo points stronger than Houdini2.0c (in a match between them,2000 games played),but stupid of me,never made a copy of those values in a pen drive,so when the hard drived crashed,i lost all the work
now,im starting from zero,just with that 5 values changed (3 values that i saw one guy called tennison changed in Houdini1.5a,and also queen and pawn values)...the original Houdini2.0z had 12 values changed from Houdini2.0c,so im trying to retreive the work and improve Houdini2.0z by at least another 15 elo points,the points that i lost in the crash
>Really Bravo...i never noticed before a such performance by other engine !!
i lost +15 elo points in the hard drive failure,so im pretty sure i can add the +15 elo points that i lost,i just need some time to test the changes Sedat...
against Houdini2.0c performs about +12 or + 13 elo points,but against other engines perform too much bette,that is why if you run Houdini2.0z in a tournament,(meaning playing against top engines and other weaker engines,not top)it will have better results than playing alone against Houdini2.0c
may i know what are the changes made? thanks
More games have been added (played by Houdini 2.0z Pro x64 6c):
1 Houdini 2.0z Pro x64 6c 3403 25 24 608 78% 3194 32%
2 Houdini 2.0c Pro x64 6c 3373 20 19 945 75% 3190 36%
3 Critter 1.4 x64 6c 3315 18 18 941 68% 3192 45%
Time Control 3min/game +2 Sec/Move
I wonder how it would look like if you confront it with new Engine of Critter ( Critter_1.6 )
And bigger Time Control F.E. 15 min or 1h per game
BTW Great Job with new modification of Houdini 2.0z
Im gonna run a match between Houdini2.0s (an improved Houdini2.0z) as soon as i will have free time to run the 1000 games match
now is 229 games
edited:i forgot,the match is between Houdini2.0s against Houdini2.0z
I expect that the Houdini 2.0z will be stabilized around + 30 elo points stronger than Houdini2.0c...time will tell if im right or not
>I expect that the Houdini 2.0z will be stabilized around + 30 elo points stronger than Houdini2.0c...time will tell if im right or not
I am not sure exactly about what will be the Elo difference between v2.0c and v2.0z
For example,after 1000 games will be more clear....but anyway it seems you created successful settings
yes,you are right,after 1000 game swill be more clear;if one engine doesnt run enough games,you can have that,for example,the 1 time i ran the match Houdini2.0z against Houdini2.0c,around the first 50 games,Houdini2.0z had an elo of -3 down Houdini2.0c,but after 1000 games,the elo was stabilized for + 13 against Houdini2.0c (running a match Houdini2.0z vs Houdini2.0c;if you run a tournament like you are running,Houdini2.0z performs better than 13 elo points because it performs very well against some engines (for example,Komodo4 and Stockfish)
Powered by mwForum 2.27.4 © 1999-2012 Markus Wichitill