I have just played a match of 50 games, time control 5+0.
Rybka was using 4-core, contempt 0, VS Robbo's single core.
I am proud to announce that Rybka on 4-core is 0 points better than
Robbo on 1-core.
Here are the games
http://www.spaadyshare.net/362m811zx1um/Robbolito 0.09 New Edition
VS Rybka 3.zip
Have a nice 2010 everyone
Damir
Robbolito 0.09 New Edition VS Rybka 3
Moderators: hgm, Rebel, chrisw
-
- Posts: 2801
- Joined: Mon Feb 11, 2008 3:53 pm
- Location: Denmark
- Full name: Damir Desevac
-
- Posts: 660
- Joined: Sat Dec 05, 2009 5:13 am
- Location: Colorado, USA
Re: Robbolito 0.09 New Edition VS Rybka 3
Where did you get a version 0.09?
Peter
Peter
-
- Posts: 2801
- Joined: Mon Feb 11, 2008 3:53 pm
- Location: Denmark
- Full name: Damir Desevac
Re: Robbolito 0.09 New Edition VS Rybka 3
This version is private, not supposed to be released yet. I am testing it.
Damir
Damir
-
- Posts: 660
- Joined: Sat Dec 05, 2009 5:13 am
- Location: Colorado, USA
Re: Robbolito 0.09 New Edition VS Rybka 3
Cool. So R0.09 1-cpu is equal with Rybka 3 4-cpu?! Impressive.Damir wrote:This version is private, not supposed to be released yet. I am testing it.
Damir
-
- Posts: 2801
- Joined: Mon Feb 11, 2008 3:53 pm
- Location: Denmark
- Full name: Damir Desevac
Re: Robbolito 0.09 New Edition VS Rybka 3
Looks like it, from the result I just posted. Robbolito was leading by 5 points most of the match. In the end however Rybka was lucky to equalize the match....
Re: Robbolito 0.09 New Edition VS Rybka 3
This would be something...the public released version of robbo* I have tested is only slightly better than rybka on a 1 core vs 2 cores. I use contempt of 15 (default).Damir wrote:I have just played a match of 50 games, time control 5+0.
Rybka was using 4-core, contempt 0, VS Robbo's single core.
I am proud to announce that Rybka on 4-core is 0 points better than
-
- Posts: 373
- Joined: Thu Aug 14, 2008 3:21 am
- Location: Albuquerque, NM
Re: Robbolito 0.09 New Edition VS Rybka 3
Am I wrong to consider this rude? I mean what's the point of announcing results of a private engine. You know from the title people are going to enter the thread looking for a download and then be disappointed to see it is private.Damir wrote:This version is private, not supposed to be released yet. I am testing it.
Maybe it wouldn't irritate me so much if the title included "New Edition (private)".
-
- Posts: 660
- Joined: Sat Dec 05, 2009 5:13 am
- Location: Colorado, USA
Re: Robbolito 0.09 New Edition VS Rybka 3
Well, CCRL has Spark 0.3 MP on the list, Onno 0.12.0 also, CEGT has private Fruits and Bright 0.5c.....bnemias wrote:Am I wrong to consider this rude? I mean what's the point of announcing results of a private engine. You know from the title people are going to enter the thread looking for a download and then be disappointed to see it is private.Damir wrote:This version is private, not supposed to be released yet. I am testing it.
Maybe it wouldn't irritate me so much if the title included "New Edition (private)".
Peter
-
- Posts: 2041
- Joined: Wed Mar 08, 2006 8:30 pm
Re: Robbolito 0.09 New Edition VS Rybka 3
... and if you had gone to 100 games, perhaps Rybka would be leading by 5 points!...Damir wrote: In the end however Rybka was lucky to equalize the match....
Please learn a bit of statistics:
on 50 games (assuming about 50% draws), standard deviation error is 2.5, or 5%
so a 55-45 outcome for Rybka could occur here anytime!
Re: Robbolito 0.09 New Edition VS Rybka 3
You are missing the big picture. Rybka is using 4 cores and robbo* is using 1 core.ernest wrote:... and if you had gone to 100 games, perhaps Rybka would be leading by 5 points!...Damir wrote: In the end however Rybka was lucky to equalize the match....
Please learn a bit of statistics:
on 50 games (assuming about 50% draws), standard deviation error is 2.5, or 5%
so a 55-45 outcome for Rybka could occur here anytime!