From what I see of the official page, Rybka didn't have very good day at Mainz. V. Shredder 1-1, v. Deep Sjeng 2-0, and a loss to Naum 0.5-1.5. Game 6 (v. Naum with Rybka white) looks particularly embarassing for Rybka; just look at the position around move 20 - phew!
Livingston Chess960 Computer World Champ, 2008
after day one:
1 2 3 4
1 Naum ** ½1 01 10 3.5/6 10.25
2 Rybka ½0 ** 01 11 3.5/6 8.75
3 Shredder 10 10 ** 01 3.0/6
4 Deep Sjeng 01 00 10 ** 2.0/6
Ah, the results are spread quite evenly; guess Rybka wasn't that bad after all. But the games, that is, the play, looks ... unconvincing.
12 games and only 1 draw! A new era in computer chess?
It seems they are having computer problems.
No excuses! It wasn´t the show of an engine, which is +200 Elo better than the rest of the world.
PS: Okay, Game 1 against Shredder is excused!
PS: Okay, Game 1 against Shredder is excused!
> No excuses! It wasn´t the show of an engine, which is +200 Elo better than the rest of the world.
>
My point too.
Rybka will come back on day two.The guys are busy at work.Maybe Monte-Carlo will help.
Actually no one said that Rybka 3 is +200 elo better in chess960 (if we exclude one of the positions, which is normal chess). In chess960 the opening phase seems to be critical, you can be objectively lost in 10-15 moves without playing weak moves. One weak thing the Rybka 960 doing is letting to opponents push the pawns and grabbing space as Jeroen already mentioned. Prbably there was no time for polishing the chess960 engine before the release.
Also they are having performance (hardware) problems as well, specially in game one as you concluded.
Also they are having performance (hardware) problems as well, specially in game one as you concluded.
> No excuses! It wasn´t the show of an engine, which is +200 Elo better than the rest of the world.
>
> PS: Okay, Game 1 against Shredder is excused!
I must have missed something but why is that game excused?
3.5/6 is clearly normal for an engine that is +200 elo better than the rest.
It is slightly less than expected(4.5/6 is near the expected result) but if you are unlucky in one game and lose it instead of winning it then it can happen.
Rybka is sharing first place.
Uri
It is slightly less than expected(4.5/6 is near the expected result) but if you are unlucky in one game and lose it instead of winning it then it can happen.
Rybka is sharing first place.
Uri
You are right! But 5/6 would be more normal for me. :-)
When Rybka loses a game last time in an official tornament? Okay, Mexico; but today she loses 2 games out of 6.
When Rybka loses a game last time in an official tornament? Okay, Mexico; but today she loses 2 games out of 6.
Perhaps the new terms that make Rybka incredibly strong at normal chess hurt at Chess960.
Why does Rybka always seem to have problems at tournaments on the first day? Does nobody do a "dry run" to make sure hardware/software are ready to go?
guys 6 games is not statistically significant even for a +200 ELO engine.
Nonetheless, watching the games is very significant: Rybka got clobbered.
It would be catastrophic if statistical significance would be the only guideline for such a product and it's quality.
All the "big things" are evaluated by statistical significance like:
1. Climatic disaster
2. Birth surplus
3. Richness
4. Welfare
5. Intelligence
...
9. Chess engines
1. Climatic disaster
2. Birth surplus
3. Richness
4. Welfare
5. Intelligence
...
9. Chess engines
Well, chess engines should be entry 2,436,825 if you are ranking them by importance! :-)
Regards,
Alan
Regards,
Alan
No!! I think, number 9 is correct. Chess is numer 8!
agreed.
Rybka humanized at 960 will be interesting to watch when games are brought forward.
haha Rybka is like Topalov in that aspect,she will come from behind and win it all.
also in that she uses her own resources to win. not like kramnik
In San Luis, Topalov makes it the other way around. :-)
Of course its interesting Turbo, but guys don't get too discouraged is all I was saying as it seems some of you have.
Mike
Mike
?? You are sure, you answered me?
Why does Rybka always seem to have problems at tournaments on the first day?
Yes Jim, I wonder what causes these. Well, we hope that she will win, anyway, yes, she will win. I remember even lsat time they had a GUI problem that did not understand castling rules. Guys must start to be well prepared when they go to tournaments :(
Yes Jim, I wonder what causes these. Well, we hope that she will win, anyway, yes, she will win. I remember even lsat time they had a GUI problem that did not understand castling rules. Guys must start to be well prepared when they go to tournaments :(
Yes, Rybka was doing some strange things against Naum. Bad evaluation at some point (due to the passed pawns?) and playing without plan.
Not to mention the operator error by Vas.I was surprised the organizers let the game resume back where the error occurred against Naum.Just goes to show that their are other programmers making progress.
>I was surprised the organizers let the game resume back where the error occurred against Naum.
This is standard in tournaments where such archaic things as "operators" are still used. :) With the time limit of 20m+5s, I'm surprised there's not more errors, really.
yes, exactly, operator errors don't decide games.
Does it seems a little strange to have Vas "Mouse Slip" Rajlich as operator for these rapid games? :-)
Well, ok, tomorrow he will have a chance to practice a bit more ;)
Powered by mwForum 2.27.4 © 1999-2012 Markus Wichitill