What a way to Win (Zappa /Rybka Article)

[Ted Summers]

For more on the Zappa vs Rybka match see the write up just posted on Chess Vibes called "What A Way To Win":

http://www.chessvibes.com/?p=1284

"During during the World Championship in Mexico, the computer match Rybka-Zappa was played. In the fourth game something happened that made me lose interest in this match (and actually in computer-computer matches in general). "

"It’s not that I want to treat Zappa’s victory with disregard – beating Rybka in a match is a fine achievement – but what happened in game 4 resulted in some scepticism on my part about games between computers."

Quoting Peter from article

[mclane]

what is the reason "Peter" is disapointed ?
It seems the operator of Zappa was very clever.
Instead of praising him, he writes very negatíve about it.
Zappa Operator showed that he knows about the mechanism that are going on.
that was a good preparation.

[Harvey Williamson]

what is the reason "Peter" is disapointed ?
It seems the operator of Zappa was very clever.
Instead of praising him, he writes very negatíve about it.
Zappa Operator showed that he knows about the mechanism that are going on.
that was a good preparation.

Hi Thor,

I agree - if he knew Rybka would blow it then play on. In fact the first human intervention was the rybka operator offering a draw - Erdo said no let the engines play.

For all Erdo knew Rybka may have had some contempt set in which case it is very good to wait for 50 moves.

Best Wishes,

Harvey

[Ted Summers]

what is the reason "Peter" is disapointed ?
It seems the operator of Zappa was very clever.
Instead of praising him, he writes very negatíve about it.
Zappa Operator showed that he knows about the mechanism that are going on.
that was a good preparation.

I can only speculate, that he thought team zappa should have accepted the draw. I think Erdo did the right thing. It's not his fault that Rybka wanted to avoid the draw by 50 move rule.

[Harvey Williamson]

This very fair reply posted on the Rybka forum by Dagh - the rybka operator:

I operated Rybka during this game. Here is some information about the rules and my recollection of what went on during this game.

Firtst, we effectively played with the "Sophia rules" during this match, and I understand that this is only normal in computer chess events. Before we could agree to a draw, the arbiter would have to allow us to do so, and he would only allow us to do so if the engines saw no way out of a drawn outcome in their PVs. So, I think Peter's description in his article is not entirely precise on this issue.

OK, what happened was that at some point, David Levy said that we were welcome to agree to a draw if we liked. I think that was around the moment when Rybka started shuffling kis king on the queen-side. It can be noted that at this point, we had been playing for several hours (game 3 was played the same day before this game). Larry and Vas were also present, Peter was, and there had been also GMs present earlier during the game (when asked what he thought about the game, one would answer "I think white is better", which was of course obvious -- the interesting questions was if he thought white was winning). So, we were all kinda tired, and I was already disappointed that Rybka had blown the win (which was pretty clear after f4-f5 and black had established a blockade). But, at the same time, Erdo had already made it quite clear that he wanted to play on in anticipation of what later turned out to happen (Rybka going on to sac pawns). So, actually, I don't think a formal draw offer was ever made. Instead, I think I asked Erdo if he wanted a draw, "are you sure you want to play on" etc. Probably also David made similar comments. And we were all kinda laughing about the absurdity of the situation: That Rybka was nominally better, had been winning, but Erdo wanted to play on anyway now for a win with black. But, please note, we all also quite respected this, even if we went on to tease Erdo a little about him always being so super-optimisitic. Erdo would respond to such comments with his own teasing like "I'm here to play chess, I can play all day long, if people want short draws, they can just watch the boring human WC!".

It's probably also worth mentioning that during the games, we would always discuss what went on on the board, and we had some small funny competitions about explaininng each other and David how things were always much rosier for our respective sides. So, naturally, Erdo would explain to us how Rybka would sac first the h-pawn, than the b-pawn etc., or in some other order. Just as he had been explaining how black could put his rooks like this or that to block white's pawns earlier on, or as he had been explaining how black would play Qb6-g6 in game 2 and get a mighty attack (after Ne7, sac'ing a pawn), which we of course now know didn't happen. It was all good-spirited fun and competition.

Did we share Erdo's opinion that Zappa might very well win? I personally didn't even like to think about the prospects of this happening, again, Rybka had more or less just thrown out a win. We all understood that it was true that Rybka might well sac at least one pawn (since her eval was significantly more than +1.00), but I don't think we seriously believed that white might go on to even lose, at least not at this point of the game. So that's why informal "draw-negotiations" were in the air.

The first one to really alert us to the seriousness of the situation was Larry, who knew that KRRB vs. KQ is a theoretical win. So once Rybka was down to only two pawns, we could just hope that she could find and would choose a perpetual, but alas.

Anyway, I absolutely in no way can see any reason to consider anything that went on during this game "unfair", and we never even mentioned such a question during the match. Erdo was rewarded for his decision to play on. On the other hand, if the game had still concluded in a draw anyway, he would promptly have been the target of teases like "hey, you know, it was a draw all along, but we could have had lunch already an hour ago!".

Also, with regard to the question of operators having undue influence: Please note, that if it had been up to the programs to determine when to accept a draw or not, the result would have been the same! Rybka would NOT have agreed to a draw on her own here, that's the entire point!

So, with the draw-rules as they were, operators will/would have absolutely minimal influence on this aspect of the game. Essentially, the rules said that we could ONLY agree to a draw when it was totally clear that the game would end in a draw anyway. If anything, the only thing questionable here was David allowing us to agree to a draw, since the position (and the engines playing it) turned out NOT to remain a dead draw. In other words, playing on was simply the correct procedure here, and we can only thank Erdo for making sure that the correct procedure was followed