Rybka vs Benjamin II - Match at Draw Odds

Introduction

Terms and comments by Larry Kaufman
Introdution by Vasik Rajlich

Conclusions

Post-match comments by Vasik Rajlich
Post-match comments by Roland Rösler

Schedule

Game 1: Benjamin - Rybka 0-1
Game 2: Benjamin - Rybka 1-0
Game 3: Benjamin - Rybka 0-1
Game 4: Benjamin - Rybka 1-0
Game 5: Benjamin - Rybka 0-1
Game 6: Benjamin - Rybka 0-1
Game 7: Benjamin - Rybka 0-1
Game 8: Benjamin - Rybka 0-1

Games

Game 1, Benjamin - Rybka 0-1: A logical game from both sides. Rybka equalized in the opening, but at the cost of allowing a queen trade and a symmetrical pawn structure. Benjamin made a few more tiny concessions to further simplify the position (20. c4) and a win to start the match didn't appear to be far away for him, but in fact the endgame was not routine and he was not able to defend it.

Game 2, Benjamin - Rybka 1-0: Another logical game from both sides, very similar to the first game. Rybka again equalized in the opening, but at exactly the same price - a trade of queens and a symmetrical pawn structure. Benjamin again made a few more tiny concessions to further simplify the position (19. Rad1) and stood nominally worse, but this time the endgame proved to be rather simple and white never really had to sweat.

After two games, the main pattern is already clear. Rybka has a fairly strong disincentive to trade pieces, strong enough to cost her around 40 Elo in self-play, but Benjamin is willing to go to greater length to force trades than Rybka is to avoid them. This is in fact quite logical, as his heuristics about this topic are better than Rybka's. I can't seriously fault Rybka's play at any point so far, although Larry and I did make one adjustment and a different version will play starting in game 3.

Game 3, Benjamin - Rybka 0-1: Benjamin accepted complications with 11. Nb5, sacrificing a pawn on the queenside to gain the initiative. The position was quite sharp and rather computer-friendly and after a series of very non-obvious inaccuracies, white's position crumbled.

Game 4, Benjamin - Rybka 1-0: The themes from games 1 and 2 returned. The queens were traded and the position was closed quite early. The Rybka version which plays in this match is averse to both, but Benjamin was again willing to go to greater lengths to achieve these aims than Rybka was to avoid them. There was still plenty of play in the resulting position, but Benjamin was able to defuse it completely by sacrificing a piece for two pawns and achieving a fortress. Rybka (mis)evaluates that position as a full pawn better for herself.

Game 5, Benjamin - Rybka 0-1: Benjamin again made opening concessions in order to lock the position, but this time he lost his grip. Rybka cracked things open with 12. .. b3 (which must have come as a surprise) and 14. .. f5, with a strong initiative for black. The resulting position is nearly impossible for a human to defend.

Game 6, Benjamin - Rybka 0-1: The openings are proving critical in this match, and this time Benjamin lost the battle quite early. He achieved a symmetrical pawn structure, but at the cost of the bishop pair, and black was able to get active play on the queenside. The resulting position is a nightmare for a human to try to deal with without computer assistance.

Game 7, Benjamin - Rybka 0-1: On move 12, Benjamin was faced with the typical dilemna for this type of match: play a solid anti-computer move, or play a good natural chess move. This time, he chose the latter and forged ahead with 12. g4. He kept the balance until move 26, when his 26. Qd2 appears to be a very complex error. The resulting position is difficult for white objectively and hopeless for a human against a computer.

Game 8, Benjamin - Rybka 0-1: The opening variation was again a bit too open and unbalanced to give the human his best chances. It seems that 20. Qd2 and 21. a3 are both inaccurate due to the surprising shot 21. .. d5, which turns the game in black's favor.

Analysis

Games 1 & 6 by K. Utzinger