Great little moves from Paderborn

At the end of December 2006, the 16th International Paderborn Computer Chess Championship was held. The tournament was played in the beautiful Heinz Nixdorf Museum in Paderborn and Rybka won with 6,5 out of 7, after she only allowed a draw against the program Spike. Second was Shredder, who who all games except the loss against Rybka. Martin van Essen shows us some nice fragments from the tournament.

By now, Rybka 2.2 has a CEGT rating of 3015 and so passed a magical boundary. The only serious competitor, Deep Shredder (6), was defeated in the mutual encouter. The rest hardly mattered. While scanning through the games, I run into some great stuff:

Gandalf X - Rybka
Paderborn, 2006

32. …, Qxf2!! 33. Rxf2, axb3 and we’re watching in complete astonishment how White is completely lost because of a decisive breakthrough on the queenside. 34. Rf1, bxc2 35. Rc1 doesn’t help because Black’s b-pawn is coming. White didn’t see anything better but 34. Qxb4, bxc2 35. Rxc2, Rxc2 when the two black rooks easily consumed two more pawns and forced White to resign.

Rybka-Jonny
Paderborn, 2006

11. …, Nc5!? Nice move. Is it theory? 12. dxc5, Nxc5 13. Bxf5! The solution: White gives the queen. 13. …, Nxb3 14. axb3, d4 15. Nxd4, Bxh2+ 16. Kxh2, Qxd4 White’s three pieces turned out to be stronger than the black queen and pawn.

A bizar game by Rybka’s competitor:

Diep – Deep Shredder
Paderborn, 2006

White just took a bishop on b7. Instead of taking back, Black played:
26. …, Ng5!?
A true Tal move. ‘Of course’ there followed 27. Qxe8+, Rxe8 28. Rxe8+, Kg7 29. Bc8 Almost every white piece is hanging and this move is directed against the nasty-looking Nxh3+ and also protects the knight, but from hindsight I can imagine White had better let his rooks protect each other, e.g. 29. Bg4, Qxb7 30. Rce1. But these dazzling complications should rather be studied by a program!

Black’s queen is now facing a superiority but still, Black throws in another piece:
29. …, Nf3+!
Of course, the pointe is 30. gxf3, Qg5+ and 31. …, Qxc1, when White still has some problems with his piece co-ordination, but how should we judge the position? White played 30. Kh1 when 30. …, Qd2 31. Rf1, Qd3! (wow how nice) started chasing the white rook. After 32. Ra1, Qd4 33. Rc1, Qf4 the time had come for White to take the knight. Perhaps a move earlier White should have accepted the move repetition. He should constantly take into account that the rook on e8 can also drop all of a sudden. Well, Black grabbed the exchange, after which the pawn avalanche on the kingside was enough to get White on its back (0-1, 49). The knight and the bishop didn’t make a single move anymore. Great game!

Another good one:

Deep Shredder-Gridchess
Paderborn, 2006

With 20. Rhd1 White allowed Gridchess to shut in his bishop with 20. …, g5. This you could call a speculative sacrifice! But note that we’re talking about chess programs. 21. Bxg5, hxg5 22. Rg4, Kd8 23. Rxg5+, Kf8 24. h4 This pawn was eventually brought to the seventh rank, and two more were added there which on the 50th (!) move (we’re taking it easy) led to a unique picture:

Position after 50. e6-e7.

White is still a piece down, but it’s clear that he has more than adequate compensation. (1-0, 55). By the way, it seems like Black has castled queenside in this game but that’s not true (13. …, 0-0).

Thanks to Martin van Essen for this report.