Here White was better in the middlegame drifted off into equality and then blundered on move 64 (overlooking a two move tactical shot) in the endgame.

5.Bxf7ch Ke7 6.Bb3 Rf8 7.d3 d6 8.Be3 Bxe3 9.fxe3 Qe8 10.Qe2 Bg4 11.Nf3 Kd8 12.c3 Nb8 13.Nbd2 a5 14.a4 Nbd7 15.Ra3 Bh5 16.Rf1 Nc5 17.Bc2 Ng4 18.h3 Nh2 19.Rf2 Nxf3ch 20.Nxf3 Kc8 21.g4 Bf7 22.c4 Be6 23.Nh4 Rxf2 24.Qxf2 Kb8 25.Nf5 g6 26.Ng3 Qe7 27.Kd2 Ka7 28.Ra1 Rf8 29.Qg2 h5 30.Rf1 hxg4 31.Rxf8 Qxf8 32.hxg4 Bxg4 33.Nf1 Na6 34.b3 Nb4 35.Kc3 Bh5 36.Kb2 Qf6 37.Ng3 Qg5 38.Qf2 Bd1 39.Bb1 Qh6 40.Kc3 Qh3 41.Kb2 g5 42.Kc3 g4 43.Kb2 Qh4 44.Kc3 Bf3 45.Kd2 Qh8 46.Ke1 Qh6 47.Kd2 Kb8 48.Ke1 Qh3 49.Kd2 Kb8 50.Ke1 Qh6 51.Kd2 Qh3 52.Ke1 Na6 53.Bc2 Nc5 54.Qg1 Ka8 55.Qf2 Na6 56.Qg1 Nb4 57.Bb1 Ka7 58.Qf2 Qh8 59.Qg1 Qh4 60.Qf2 Qh7 61.Qg1 Qh6 62.Qf2 Ka8 63.Qf1 Nc6 64.Bc2?? (Nf5) Nd4 0-1

sloughter wrote on 01/07/13 at 01:53:09:


Were you playing white or black?

The Wilkes Barre as it is called in the US or the Traxler as it is known in Europe is probably the most wildly tactical opening in all of chess. The starting moves are 1.e4 e5 2.Nf3 Nc6 3.Bc4 Nf6 4.Ng5 Bc5?! Black sacs the Rook for a powerful attack. Only the foolhardy now play 5.Nxf7? because 5…Bxf2ch! gives Black a huge attack. According to Pinski in a review of 2,500,000 games Black won 77% of the games where 5.Nxf7 was played! 5.d4 d5 is slightly better for White but the real refutation occurs after 5.Bxf7ch. You might think that Black’s position is unsound until you realize that Anand, a future World Champion lost on the White side of this variation and  Karpov, a World Champion only drew on the White side of this opening.

I have discovered three simple moves in the opening that completely refute this variation. When I tried them out on Fritz 12, I got a completely winning position then botched a simple ending and only drew. When I tried this opening for the first time against Houdini 3, it hit me with a combination 15 moves deep and even though I had two minor pieces for a Rook and two pawns I knew I would lose.

The second game I drew in 24 moves (see supplemental game), lost the third and here is the fourth game. Houdini plays an unsound sacrificial combination because it realized passive play was hopeless: 5.Bxf7ch Ke7 6.Bb3 (This is a simple way to a solid plus; the other popular choice is 6.Bd5 e.g. Karpov-Beliavsky) 6…Qe8 (Rf8 is somewhat poorer. I got a solid plus against Fritz 12 with 7.d3/Be3/fxe3/Nf3/Nbd2/Qe2/O-O-O & Black has no compensation for the pawn. 7.d3 h6 (Qg6? 8.Bf7 +-) 8.Nf3 Ng4?! 9.Rf1! When I looked at this variation with Houdini before playing this game I noticed that the machine didn’t castle out to move 24 so I opted for 9.Rf1, not 9.O-O. 9…Qh5 10.h3 Bxf2ch 11.Rxf2 Nxf2 12.Kxf2 White is much better with two pieces for a Rook but the game is young. 12…a5 13.a4 d6 14.Be3 Bd7 15.c3 (to save the Bishop) 15…Raf8 16.Nbd2 Rf6 17.Qf1 (To free the d1 square for the King) Kd8 18.Ke2 (Now the King strolls to the sanctuary of the Queenside) 18…Kc8 19.Kd1 Qe8 Black shows an absence of planning and is just making disjointed moves. 20.Kc2 Rhf8 21.Kb1 Kb8 22.Ka2 (This doesn’t detract from my winning chances and is superfluous---we are already in what pros call a “trial and error” position i.e. don’t do anything committal and just build on the position) 22…Rg6 23.Qe2 Ne7 24.Kb1 Rgf6 25.Bc2 Ng6 26.b3 Nf4?! (This is dubious---the fewer the minors the less opportunity to complicate; Houdini is driven by evaluation. It clearly fears my pawn levers in the center which I had no intention of playing for awhile. The position is deceptive---White has a big advantage but swapping tactical shots with a machine that sees 15 moves ahead seemed dangerous.

27.Bxf4 Rxf4 28.Bd1 Be6 29.Kb2 g5 30.Rc1 Qf7 31.Rc2 (It is important to defend along the second rank)  Qg7 32.Qe3 g4 33.hxg4 Rxg4 (At this point I decided to build a “Great Wall” around my King. There are no useful pawn levers on the Queenside for Black or in the center and I can keep the majors from invading with the simple piece placement of Rf2/Ne1/Bf3/Qe2 34.Ne1 Rg3 35.Bf3! h5 36.Nf1 Rg5 37.Rf2 h4 38.Nd2 Qe7 39.Nc4! (Going for the draw) 39…Bxc4?! (Giving up any hope for a win) 40.dxc4 (Freezing the Black pawns on the Queenside) Rg3 41.Qe2 Rf7 42.Kc2 Qf8 43.Kb2 h3 44.gxh3 Rxh3 =`                                                                  

White has constructed a fortress than cannot be breached and draws this ending by simply shuffling his King from a2, b2 and c2. The only attempt to win by Black is to play Rg1 or h1 to overload the Knight on e1, when I simply responded Nd3 defending the Rook on f2. Black can make no progress and the fifty move rule kicked in on move 129.

After analyzing the Wilkes-Barre/Traxler with Houdini 3, I noticed that Houdini didn't castle out to move 24. Here is the second game I played against Houdini at tournament level to see what Houdini could do:

The WBT is definitely cooked. Houdini played a sacrificial combination no Grandmaster would play stepping into a middlegame favorable to White. I thought I had a big advantage when I took the draw but Houdini disagreed and only had the advantage as +.27.

1.e4 e5 2.Nf3 Nc6 3.Bc4 Nf6 4.Ng5 Bc5 5.Bxf7ch Ke7 6.Bb3 Qe8 8.d3 h6 9.Nf3 Ng4 10.Rf1 Bxf2ch!?! 11.Rxf2 Nxf2 12.Kxf2 Rf8 13.Be3 d6 14.c3! Kd8 15.Nbd2 Bd7 16.Kg1 a5 17.a4 Kc8 18.Qe2 Rf6 19.Bc2 Kb8 20.b3 Qe7 21.Bd1 Rf8 22.Bc2 Rf6 23.Bd1 Rf8 24.Bc2 1/2-1/2

Obviously, against a human opponent I wouldn't force the draw.