When Diagrams are Wrong

An interesting study caught my eye on the blog Brooklyn 64 yesterday: "A Problem Worth Solving." I worked on it a few minutes, then read the rest of the post where an elegant solution was presented. Then I went on to reading other blogs. Later, on the way to a school's chess club, I was haunted by the solution. I remembered the position, and a quick mental reconstruction of the diagonal a6-f1 while driving the Jeep revealed that the White king on g1 was outside the square of the pawn on a6.

The diagram as presented on Brooklyn 64.

White to move

White must have a pawn on the a-file, but not on a2.

The position as I found it in my database.

White to move

The solution given at Brooklyn 64 was played in the game.

Kosikov,Alexey I (2440) - Bezman,Vadim [C69]
URS-ch1 Soviet Union, 1986

50.Rf2 Rxb3 51.Rg2 Rb1+ 52.Kf2 Rb2+ 53.Kg1 Rxg2+ 54.Kxg2 Ng6 55.Kf2 Nf8 56.Ke3 Ne6 57.Kd3 d5 58.Ke3 1–0

After 50.Rf2, Black cannot take the knight. 50...Re7+ gives him a chance to fight on.

I made a similar copying error on one of my flash cards from Dvoretsky's Endgame Manual. Alas, I failed to set the card aside for replacement, and must find it the hard way again.


Thursday Edit

After an email exchange Tuesday evening and Wednesday morning, Brooklyn 64 corrected their diagram. I used this problem with some of my students yesterday. Winning with White after the exchange of rooks requires watching out for knight forks, as well as two notable stalemate dangers. Capturing the Black knight on f8 with White's king after eliminating Black's d-pawn = stalemate. Promoting the a-pawn to a queen (or rook) with the knight already on f8 = stalemate. It was an instructive position for elementary chess students. In one catastrophic effort, one of my chess students managed to lose all White's pawns, gaining my knight and two Black pawns in the process. He concluded that he was lost, only to learn that the White king could race to h1 in front of Black's remaining pawn = another draw.