Kamsky's Endgame Play

The Philadelphia Open recently concluded in my home city with the victory of the top seed, former world championship challenger, Gata Kamsky.

Kamsky has a very unusual style, a style which I think many people neither understand or appreciate. I spoke to one of his opponents after an earlier round, GM Magesh Panchanathan, who said something to the effect of: "I would be thinking that some move was forced, and while I was calculating some long line, Kamsky would find a completely different first move, which I hadn't even considered." This interesting game ended in a draw:

Kamsky's style is distinguished by a deep understanding of the basic mechanics of chess. He is known as a great technician, but you can find examples of crazy, complicated games played by him - such as his recent contests against Shakhriyar Mamedyarov from the World Cup. While he played very theoretical openings in the past (and was also criticized at that time by Kasparov for being "unimaginative"), since his retirement and return to chess, he has mostly played obscure openings. A main feature of his play seems to be the relentless setting of problems for the opponent. My own game with him from some years ago left a strong impression on me:

In this position, arising from the Dragon, Black's queenside pawns are broken up, as usual, but Black has some piece activity for compensation. I expected 22.Nb2, but Kamsky almost immediately played 22.Rxd8+! Rxd8 23.Nxb6! axb6, not only giving up the d-file, but also unifying Black's pawns. Then followed 24.Rd2 Rxd2 25.Kxd2, with the following position:

Black's problems are mainly that White, by a4, b4, and a5, can create a passed pawn on the queenside. While I understood that White was better, what was impressive was that Kamsky took this decision almost immediately despite the risk of a draw that the simplification posed, confident of his evaluation of the resulting ending.

It takes a very deep understanding to even consider the - on the surface, paradoxical - sequence of moves (unifying the black pawns), and to be perfectly confident in the evaluation. And additionally, it seems he was right, since I couldn't find a draw, and I doubt that one exists:

The main game of the article will be Kamsky's final round game against GM Ioan Chirila, of Romania. As above, we will see Kamsky's play in the endgame, including his understanding of when chances remain after considerable simplification.

While the final blunder, 42...b5??, was clearly the immediate reason for Black's demise, it is very possible that the position is objectively lost anyway. Note that I could hardly condemn any of Black's earlier moves as outright "mistakes" - just as "questionable decisions," such as 17...cxd4, 23...Rcc8, and 27...dxe4, which added up to give Kamsky enough to create his chances and take the clear first prize.