There is one undeniable truth when it comes to rook endgames: everybody goes wrong in them. Even in the «simplest» theoretical positions the greatest champions have made «incredible» mistakes.

It is comforting to know that even the best players sometimes go wrong in theoretical endgames. This only means that they are not so easy to master and play. Which of course shouldn’t put us off from studying them.

Here’s a small selection of the greats going wrong.

This theoretically drawn position arose in Capablanca-Menchik, played in Hastings in 1929/30. Black draws with 55…Rb8, controlling the important eighth rank. However, Menchik played 55…Ra6? and now White is winning.

The great Cuban, known for his great technique, returned the compliment. Instead of winning with 56 Kf8 Kg6 57 f7 Kf6 58 Kg8, he played 56 Rd7? after which it’s a draw again.

Now Menchik got it right with 56…Ra8, but after 57 Re7 she went wrong again: 57…Ra6? repeating the position instead of keeping the rook on the eighth rank by moving it to b8. Capa pounced with 58 Kf8 Kg6 59 f7 Ra8 60 Re8 Ra7 61 Re6 Kh7

But now instead of 62 Re1 he shockingly played 62 Ke8? after which it’s a draw again after 62…Ra8 63 Ke7, but Menchik missed it for one last time – instead of keeping control of the eighth rank with 63…Kg7 she let it go with 63…Ra7? and resigned after 64 Kf6.

You would expect Capablanca to know these things, the number of mistakes in this game is quite surprising.

This is the position from the game Tal-Zaitsev, USSR Team Championship in 1968. Black’s pawn is still not advanced so the defence from the front is called for. Hence 73 Rb1 draws. Tal however hasted to get closer with the king (which cannot be prevented, so there was time for the move 73 Rb1) and played 73 Kd3? and after 73…Re1, preventing the defence from the front, he resigned.

The position from the game Piket-Kasparov, played on the internet in 2000 (at the time control of 60 minutes per player per game). Black didn’t have to allow all this, but even here he should draw without too many problems. And one would expect Kasparov to know how. And yet he errs with 44…Re1? missing White’s idea of 45 Rc7 followed by the pawn breaks f5 and/or e6. The correct defence is the prevention of this by 44…Ra3 45 Rc7 and now either 45…Ra6 or 45…Ra5.

Very similar to the Capablanca-Menchik game above, this one is from Aronian-Carlsen, Tal Memorial 2006. Here Black controls the eighth rank, so he only need to maintain that control with 73…Kg6. Instead he commits the same mistake as Menchik: 73…Ra7? and after 74 Ke8 he resigned.

This is only a small selection that came to mind out of the many examples you can find in the databases. I have found that these mistakes are made when there is lack of clarity in the understanding how the draw (or win) is achieved. Therefore, when you study such theoretical endgames, do not rush, take all the time you need to understand why a move is made and what it prevents (or allows). Repeat and revise often, to make sure the ideas stick in your memory. You will be grateful for this work when you win (or draw) with few seconds left in an important game.