How to Actually Improve at Chess Endgames

Capablanca's advice to study endgames before openings is quoted constantly and followed almost never, because openings feel productive and endgames feel like homework. But the advice is right, and the reason is specific: the endgame is the one phase where the correct method is finite, learnable, and permanent.

An opening line can be refuted, goes out of fashion, or simply never appears in your games. The technique for winning king and pawn versus king has not changed in two hundred years and never will. Every hour spent there keeps paying, in every game you ever play, forever.

Start with the checkmates, in this order: king and queen versus king, king and rook versus king, then two bishops. These are not optional. Failing to convert them is not just a lost half point — it is the kind of loss that stays with you, and it happens to rated players more often than anyone admits.

Then learn the opposition in king and pawn endings. This single concept — the geometry of which king gets pushed back — decides an enormous number of club games, and most players who lose these endings never realise the result was determined several moves before the pawn fell.

After that, rook endings, because they are by far the most common ending in real play. You need two positions cold: the Lucena (how to win when you're up a pawn with the enemy king cut off) and the Philidor (how to draw when you're not). Between them they cover a remarkable share of the rook endings you'll actually reach.

Notice what is not on this list: hundreds of theoretical positions. The goal is not encyclopedic knowledge. It is recognition speed — being able to look at a position and think "this is a Philidor, I know what to do" rather than calculating from scratch with five minutes on the clock. That recognition is what technique actually is.

The most efficient way to build it is to play the positions out against an engine rather than reading them in a book. Reading tells you the method; playing it against something that defends properly tells you whether you can execute it. Those are different skills, and only the second one shows up in your games.

One more reason to do this now rather than later: endgame skill changes how you play the middlegame. Once you know which endings you win, you know which trades to make. Players who avoid endgames are effectively playing the middlegame blind — they can't tell a good simplification from a bad one.