On Monday I left you with this position which came from a game played in our lichess arena a couple of weeks ago: I had the black pieces against Steve and had to decide what move to play.

(As it happens, White’s last move was g3xf4: the more natural Kxf4 would have led to a draw with best play.)

I asked you which of Black’s possible moves would win, which would draw and which would lose.

Black has two winning moves here

38… a3!, followed by a king invasion down the a-file. White can’t hold the queenside and stop the h-pawn at the same time.

Or, if you prefer,

38… Kc6! when, if White now plays 39. a3, your king will have time to defend the h-pawn which will eventually win the game.

Black also has two drawing moves: b4 and Kc5.

The other three moves lose: I chose one of them.

38… h5??

Now 39. a3! wins, stopping any nonsense on the queenside and then winning the h-pawn.

But instead we continued:

39. Kf2?

Here 39… a3! now draws: Black promotes first but can’t stop White’s pawn on e7. (Play it out yourself if you’re interested.)

39… h4?

40. Kg2

40. a3! is the easy way to win, but this is also good enough.

40… b4

This is what I’d seen several moves earlier, which explains my previous play. My idea was to tempt his king outside the square of the c-pawn and then play the sacrificial breakthrough. But if White remembers the trick he just plays 41. a3! and Black can resign. Unfortunately he missed it.

41. cxb4?? a3!

And now it’s White who had to resign.

I was just fixated on this idea rather than trying to play sensible moves. As it happened it worked, but, given how poor my 38th and 39th moves were, I really deserved to lose this game.

I hope you’ll learn something about pawn endings from this example. You might find it instructive to play out the variations. I can’t emphasize too much how important they are in modern chess.