There is a bit of a story behind this endgame as far as I am concerned.   After 55 moves playing Black, I reached the position below against Richard Allis in a B&DCL match.   I knew that two bishops against knight was a technical win and that two bishops against knight and pawn was normally won too.   I had never studied either endgame and had always assumed that winning the pawn was the hard part of converting and that finishing off BB vs N was easy.

The game continued and I won the pawn in a few moves.   Then I spent the next 40 or so moves trying really hard and getting nowhere very fast against the lone knight, whilst Richard confidently and quickly made reasonable looking moves.  Realising that I could no longer win within the 50 move rule I offered a draw which Richard accepted. I went home feeling like a complete chump for failing to win! Fortunately it did not affect the match result.

The following day I started studying the endgame.  What I discovered was that:

• winning the pawn is the easy part (just as I had found during the game)
• winning the BB v N is not easy and even grandmasters often fail
• Botvinnik resigned BB v N against Tal as soon as it was reached – against Tal that might be OK but is not generally recommended!

To illustrate the best approach and the extreme difficulty of the endgame it is worth looking at a couple of key decision points from my game against Richard.

Richard has just played 63.Kg3 and there seemed to me to be two valid approaches:
1) I could play 63…Bh5 and start the process of driving the White king to the edge whilst my pieces are nicely coordinated
2) With the knight separated from the king, I could start chasing it, with the idea of winning it by force.

I chose the latter, but the knight slipped passed my pieces easily and soon reconnected with the king and I regretted my decision.   Indeed, when I studied the endgame at home what I discovered was that the following rules of thumb seem pretty good:

• Only go chasing the knight when it is separated from the king if you can definitely trap it – this is often when the opposing king is already confined to a corner or an edge.   Very often chasing the knight when it is separated from the king but the king has some room is not productive.
• The first step is normally to drive the king to the edge or a corner before worrying about the knight.    If you can do this then the knight may be trapped in the corner/edge or separated from the king under favourable circumstances, i.e. when it is hard to reunite with the king.

The second decision point I remember was in the diagram below.  This is just after Richard had played 84.Nf6 (you can see I haven’t made any progress at all).

My candidate moves here were 84…Ba3 and 84…Bb4 and I chose the latter, largely by instinct as I had little time left to analyse.   Tablebases tell us that after 84…Ba3 it is mate in 38 with perfect play, but after 84…Bb4 it is mate in 60.   When such a small difference in the move makes such a huge difference in the distance to mate you know you have a tricky endgame on your hands!

My study of the endgame made me feel a bit better about not scoring the full point.  And if I reach this endgame again I will be a little better equipped to handle it, but I am rather hoping to avoid it all together!