Forums

Improvements for white in this endgame?

Sort:
zezpwn44

This is from a game I played tonight on chess.com that lasted about 2 hours. I was better from the start, but just couldn't convert this endgame. Houdini kept showing that I had a +.8 throughout until it very slowly decreased at the end, and indeed I thought I was winning most of the way through, but all Houdini was able to suggest was cryptic computer lines and I'm still not sure where I went wrong.

I accepted a draw at the end. 

I feel that I really lack a deep understanding of these minor piece endgames, and engines do not at all seem to be a useful training tool for them. Strong endgame players out there - any advice?

sammynouri

I'm not a strong endgame player but 1. Kd3 was in my opinion a good move because it doesn't allow his king too close, and gets your king closer to the center. 7. Bd2? I have no idea why you did that, better may have been 7.g5 Kf5 8. Kd3. You would proceed by trading off the pawns on the kingside, yes its fine if you lose that extra pawn because your king would move over to the queenside and take blacks remaining pawns. Knights aren't good at defending pawns. Again, I'm no expert so take my words with a grain of salt.

sammynouri

Actually, 1.Kd3 loses a pawn to Nb4+ so I take that back. I still have trouble with knight forks.

sammynouri

1.Kf3, I think that might work better.

sammynouri
sammynouri wrote:

Actually, 1.Kd3 loses a pawn to Nb4+ so I take that back. I still have trouble with knight forks.

No, it doesn't! Black loses his knight to 1.Kd3 Nb4+ 2.Kc3 Nxa2+ 3.Kb3 and the knight is trapped. It's funny how endgames are often more complex that middlegames. Sorry about 4 posts in less than 10 minutes, I have a thing for endgame positions.

AdolfAnderssen92

 
I don't have access to a engine right now so i don't know why your houdini only gives +0.8 in this position. To me the initial position looks winning for white.

Now let me put my thoughts into words.

why do i think white has a winning position?-

1.White has the bishop.Bishop is considered superior to a knight when the pawns are present on both sides.

2.Black kingside pawns are weak.

In general any plan in any chess position involves  attacking the weakness(i.e the kingside pawns).Here you can only attack the weak pawns with the king

..so the best move should be 1.Kf3 and now you should be able to win easily(or so i believe).

The plan should ideally work as follows

1)first tie your opponents pieces to the defense of these pawns and

2)then advance your kingside pawns creating a passed pawn.

3)Now your opponent might need to give up the knight for the passed pawn in which case you will win easily...But if your opponent manages to set a blockade you will sacrifice the passed pawn for time which you are going to use to rush to the queenside with the king.

1.g4 is a blunder(allowing Ke4 and not activating your king) after which i am no longer sure if white has a win.

Note: Try to always activate your king first and then push your pawns.(i.e in king and pawn endings and some minor piece endings).

zezpwn44

There's no time to prevent Nf8 though. 13. Bb4 gh and black maybe even has the better chances, and white certainly can't win.

caveatcanis

White has the following winning plan: g4 f3 Bd2 Ke3 Bc3, followed by marching the king through the centre to eat Black's queenside pawns.

We use the threat of creating a passed pawn on the kingside to distract the Black pieces. If necessary, this pawn can be sacrificed.

That's why Houdini likes 7. g5. After 7...Kf5 8. Kd3 Black is in zugzwang and has to allow the White king through. For example 8...a5 9. g6 hxg6 10. h6 Kf6 11. Ke4 and it will soon be all over.

[Edit] I just notice that Houdini prefers 8. Kf1 rather than 8. Kd3.

After consulting Stockfish, the reason is that it's playing for a different zugzwang. The immediate 8. g6 hxg6 9. h6 Kf6 isn't very clear, but it would be devasting if White could safely play Bd4+ here.

So after 8. Kf1, Black can't move his knight. After 8...Ke5 9. Bd2 Kf5 would be met with 10. g6, so Black has to play 9...Ke6, at which point White can play 10. Ke2 and advance his king through the central White squares.

This finesse clearly makes sense to engines (so there must be some point to it), but the naive human move 8. Kd3 is also seems to win easily.