PDA

View Full Version : Bizarre minor piece ending



Kevin Bonham
03-02-2006, 04:08 AM
Black to move.

6k1/5pBp/2p2P1P/1p6/3K4/2nB4/8/8 b - - 0 40

Ivkovic-Saunders. HICC One Hour 2005-6.

White should have won this in a canter after winning a piece in the middlegame (Black could have averted this by putting his Q en prise then recovering it with a knight fork, but that's another story.)

Instead, he's pushed his f-pawn up and walled in his bishop on g7, which means he has to be careful because of stalemate defences. In the position above Black played 1...Nd5 and White responded 2.Bxb5 and a draw was agreed.

Is it won for White after 1...Nd5 with best play? What about had White played 1...Na4 instead?

Comments welcome. This one really is tricky.

eclectic
03-02-2006, 07:19 AM
the task at hand irrespective of black's move is for the white king to settle on either e7 or e8 so that it can support either or both of 2 tasks ... the light squared bishop mating on f7 and/or the dark squared bishop coming to f8 and thus being activated

white has to be careful to avoid forking knight checks; doubly so if both bishops come into play

it might be wise not to eliminate both black passed pawns as temporising might be better enabled for white with spare black pawn moves not to mention the need to deprive black of stalemating opportunities

perhaps eliminate one pawn lest an advanced pair become too troublesome

indeed it doesn't not matter if one of these pawns queens so long as white has the opportunity to inflict mate first

the black knight will be too stetched to harrass both the king on e7 or e8 and the light squared bishop on b1 aiming to shift over to a2 and threaten mate on f7

it might hope to rest on c3 driving the bishop off b1 and deprive it of access to a2 but then the dark squared bishop is free to come out of hiding and put further pressure on the knight and the pawn

this is assuming that the b pawn remains so adjustments in strategy would be needed to cope with the c pawn were it to remain instead

however i think that the knight could not continue to cope with protecting both its pawn and harrassing the light squared bishop

forget it if 2 bishops are there

if the black king stays locked in and the white dark squared bishop gets out it is basically 3 pieces v 1 piece and 2 (connected) ,1 or 0 pawns

(i hope i do not sound like either karsten mueller or the late mikhail botvinnik :rolleyes: :eek: :confused: )

pax
03-02-2006, 01:31 PM
After 1...Nd5, it looks like a pretty comfortable win for white:

e.g 2.Kc5 b4 3.Bc4 Ne3 4.Bb3 Nf5 5.Kxc6 marching into e7/e8 for the killing blow.

1...Na5 is better, because the key black squares are covered for a little while, but it still looks like a win for white:

e.g 2.Ke5 Nc5 3.Bf5 Nb7 4.Bc8 Nc5 5.Kd6 Ne4+ 6.Kxc6 b4 7.Bf5 Ng3 8.Bc2 ...

or 3... b4 4.Kd6 Nb3 5.Bd3 Nd2 6.Be2 Ne4+ 7.Ke7 Nc3 8.Bc4 Nd5+ 9.Ke8 Nc7+ 10.Kd7 Nd5 11.Kxc6 ...

Kevin Bonham
03-02-2006, 02:03 PM
After 1...Nd5, it looks like a pretty comfortable win for white:

e.g 2.Kc5 b4 3.Bc4 Ne3 4.Bb3 Nf5 5.Kxc6 marching into e7/e8 for the killing blow.

5...Nd4+.


1...Na5 is better, because the key black squares are covered for a little while, but it still looks like a win for white:

e.g 2.Ke5 Nc5 3.Bf5 Nb7 4.Bc8 Nc5 5.Kd6 Ne4+ 6.Kxc6 b4 7.Bf5 Ng3 8.Bc2 ...

Can't find a forced win yet after 8...Nf1. eg 9.Kd7 Ne3 10.Bb3 Nf5 harassing the crucial squares. It seems to be really hard to win after eliminating both pawns. The knight just hops around forever and as soon as it is taken it is stalemate. I don't think the king and LSB alone can screen it out from policing e7/e8.


or 3... b4 4.Kd6 Nb3 5.Bd3 Nd2 6.Be2 Ne4+ 7.Ke7 Nc3 8.Bc4 Nd5+ 9.Ke8 Nc7+ 10.Kd7 Nd5 11.Kxc6 ...

7...b3 draws.

I suspect it is a draw with either 1...Nd5 or 1...Na4.

pballard
03-02-2006, 04:32 PM
It seems to be really hard to win after eliminating both pawns. The knight just hops around forever and as soon as it is taken it is stalemate.

It's not that easy. If the N goes en prise, white can often respond with Bf8, winning. I'm not sure if that ending (after winning both pawns) is actually winning, but black has lots of chances to go wrong.

pax
03-02-2006, 04:41 PM
5...Nd4+.
Oops, I interposed my lines. In the above line, the knight was on d5, not f5.




Can't find a forced win yet after 8...Nf1. eg 9.Kd7 Ne3 10.Bb3 Nf5 harassing the crucial squares. It seems to be really hard to win after eliminating both pawns. The knight just hops around forever and as soon as it is taken it is stalemate. I don't think the king and LSB alone can screen it out from policing e7/e8.


After the above:

11.Bc4 b3 12.Bxb3 then black loses rapidly e.g 12...Nd6 13.Bf8! or 12...Nd4 13.Bd5 Nf5 14.Ba2 Nd6 15.Bf8!

This Bf8 idea kills most of black's drawing plans dead, because he cannot longer leave his knight en prise.



3... b4 4.Kd6 Nb3 5.Bd3 Nd2 6.Be2 Ne4+ 7.Ke7 b3 draws

No, 8.Bc4! Ng5 9.Bxb3 c5 10.Bc4 Nmoves 11.Bxc7mate

Phil Bourke
03-02-2006, 04:44 PM
Don't you mean 1 ... Na4, 1 ... Na5 is illegal :)

pax
03-02-2006, 04:49 PM
Don't you mean 1 ... Na4

Yes! (oops again)

Kevin Bonham
03-02-2006, 05:15 PM
11.Bc4 b3 12.Bxb3 then black loses rapidly e.g 12...Nd6 13.Bf8! or 12...Nd4 13.Bd5 Nf5 14.Ba2 Nd6 15.Bf8!

This Bf8 idea kills most of black's drawing plans dead, because he cannot longer leave his knight en prise.

Agreed (and agreed with pballard's comments too - Bf8! is a big threat if the knight is not close enough to deal with it).


No, 8.Bc4! Ng5 9.Bxb3 c5 10.Bc4 Nmoves 11.Bxc7mate

Now I'm mixing up lines, I had the black king on d7. So after 7.Ke7 Nc3 8.Bc4 Nd5+ 9.Kd6 and white can clean up the pawns.

Looks a lot like it's a win in both lines by the same sort of method now.

qpawn
11-02-2006, 05:07 PM
Thanks sincerely for a most intrigueing position. I haven't put it on Fritz. Without looking at specific lines my intuition, for waht it is worth at my meagre standard of play, is that it's drawn. I can't see white getting a pawn queened or mating with 2 bishops. Black has stalemate prospects unless white sacs a bishop. Hence with the threat of stalemate black's knight can go enprise as it pleases. I think that the black pawns can be rounded up easily bearing in mind the stalemate mentioned. :)

Kevin Bonham
11-02-2006, 05:34 PM
Fritz is pretty useless in an ending like this because it takes it way too long to smell the coffee either way. However when using Fritz to play around with this ending I did find that once its evaluation went into the +5s it was invariably a win, whereas a +4 was no guarantee.

pax
11-02-2006, 06:45 PM
Fritz is pretty useless in an ending like this because it takes it way too long to smell the coffee either way. However when using Fritz to play around with this ending I did find that once its evaluation went into the +5s it was invariably a win, whereas a +4 was no guarantee.

Yes, you have to be careful with computer engines in endgames. If you find that all the lines are coming back with the same score (even as high as 4 or 5), it's a pretty good indication that it doesn't understand the position.

pballard
11-02-2006, 08:54 PM
Thanks sincerely for a most intrigueing position. I haven't put it on Fritz. Without looking at specific lines my intuition, for waht it is worth at my meagre standard of play, is that it's drawn. I can't see white getting a pawn queened or mating with 2 bishops. Black has stalemate prospects unless white sacs a bishop. Hence with the threat of stalemate black's knight can go enprise as it pleases. I think that the black pawns can be rounded up easily bearing in mind the stalemate mentioned. :)

But as had already been mentioned, the black knight cannot go en prise "as he pleases", because in many cases white can respond Bf8. If black takes the B, white takes the N and it's an easy win (B + 2P vs. 2P). If black doesn't take the B then it's 2B + 2P vs. N + 2P which is still (I think) an easy win, at least in this position.

But when I looked at it a week ago, it was not clear whether Black was forced to allow this to happen. There were lots of ways to go wrong, but it did seem (to me) like he could hop his N merrily around the board, and not get trapped by the white K + white-squared B, with best play.

(All this is assuming white can mop up the 2 Q-side pawns first. I think there is agreement on this).