Saturday, December 28, 2013

Deep calculation endgame tactic



White to move. Is this a win, draw, or loss for White? How should White proceed?

8/2pp2pp/8/2PP1P2/1p5k/8/PP4p1/6K1 w - - 0 1

21 comments:

Oleg Mezjuev said...

1. Kxg2!

The only move. Not 1. a4? which leads to forced checkmate after 1. ... Kg3 2. a5 h5 3. a6 h4 4. a7 h3 5. a8Q h2#

2. ... Kg4
2. a4 bxa3
3. bxa3 Kxf5
4. a4 Ke5
5. d6! cxd6

White is starting to create a blockade in the path of the black King.

6. c6! dxc6

The blockade is complete. Black King now has a longer way to go to a8.

7. a5 Ke6

Or 7. ... Kd5 8. a6 Kc5 9. a7 Kb6 10. a8! and white wins.

8. a6 Kd7
9. a7 Kc7
10. a8Q 1-0

Anonymous said...

A4 and keep pushing the passed pawn

Anthony Basilicato said...

a4 does not look right. After Kg3 it looks like white gets mated by h2.

Anonymous said...

8:26

A4 loses to Kg3 trapping the King; even though White Queens he can not stop h7-h2 mate

after 1Kg2 Kg4, Black's King makes it into the queening square, so white needs to find a way to block Black's King from beeping to a8.

Anonymous said...

8:26

A4 loses to Kg3 trapping the King; even though White Queens he can not stop h7-h2 mate

after 1Kg2 Kg4, Black's King makes it into the queening square, so white needs to find a way to block Black's King from beeping to a8.

Bob said...

It's a win for White, but only with precise moves. For example, 1. a4?? loses after 1. ... b4xa3; 2. b2xa3, Kg3!; 3. a4, h5; 4. a5, h4; 5. a6, h3; 6. a7, h2 mate.

So White has to waste a move by taking the g pawn before Black plays ...Kg3. But he can't do it immediately.

If 1. Kxg2, Kg5; 2. a4, b4xa3; 3. b2xa3 then Black loses by playing ... Kxf5?

It's too early for the interference theme with 4. d6, cxd6; 5. c6, dxc6; 6. a4 when ... Ke5?; 7. a5, Kd5; 8. a6 and the c6 square is unavailable to the Black King so the a pawn queens, but ... Ke6!; 7. a5, Kd7; 8. a6, Kc6 catches the a pawn and wins. Of course, 6. ... Ke6 was only possible because the White pawn was no longer on d5.

So White first plays 1. Kxg2, Kg5; 2. a4, b4xa3; 3. b2xa3, Kxf5?; 4. a4!, Ke5 and now the interference theme works -- 5. d6, cxd6 (5. ... c6 is no better; 6. a5, Kd5; 7. a6 and the c6 square is not available to the Black King); 6. c6, dxc6; 7. a5 and Black can't catch the a pawn -- the c6 square is blocked and the Black King is a move farther away from the e7, d7, c7 route.

Alas, after 1. Kxg2?, Kg5; 2. a4, b4xa3; 3. b2xa3, Kf6! avoids those problems and wins for Black: 4. a4, Ke7; 5. a5, Kd8; 6. a6, Kc8. The interference theme can't stop that.

After all that, the correct sequence is obvious. 1. f6!, gxf6; 2. Kxg2!, Kg5; 3. a4!, b4xa3; 4. b2xa3, Kf5; 5. a4!, Ke5; 6. d6! (6. c6?, d6! and c6 isn't blocked so 7. a5, Kxd5; 8. a6, Kxc6 wins for Black) , cxd6; 7. c6, dxc6; 8. a6 and White wins. But this is a very narrow path and any deviation by White lets Black win instead.

A fun puzzle.

Luc Despontin said...

Not so easy anonymous, because after a4?? black plays Kg3 and push the A pawn until mat.

Luc Despontin said...

Sorry, read: black push the H pawn...

Anonymous said...

kg2 to stop mate and push a pawn
black king comes close play d6 and c6 to obstruct its way. white wins

chessdrummer said...

a4 immediately doesn't work. Black will take ep and your pawn will be on a3. You'll get mated with Kg3 and ...h5...h4...h3...h2 mate is too fast.

Anonymous said...

If immediate 1.a4 Kg2! and white will be checkmated first by the black's h pawn. White must therefore play Kxg2 first and then followed by a4.

mshroder said...

First try:

1. a4 bxa ep
2. bxa h5
3. a4 Kg3
4. a5 h4
5. a6 h3
6. a7 h2#

Let's try that again

1. a4 bxa ep
2. bxa h5
3. Kxg2 Kg4
4. a4 h4
5. a5 h3+
6. Kh2 Black is doomed

Can black improve on move 2?
Third try:
1. a4 bxa ep
2. bxa Kg3
3. a4 h5
4. a5 h4
5. a6 h3
6. a7 h2#

So white can't allow the black king to move to g3, hence
Fourth try:
1. Kh2 Kg4
2. a4 bxa ep
3. bxa h5
4. a4 h4
5. a5 h3
6. a6 g1/Q+
7. Kxg1 Kg3
8. a7 h2+
8. Kh1 and white wins.

I think that's the solution.

Mark

Menno van Koningsveld said...

Well, after bxa, bxa, Kg3 white gesTs mated doesn't he?!

Anonymous said...

1. f6! gxf6 forced (1...Kg3, intending h5, etc to h2# does not work due to 2. fxg, f8=Q+)
2. Kxg2. Kg5
3. a4 bxa3 ep
4. bxa3 Kf5
5. a4 Ke5
6. d6!! cxd forced but cutting off one access to c6 for Black king
(If 6... c5 Black King cannot catch the pawn
7. a5. Kd5
8. a6 Kxc5
9. a7 and white queens)
7. c6!!! dxc forced but cutting off the other access to c6 for Black king; (otherwise c7 and c8=Q)
8. a5 Kd5
7. a6 Kc5
10 a7 and white queens.

All other Options lose for white!!!
eg.,

In above variation,
A 6. a5 Kxd5
7. a6 Kc6!
8. Kg3 d6!
9. cxd6 cxd6
10 Kg4 d5! and black catches white a pawn and black queens one of d, f and h pawns!

In above variation,
B 6. c6 d6
7. a5 Kxd5
8. a6 Kxc6 and again black catches white a pawn and queens one of his d, f or h pawns!

Finally

C
1. a4. bxa ep
2. bxa3 Kg3
3. a4 h5
4. a5 h4
5. a6 h3
6. a7 h2#

D
1. Kxg2 Kg5
2. a4. bxa ep
3. bxa3 Kxf5
4. a4 Ke5
5. a5 Kxd5
6. a6 Kc6! as in variation A and black wins!
7. Kg3 d6!
8. cxd6 cxd6
9. Kg4 d5! and black catches white a pawn and black queens one of d, f and h pawns!

Harry

Awfulhangover said...

Not 1.a4, coz 1.Kg3 followed by h5-h4 etc, so white must play 1.Kxg2
But then it looks like black will catch a-pawn by going to the queen side. So, there must be something tricky with the c and d-pawns that makes black's king route longer

Anonymous said...

If 1.a4 ba3 2.ba3 Kg3 and pushing the h-pawn, black will mate after white's a7
I'm not sure in detail, but i think the idea is something like blocking the black king by his own pawns. for ex.: 1.f6 gf6 2.Kg2 Kg5 3.a4 ba3 4.ba3 Kf5 5.a4 Ke5 6.d6 cd6 7.c6 dc6 8.a5 and the black king should be late to stop the white pawn - but there might be something i miss...

Anonymous said...

1. a4 Kg3 and white can't prevent h5-h4-h3-h2 mate, even though white promotes his pawn

Anonymous said...

f6 is obvious because it's the least obvious move.

Anonymous said...

f6 is obvious because it's the least obvious move.

Yancey Ward said...

The solution is the anonymous commenter at 4:55 AM CST and Harry.

Cortex said...

Troitzky, Deutsche Schachzeitung, 1913)

1.f6 correct move.

Alas, 3.a3 does the trick too. (Mario Campioli, Finales y Temas, 2001)

So, this interesting study is cooked!