PDA

View Full Version : Pawn ending from tournament game



Kevin Bonham
17-11-2009, 08:20 PM
8/7p/p2k2p1/1p4P1/1P1K4/7P/P7/8 w - - 0 1

I had this in a G60 flat time scramble in 2001 with both players very short of time (opponent was more so), got it wrong :wall: , but ended up winning in a rather odd manner.

White has better king position and two pawn moves to burn, so should be winning this, and is. There are three candidate moves here: 1.a3, 1.h4, 1.Ke4.

Two win overwhelmingly but the third does not (in fact, it leads to a tablebase draw). Which is the odd one out, and why?

If you want to test your time-scramble skills, try solving this in a minute or less.

Answers in white text please.

Alexrules01
17-11-2009, 09:18 PM
1. h4 wins

1. h4 Ke6 2. Kc5 Kf5 3. Kb6 Kg4 4. Kxa6 Kxh4 5. Kxb5 Kxg5 6. Ka5 Kf5 7. b5 g5 8. b6 g4 9. b7 g3 10. b8/Q and white should win.

And I believe 1. Ke4 Ke6 2. h4 Kd6 3. Kd4 Ke6 4. Kc5 leads to the same line as above. But there might be something better

1. a3 draws

1... Ke6 2. Kc5 Kf5 3. Kb6 Kxg5 4. Kxa6 Kh4 5. Kxb5 Kxh3 6. Ka5 g5 7. b5 g4 8. b6 g3 9. b7 g2 10. b8/Q g1/Q

Tad longer than a minute

Phil Bourke
17-11-2009, 09:33 PM
I took longer than the 1 minute, more like 15 :), I think that h4 and Ke4 are the choices. I admit that my original liking was for a3, and in the the given circumstances, probably would have played it.

Capablanca-Fan
18-11-2009, 04:12 AM
Alexrules01:


1. h4 wins

1. h4 Ke6 2. Kc5 Kf5 3. Kb6 Kg4 4. Kxa6 Kxh4 5. Kxb5 Kxg5 6. Ka5 Kf5 7. b5 g5 8. b6 g4 9. b7 g3 10. b8/Q and white should win.
Right conclusion, but you forgot the rook's pawns. 6. a4 is correct, queening and controlling the square of the opposing P. 6... h5 is correct with both Ps queening.

Spiny Norman
18-11-2009, 04:24 AM
Since I only have 1 minute up my sleeve, I will say (without calculation):

The move to win is 1.h4 ... and my reasoning (possibly dodgy) is:
* that the pawns on the a-side are closer to Black's queening squares, so if there be any hope for Black it would be that they get into a race to queen a pawn. By holding back the a-pawn, Black would have to make one additional king move in order to capture that pawn before trying to queen his own pawns.
* after 1.h4 Black's king must give way, allowing White to penetrate the position.

So in answer to the question, I will propose that 1.a3 is the odd move out and allows Black to draw, given that 1.Ke4 just maintains the status quo.

Kevin Bonham
18-11-2009, 09:34 PM
I'll post the answer to this and the simplest way to explain it (as well as the peculiar finish) tomorrow night.

Jesper Norgaard
19-11-2009, 04:32 AM
I'll post the answer to this and the simplest way to explain it (as well as the peculiar finish) tomorrow night.

I'll give it a try. The move 1.Ke4 just "passes the ball" so to say. The real choice is between 1.a3 and 1.h4. It is clear that white cannot by force go to capture g6 and h7, so if he capures a6 and b5, if the a-pawn is on a2 or a3 doesn't matter in the race. In the end it comes down to if g5,h4 or g5,h3 is the optimal pawns position on the kings side.

It turns out that g5,h4 is optimal because of a peculiarity in the position - white a-pawn promoted can prevent black h-pawn from promoting. I guess all these intellectual obstacles makes it almost impossible to grasp why g5,h4 is optimal from a common sense perspective, you need to calculate, but the variation is clear 1.h4,Ke6 2.Kc5,Kf5 3.Kb6,Kg4 4.Kxa6,Kxh4 5.Kxb5,Kxg5 6.a4!,h5 (6...Kf5 and 7.g5 loses a tempo in the race) 7.a5,h4 8.a6,h3 9.a7,h2 10.a8=Q and curtains because of 10...h1=Q 11.Qxh1.

On the other hand 1.h4 doesn't win a tempo as I intuitively thought because black needs an extra tempo to play Kg5-h4xh3, however, now the g-pawn has a free ride instead of the h-pawn and black draws (if 1.a3,Ke6 2.Kc5).

Kevin Bonham
20-11-2009, 12:02 AM
OK, people have been getting the right answer but some of the early answers had some slightly incorrect reasoning. Jesper and Jono are both completely correct.

The easy path to the idea that 1.h4 is right and 1.a3 is wrong is (in my view) as follows.

If black just shuffles his king between d6 and e6, then white can only win by gaining the opposition then going via c5 to grab the black a and b pawns.

So white moves a pawn, and black plays Ke6. Now white queens in nine moves no matter which pawn he moved:
Kc5, Kb6, Kxa6, Kxb5, a4, a5, a6, a7, a8.

Black also queens in nine moves no matter which pawn white moved. If white moved the a-pawn then Kf5, Kxg5, Kh4, Kxh3, g5, g4, g3, g2, g1

If white moved the h-pawn then Kf5, Kg4, Kxh4, Kxg5, h5, h4, h3, h2, h1

But in the first case the black queen lives. In the second case it immediately perishes to Qxh1.

Therefore 1.h4 wins (1.Ke4 followed by h4 also wins) but 1.a3?? leads to KQP vs KQP which is drawn. The issue is not speed of queening but control of the black queening square.

This is how the game actually finished.

1.a3?? Ke6 2.Kc5 Kf5 3.Kb6 Kxg5 4.Kxa6 Kh4 5.Kxb5 Kxh3 6.a4 g5 7.a5 g4 8.a6 g3 9.a7 g2 10.a8=Q g1=Q 11.Qc8+? (better 11.Qf3+ when black might lose the pawn and have some work to do) 11...Qg4 12.Qxg4+ Kxg4 13.Kc5 h5 14.b5 h4 15.b6 h3 16.b7 h2 17.b8=Q h1=Q

Black could have reasonably claimed 10.2 in the last few moves but he did not do so.

As Black played 17...h1=Q his flag fell. I noticed this immediately but he didn't.

Under the 1997 Laws, at least as we interpreted them, G60 flats were covered by the Rapidplay rules. (The 1997 Rapidplay laws called a rapid "between 15 to 60 minutes", and clearly G15 flat was a rapid since blitz was "less than 15 minutes", so it seemed logical that the "between" was inclusive at the other end too.)

So although my opponent had flagged the game wasn't over unless I claimed it, which with KQ vs KQ on the board would be mean. On the other hand, agreeing a draw when my opponent could have claimed 10.2 and had instead kept playing and flagged didn't much appeal to me either.

I made a snap decision to postpone the dilemma and played 18.Qg8+. After ...Kf3?? (anything non-losing here and I would have felt obliged to concede the draw) 19.Qa8+ black resigned.

Desmond
20-11-2009, 10:07 AM
I made a snap decision to postpone the dilemma and played 18.Qg8+. After ...Kf3?? (anything non-losing here and I would have felt obliged to concede the draw) 19.Qa8+ black resigned.Something to be said here for the player who makes the second last mistake. That's the stuff of nightmares.

Kevin Bonham
20-11-2009, 12:33 PM
Something to be said here for the player who makes the second last mistake. That's the stuff of nightmares.

The whole game was atrocious. I was totally winning out of the opening (pawn up with huge attack on move 12) and my play after that was one bungle after another against a lower-rated player who I would normally have put away in <25 moves from such a position.

He actually played really well to get back into it, albeit with a lot of help from me. The only reason for the blunder at the end is he was moving instantaneously thinking his flag was teetering and unaware that it had actually already fallen.

Phil Bourke
20-11-2009, 02:23 PM
Pinched it :) Going to use it as a lesson for the school students, lots of good little things in it, hope it sticks in their heads for their own games.

Kevin Bonham
20-11-2009, 02:37 PM
Pinched it :) Going to use it as a lesson for the school students, lots of good little things in it, hope it sticks in their heads for their own games.

I think it's quite a good one for this sort of purpose.

The most important lesson from one of these is that when you get into a pure pawn ending (or are even thinking about swapping into one) with lots of time on your clock, be willing to spend a lot of time calculating to make sure you get it right. Of course, in this case I didn't have that option, which is why I made the mistake.

So many games are won or lost by players playing too fast in pure pawn endings.

Jesper Norgaard
22-11-2009, 07:29 PM
So many games are won or lost by players playing too fast in pure pawn endings.
Gustavo Falcon - Jesper Norgaard 1994. Actually this is an example of how tenacious pressing on can sometimes give bonus even if it should have been a draw. Black is a pawn up but 3 pawns vs. 2 on the same flank and rook ending should be a draw. I continue to set problems for him though:

28.Rd7! (28.Rd8+ Kf7 29.Rd7+ Re7 would be much worse of course) 28...b5 29.Rc7 bxc4 30.Rxc6 cxb3 31.cxb3 a5! {still trying to avoid pawn exchanges too early} 32.Ra6 Re5 33.Rb6 h5! 34.b4 a4! {no changes as long as it can be helped} 35.Ra6 Re1+ 36.Kf2 Ra1 37.b5?! (37.h4! would stop my initiative on the King's side, and it must be a draw) 37...Kf7 38.b6 g5! 39.h4!? g4! 40.Ra5! Rb1 (40...Kg6?? 41.Rb5+- would simply lose, so black can no longer avoid the change of a-pawn for b-pawn) 41.Rxa4 Rxb6 42.Ra2? {passive} 42...Rb3 43.Rc2 g3+!? {trying to put a wedge in between the pawns to maximize the chance that the h4 pawn will fall} 44.Ke2 Ke6 45.Rc6+ Ke5 46.Rc5+ Ke4! (yet another trap, if 47.Rxh5?? Rb2+ 48.Kf1? Ke3 or 48.Kd1 Rxg2 white loses) 47.Rc4+! Kd5 48.Rf4 Ke5 49.Rf3 Rxf3 50.Kxf3 Kd4! 51.Kf4 Kd3 52.Kf3 Kd2 53.Kf4 Ke2 {the moment of truth, now white has to take the pawn} 54.Kxg3 Ke3 {again my opponent spent a long time deciding between Kh3 and Kh2, and again he chose right!} 55.Kh2! (55.Kh3?? Kf4-+ 56.Kh2 Kg4 57.Kg1 Kxh4 58.Kf2 Kg4 59.Kf1 Kg3 60.Kg1 h4 61.Kf1 h3 62.Kg1! Kg4! 63.Kh2 hxg2 64.Kxg2 f5-+) 55...Kf4 56.Kh3 {only move} 56...f5 57.Kh2 Kg4 58.g3??-+ (58.Kg1!= Kxh4 59.Kf2 Kg4 60.Kf1 Kg3 61.Kg1 h4 62.Kf1 h3 63.gxh3 Kxh3 64.Kf2 Kg4 65.Kg2 would draw but after surviving one nasty trap after the other, white finally perishes) 58...Kf3 59.Kh3 {all fine and dandy, on Kf2 Kh2, and on Ke2 Kg2 but ...} 59...Ke3! 60.Kg2 Ke2 61.Kh1 Kf3 62.Kh2 Kf2 63.Kh3 Kg1 {after 64.g4 fxg4+ there is really no hope left, so 0-1}

Desmond
23-11-2009, 09:04 AM
Nice win