Bocajaka wrote on 08/14/08 at 19:46:35:


Pawn endings tend to be either pretty trivial (as long as you know, understand, and have practised till automatic, half a dozen basic positions and are familiar with some methods such as the opposition and triangulation) or quite tricky, sometimes fiendishly difficult. I tell students NOT to simplify into a pawn ending unless it is clear that it is trivially won/drawn, or unless the alternatives are clearly worse.

Soltis (in “Grandmaster Secrets - Endings”) laid down the commandment:

“Thou shalt not trade down to K+P unless you can safely bet your first-born child on the result.”

A bit over-the-top perhaps, but he gets the point across.

I certainly would not categorize this one as “very very easy”; it’s a bit tricky in fact.

I suggest you study an endgame primer before tackling these Polgar puzzles, otherwise you are definitely going find the going tough and become discouraged. It's like trying to solve tricky problems in geometry before you've mastered the basic theorems.

The position first of all requires one to know that if it reduces to a single black pawn on g5 plus kings, with White's king somewhere in front of the pawn it’s a draw unless Black can get his own king a) ahead of the pawn and b) with the opposition.

Pushing the e-pawn does not seem to achieve anything, so it is clear that White must either resign and head for the bar, or try to get counterplay by activating his king. However, 1 Kg3? fails to 1...h5!, suppressing counterplay. 

So 1 Kh3, the first real point of which appears when you see that it allows White to answer 1...h5 with 2 g4+! hxg4+ 3 Kg3 Kxe5 4 Kxg4 reaching one of the basic position-types (a draw in this case) that you need to learn. (So in a practical game 1 Kh3 could almost be chosen by elimination in fact.)

So then we examine some other tries for Black: 
1...g4+ 2 Kh4 seems fine for White.
1...Kxe5 2 Kg4 Kf6 allows 3 Kh5 since if 3...Kg7 4 g4 Kh7 stalemate!
1...Kg6 can be met either by 2 Kg4 h5+ 3 Kf3 Kf5 4 g4+! hxg4 5 Kg3 etc., or even 2 g4 as long as White can gain the opposition after Black eventually captures on e5, and it turns out he can: 2...Kf7 3 Kg3 Ke6 4  Kf3 Kxe5 5 Ke3.

If I've made an error or missed anything important I expect someone will soon tell me.

I have already started to notice some endgame patterns and it's getting easier to solve some of the positions. Thanks to all of you.

However, I shall keep on bothering you with another position:

White to move 



In this position I notice that I am going to lose e5 and thus will have to try to get a draw (if my pawn was passed I would try to win!).

So here I tried first 1.b4! with the idea that if he takes it I win, but ofcourse he does not. But I still seriously thought that perhaps I could draw anyway after 1...Kxe5.

Then I looked at 1.Kg3 emmediately 1...Kxe5 2.Kg4 Kf5 3.Kh5 and I felt I could not reach a draw.

So I got discouraged and gave up and looked at the answer, 1...Kh3. 

As you can see I did not look further than 3 moves (which is very little for being an endgame) but again I can not blame myself. I remember when I was starting out and tried to solve a middlegame position, the same thing happend to me, I got discouraged because I did not have a clue about what was going on. Namely I did not see any "ideas", that is tactical motifs etc. Nowadays, middlegames are not very mystical too me, they are more of an calculation exercise if you understand what I mean.

But the question is whether or not I can learn endgames such as in the same way. How easily do you guys see 1.Kh3. Do you even bother to calculate 1.b4 and 1.Kg3 or is it similar to you like calculating a really passive move when you have a mating attack in the middlegame? I'm very curious.

Btw, in this exercise I did not find any use for "corresponding squares". Should I have been able to use it or is this a completely different type of exercise.Don't get me wrong, I have had great use for this in lots of other excerises and it has helped my tremendously.

Willempie wrote on 08/12/08 at 07:42:36:


I will get to rook endgames in a few hundred problems or so and then you can be sure I will terrorise you some more 

So basically now I know corresponding squares and opposition.  How many important thechniques do you believe there are in pawn endgames? 

I suppose wer are all different but personally I like to do it this way, by "doing and learning", instead of reading an instructional endgame book.

Thx guys for the corrections:

Indeed:
1.Kd5, Kf7 2.g6+, fxg6 3.fxg6 (instead of my error 3.hxg) wins.

1.f6+,xf6+ 2.xf6,Kf7 3.Kf5, And now either Ke8 or Kg8 draws
(iso of 3....h6 which is a clear error).

I'm afraid I already took the fun away three posts back in the mirror-image position.

Yes, you are right; I followed Micawber's line rather uncritically. Moving the h-pawn is a terrible mistake as it allows W to set up the winnign maneuver, but W can accomplish nothing if the h-pawn stays on it's original square.

As an example of corresponding squares, just take my line after 3...h6.

As to your general question about endings, my answer would be that endings are all about building blocks. If you look at some of Micawber's terrific analyses, you will see how fruitfully he makes use of this method. These building blocks serve to orientate you; without them, you are looking at deducing everything from first principles, which is time-consuming and prone to error.

Problem 1:

1 f6 wins, as Black gets zugzwanged.

1 f6 gf; 2 gf Kf7; 3 Kf5 h6; 4 Ke5 Kf8; 5 Kf4 Ke8; 6 Ke4...zugzwang.

6...Kf7; 7 Kf5, or 6...Kf8; 7 Ke5.

This is a simple example of corresponding squares. if White plays to one of the key squares, Black has to play to the corresponding square to hold the position.

the corresponding squares are f5 and f7, e5 and f8, e4/f4 correspond to e8/g8. hence Black's problem, as e8 and g8 are noncontiguous (Black would draw if he could play 6...Ke8-g8, maintaining the correspondence.

1...Kf7 is left as an exrecise for the reader