Tricky K+P endgame

White to play and win.

Hey.  I guess that blocking technique would work.  1 Ke5!  1...Kc3 does not work b/c of 2 f4.   1...Kb5 would be correct but 2 Ke6 wins.  I'll wait for other posts to find out actual solution.  Thanks for posting.  Cheers.

I'd guess Kd4.  The pawn doesn't want to move closer and the king can't approach.  If Kb3 Ke5 and if Kb5 then Kd5.  The first move is a nice example of zugzwang.

The idea, of course, is to capture the pawn when the black king can't land on any of the key squares (usually simply the three in front of the pawn one rank away e.g. in the starting position e4, f4, and g4).

This is quite fun.

Now for another diagram explaining everything that could go wrong.

Key idea here is to keep the black king far enough behind so that he can't "touch" the white pawn before the black pawn is captured. To do that, he must be pushed away (shouldered) until white has a big enough lead to push his pawn forward and take black's pawn.

Waffllemaster got it. [EDITED SINCE MY EARLIER ANALYSIS FAILS]. The idea is not the e4-f4-g4 squares, but rather knowledge of black's saving idea to play "tag" and touch the white pawn just before his pawn falls. To do that, white must block black off for tempi, not position.

Well, it was too tricky for White in Malakhov-Najer (Moscow 2007). Malakhov played the natural 1. Ke5? (which would have been my first choice too) and threw away the win. The position is from Dvoretsky's Tragicomedy in the Endgame.

Remellion doesn't mention the refutation of my second choice 1. Kd5? Black attacks the pawn from behind with: 1...Kc3 2. f4 Kd3 3. f5 Ke3 4. Ke5 Kf3 5. f6 Kg4 6. Kd6? Kf5 7. Ke7 Kg6

Also 1. Ke5? Kc4? 2. f4 Kd3 3. f5 Ke3 4. f6 Kf3 5. Kd6 Kf4 6. Ke7 wins for White.

The right defence is 1. Ke5? Kc5 2. f4 Kc6 when neither 3. f5 Kd7 nor 3. Kf6 Kd5 work for White.

Ooops. Analysis fail on my part with 1. Ke5? Kc4? there. Sigh, complicated stuff! XD

Ah, 1. Kd5? was actually the first line I looked at (and evaluated as such, couldn't remember to refute it.

And the illustrative 6. Kd6?? shows how badly even a simple K+P endgame can go. Instead, 6. Ke4 holds the draw with opposition: 6. Ke4 Kg5 7. Ke5 Kh6 8. Kd6 Kg5 9. Ke5 Kg6 10. Ke4! Kxf6 11. Kf4, =

This time, the analysis shows the highly instructive 1. Kd5? line. Aside from the earlier-mentioned "tag the pawn" idea, here we also see the themes of opposition and mutual zugzwang, as well as the need to play "only moves" for both sides. Hopefully this is useful for those still learning K+P endgames. (Well, I'm still learning too, Basic Chess Endings is way hard.)

Sometimes king moves can be confusing.  A trick I've found for myself (FWIW) is to think in terms of either moving by ranks, files, or both.  Then you just find out which is important...

That may sound weird but in this example after 1.Kd5 we may not think 1...Kc3 is very intimidating because it's moving down... but files are what's important, this move only loses time if ranks were critical.  Regardless of losing a rank it's still makes progress.  And as this shows (and in a number of different endgame actually) as long as you're within two ranks of the enemy king (come from behind) then you're still ok.

Now if ranks and files were critical you'd have to do a diagonal move of course.  But anyway, this has helped me simplify some odd king maneuvers in the past.

Hey wafflemaster.   Would you mind developing you explanation a bit? Although it is interesting and read it a couple of times it isn't clear to me.  Thanks.

Hold on.  Do you mean opposition must be won on d-file because e6 is not accessible for white king?

If 2. ... Kd3, then 3. f4 Ke4 5. f5, and 6. Kxf7.

@Gshapiroy : 2...Kd3 3.f4 Ke4 4.f5 Kf4! 5.Kxf7 Kxf5 - that's the defensive idea

Oops, just now saw this.

Maybe it's too simple heh.  All I'm saying is with a king move you're either moving left/right, up/down, or one of each (diagonal).

For the original position white's first move doesn't allow black's king to move to the right.  Some may have chosen 1.Kd5 where the response 1...Kc3 doesn't seem threatening, but 1...Kc3 moves to the right so it's making progress.

In this problem this idea doesn't independently explain the strength of white's move.  White's move puts black in zugzwang, which you would have to calculate to know.

Thanks.  I actually thought it was more complicated than that so I appreciate your clarification ---I thought the wafflemaster dude has a unique way to look at these endgames and I definitely want to know about it---.  Let me suggest an idea about the OP's problem that I would appreciate you let me know your thoughts.  I am trying to think a different possibility to shouldering.  e6 is not accessible for white's king because is attacked by black's pawn so  white needs to win opposition on d-file and that is why Kd4 works but not Kd5.  

That's more or less right. Kd4 and Kd5 both let white take the same number of moves to capture the f-pawn (since e6 is not available as you said.) But it also delay's black's king from getting to the important part of the board by one move ("opposition" so to speak; white waits for black to say "I'm going up/down" and reacts accordingly so the king is too far off either way.)

Of course, that applies to this position. Most of the time, shouldering can be done without the explicit opposition, like in this cute position:

(Taken from GM Serper's article on this site on K+P endings. It's a great article, and the last 2 positions are relevant here.)

yeah need to learn those endgame thingy.so weak in them.

@Remellion.  Thanks!

I don't particularly like thread necromancy, but a little something to repay caveatcanis's endgame puzzles. Plus it's another K+P v K+P. Again, white to play and win:

1. Kd4 Kb4 2. f4 or 1...Kc6 2. Ke5 Kc5 3. f4

Instructive position.