Pawn Endings on Autopilot (1)

White to play and win without working hard.

As you may remember from earlier in the week, I left you the above position (White to move) as a test of your abilities in Pawn Endings. Suffice it to say that it's one of the easier pawn endings to win that you may see, however there are still tricks and nuances you must be familiar with, and this remains true if you're playing it against a beastly engine to test yourself. I can thoroughly recommend this, in fact. Take a batch of positions you reached in a certain sort of ending, and play them against your engine, at a reasonable time control that mimics how much you are likely to have left on your clock in real life.

We will come back to the idea of engines endings training later in the week, and I will post a batch of positions for people to use in precisely this manner. For now though, let's take a stroll through the above ending, and see how I got on against Houdini 1.5a earlier tonight at 25 minutes each. Houdini unleashed rates itself at about 3182, though we know certainly that this isn't representative of a World Champion level of endgame knowledge. This will make for some interesting posts in the next few months as I dig out some very difficult and some deceptive pawn endings from my own play. Even good club players fall foul of subtleties all the time, so the more you can play such positions by hand, the better.

Before you enter a pawn ending, you should always know whether it's won, drawn or lost, or at least be able to eliminate one of those options. Another tip I can give you is to always treat the first move of a pawn ending as critical. It is a good chance to break the shackles of a middlegame mentality, and move towards pure calculation. In some such endings (as we will see in later entries) there is a critical first move, which makes your task much easier or that of your opponent much harder. This position is more a case of working out the possible ways things might unfold, and reacting most flexibly to that. This and subsequent posts assume that the reader has at least a basic knowledge of pawn endings no matter his grade.

White having a Kingside majority, is winning, especially so since all of Black's poor pawns are isolated. This isn't to say they aren't useful though. His f-pawn and c-pawn might be used as splitters, to try to fracture my structure on King or Queenside, to make things more difficult or set traps. One other advantage is that White has spare tempo on his side: he has more optional pawn moves which do not compromise the structure than Black. This alone indicates that if we get to an opposition situation, it may be winning for White anyways. Knowledge such as this can help a beginning club player to maximise his chances from such positions in the face of stronger opponents.

Let's imagine you had five minutes left to play this as White, and therefore couldn't afford the luxury of taking all the time you needed to calculate everything you might have difficulty with. Where should you go first?

Since Black's pawns are so weak it stands to reason that we should see what happens if he goes route 1, and tries to promote on the Queenside. To do this, he needs to play Kd5, c4, and after the exchange, get to the back of the a-pawn, eat it, and push the a-pawn all the way himself. In the meantime, if allowed, White would eat the f4-pawn, and push his f-pawn all the way up the board. 

Pure counting shows that White would need five moves to eat the pawn and move out the way, five to promote his pawn, and one to eat the pawn that will come to c4, for a total of 11. Black meanwhile needs three moves to get his King to c4 following said exchange, five to promote his pawn, and four more to get his King off a2 once he eats the pawn there, which comes to 12. So we needn't worry about this plan from Black at all, without any calculation.

There are other ways for White to play though. He could play g3 straight away, and if Black brings the King forward, then follow up with f3: forming a barrier against further King penetration. A player who isn't as strong at deeper calculation (or equally has very little time left) might worry that Black would try 1. g3 f3 to create disruption to White's optimal King path. Indeed there are difficulties then. If White sees the idea of blocking off the Black King on the Kingside and follows up with 2. h3 Ke5 3. Kf1 Kd4 4. Ke1 Kc3 5. Kd1?! a5, he may well be too slow to win any subsequent race.

(Remember - we're looking at the safest way to play the position so as to avoid leaving any traps that can be set. It doesn't matter that however many moves actually win here. What matters is being able to bosh out your moves in however severe time trouble.)

Equally, White should realise that playing f3 will leave e3 weak, as a potential nagging square for the Black King if he tries to turn the other way. As a practical, defensively-minded measure, White could therefore avoid such things if he were playing safety first. So we play 1. Kf1 Kd5 2. Ke2 Ke4.

Now White's extra pawn moves come in handy to force clarification, since obviously we cannot make progress with the King position in the meanwhile.

If you were being lazy, you might well see that Black threatening ...f3+ gxf3+ Kf4 is annoying (again no calculation here!), so White plays 3. g3 f3+ 4. Kd2 Kd4 to cut that out. By keeping his f-pawn back, he creates a forbidden zone into which the Black King can never sneak. He is also stopping any ideas Black might have had, of pushing the h-pawn to h4 to try to eat up some of White's spare tempi.

 
Next comes 5. h3 a5 6. g4 Ke4 ensuring that the h-pawn will remain frozen, and planning to just advance the pair together, up the board. Black meanwhile is maximising his chances to set traps Queenside. Now, Black will not be able to come behind the g/h-pawn pair until White is far enough advanced that it won't present a problem. 

Let's say you'd spent 2 of your 5 minutes on the initial problem of counting moves to direct Queening for both sides. Continuing on autopilot, you should now notice that Black is threatening to come around the back, via Kf4-g5-h4xh3-g2xf2, when he would, if we err, be able to promote his hard working f-pawn. Of course if you had time you could calculate that even this fails to 7. Kc3 Kf4 8. Kxc5 Kh4, BUT only if you follow up with the immediate 9. g5 or 9. Kd4. The easier way to approach this is to play 8. Kd5 Kh4 9. Ke4 Kxh3 10. Kxf3, ensuring that Black's King has nowhere to go to play his tricks.

Black therefore tries his last trick, playing 10...Kh2 and giving you a hard stare. After the now obvious 11. g5 Kh3 12. Kf4 Kg2, he is hoping you will play 13. Kg4??

which only draws, thanks to 13...Kxf2 14. Kh5 Kf3 15. Kh6 Kf4.

Instead, White will win after playing 13. f3 Kf2 14. Ke4 Kg3 15. f4 Kg4 16. f5 Kxg5 17. Ke5 h5 18. f6 Kg6 19. Ke6 h4 20. f7 Kg7 21. Ke7 and promoting next move.

I hope this helps some of my younger/less experienced readers to get some practical grips on pawn endings in a way that looks at it primarily from a strategical viewpoint. Let me tell you though, that to play the game to its fullest, there is no substitution for calculation, and training yourself to be able to visualise clearly and properly to correctly assess such positions and their evaluation. A lot of the time, you will be able to verbalise your thoughts in such a way only because your calculation of the underlying variations is fluent. If you can't do that, you may find you have to switch the brain on completely and see everything you can, only after the event finding ways to codify your learning.