Wednesday, February 20, 2008

How hard is it to see a move?

What is the best move for White in this position? The computer will tell you instantly, but you're not using a computer. You're on your own and you have two minutes to decide. Black's last move was 29...Qg3-c3.

Let's insert a little illustration so that the answer isn't immediately visible to anyone who sees the diagram.

The move White played - the move I played - was 30.Rd5-d2. The winning move - very winning, since Rybka gets to +11 in less time than it takes to type it - is 30.e5-e6.

My opponent saw it. I didn't. Didn't see it, didn't reject it, didn't consider it. Why not? It's an easy win. But does that mean it's a win easily seen?

It's true, of course, that at the board there's no hint that you have a winning move at a given juncture. You have to find it - and then you have to make yourself sure of it. I should think that many or most players would consider 30.e6 in the position, provided it was presented to them as a puzzle. In other circumstances, many fewer would consider it and fewer still would play it.

Moreover it's not just a question of having a heavy hint that there's a winning move available - when you attempt a puzzle it's open to you to speculate, to suggest that it might be such-and-such a move without being sure, to then look up the answer and find out. In a game you don't have the luxury of doing that. You're putting pieces en prise and opening up your king. You need to be pretty sure. Additionally, in this particular game there's the luxury of a safe move being present and obvious, that move being the one I played. In fact, it's more than safe, since it ought to win - though it was neither safe nor winning the way I played it, given that my thirty-first move threw away the win (31.Rc1xc2!) and my thirty-second, the draw (32.Kg1-f1). Are you going to play 30.e6 - maybe winning immediately, but surely losing if it doesn't - when you have the safe 30.Rd2 available?

So, in two minutes - which is how long I had left - you need to see 30.e6. You also need to check it. To see that promotion is unstoppable after 30...fxe6 31.f7 and that the opening of the b8-h2 diagonal does nothing damaging to White's position. You need to check that Black hasn't got any tricks like putting something on d8. You need to be sure of all of this before you play it. And then you have to play it.

Is that easy? I don't know. It probably makes it harder that you need to be looking for it, too. I wrote something last week that might be pertinent here, involving the missed win 38..Bf3! in the featured game: that move being
an aggressive, attacking move when what you're thinking of doing is keeping it tight and not dropping a piece with your clock about to run out.
Possibly that might explain why my opponent was able to see it while I was not: both of us were trying above all to stay out of trouble. Black because he was plainly on the brink of defeat, White because provided I didn't get mated or drop material I was plainly going to win - or so I thought. And in those circumstances you look out for your opponent's potentially dangerous moves, before your own. Or instead of your own. I wasn't playing over-cautiously in the game - I had, after all, sacrificed a queen. But I didn't think I needed to find something dramatic.

And yet - it was a very easy win indeed. And maybe you have to look for wins as well as staying out of range for long enough for the win to make itself manifest. You don't just have to find, you have to look - and that's something I've never been much good at.

Can this actually be addressed? It's easy, when looking at your games and your mistakes, after you find (or are shown) the wins you missed, to resolve that you will try harder in future and hope to see the win the next time. But you don't. I doubt that simply resolving makes the slightest difference. But how, then, do you do it?

In the past few years, when I've been trying to improve my results, I've generally succeeded in doing so by managing to do less, not more. It is hard to add to your game, when you're in your forties. It is maybe easier, or better, to see what you can take away. To employ the capacity for self-knowledge to reduce one's tendency to blunder, to panic, to throw the game away, to see attacks where there are none. Less is more.

Now maybe it would be possible, even so, to employ that same self-knowledge and declare: look, this situation recurs and recurs, and you know that you miss wins a lot at the climax of a game, when both players are short of time. So would it not be possible to look? To try and adopt a certain mechanism, a discipline? To start out, at least if there are a couple minutes left, by briefly asking myself - has my opponent let me in here? Do I have a win?

I should. But even then, saying is a very long way from doing. And looking is a very long way from seeing. I suspect I shall be missing wins like this, far more often than I see them, for as long as I play chess. Two minutes is not a very long time to see a win: it's not a very long time to check it.

And yet.... it was a very easy win indeed.


Tom Chivers said...

Great illustrations! I looked for two minutes & didn't consider e6 at all.

In this kind of position I think you have to ask yourself 'What would Robin Haldane do?' (Answer: give up a pawn.)

Jonathan B said...

The diagram reminds me of the famous Polugaevsky-Torres game from Moscow, 1981.

ejh said...

Incidentally, Rybka reckons the queen sacrifice is unnecessary - and maybe not even sound - because 21.Qc2 seems to give White a huge advantage (better pieces, safer king, dangerous pawns).

Which is true but I'd probably be a little less unhappy about missing an easy win if I hadn't given up the queen.

Anonymous said...

Well i think it's pretty obvious ;)
Personally i can't see why you wouldn't at least consider it with a big fat pawn sitting on e7. 20 seconds to note that Qg3 leaves no threat and get on with it.


ps. incidentally i can't believe anybody would be put off by the thought that Black might put a piece on d8 for some unknown reason.

ejh said...

I think that probably the thought of moving the pawn didn't occur to me because it was blocking the bishop - and as the game showed, the bishop and queen on the dark squares were very dangerous indeed.

Of course logically this makes incomplete sense because Rd2 abandoned the e-pawn anway, but maybe you seen what I mean. Possibly I should have been alerted to the change in circumstances by the fact that the queen had left the b8-h2 diagonal, asked asked "what does that give me?" - but I was busy asking the other question, "what do I need to stop?"

No doubt also I was too fixated on trying to clean up on the queenside. And I think I was surprised by Qg3 - I think I was expecting the bishop to take on e5.

Or whatever.

Re: the d8 idea - getting rid of White's e7 pawn might very well be worth a piece in some circumstances - this is the sort of position (and variation) where material isn't the primary consideration. It's the sort of theme that one should be aware of, anyway, even if in the specific position it doesn't do (or mean) so much. Besides, given that you've already sacrificed a queen and now you're putting a rook en prise as well, it's surely chess common sense to ask if there any counter-sacrifices present. Chess positions are full of surpirses.

Anonymous said...

I found it under 10 seconds - I expect any 200 player would. But I was also thinking 'I remember this - it's Ivanchuk - Shirov or similar.