Monday, December 15, 2008

Sac, sac . . . fail!

Chessbase's recent article about the philosophy of chess improvement was loooooong on wordy analysis, short on practical advice. Nonetheless, practical advice it did have - right at the end of the article. In case you missed it, or your mouse broke trying to scroll down that far, here it is:
There is . . . a foolproof method for increasing playing strength . . . [as per] Alexander Kotov’s Think like a Grandmaster [and] it goes as follows:
  • . . . find annotated games from tournament books or magazines and play through them till you to come to the point with the greatest number of variations.
  • Cover up the annotations with a sheet of paper and, without moving the pieces, just like a tournament game, analyse the position from 30 minutes to an hour. If the variations are extremely complex, you might write down your analyses while analysing.
  • When time is out, stop analysing and uncover the annotations in the book or magazine, and compare your notes with the annotator’s.
When starting out, there might be a great discrepancy between your analyses and the annotators’ but with time, one learns to delineate relevant moves and variations as this [method] will exercise and target the mind’s ability to perceive chess positions and produce high quality moves.
Sounds plausible to me. I tried it out with this position:

Perhaps you recognize this position. It's a variation from the game Fischer-Larsen (Portoroz, 1958), which is game number 2 in Fischer's My 60 Memorable Games. In the game black didn't play 17... h7-h5 as indicated, but instead chose 17... Qb5. The longest note of Fischer's analysis concerns 17...h5 instead, which is why I chose this position for my exercise. Perhaps you'll want to try the same before reading on.

Now, it may strike you that this position is rather thematic. Correct. As Fischer wrote, "I'd won dozens of skittles games in analogous positions and had it down to a science: pry open the h-file, sac, sac . . . mate!" Interesting concept of science aside, there is not much to add to that: white is obviously going to play g2-g4, try to line the heavy artillery up on the h-file (and maybe the g-file); the only remotely subtle point is that black's f-pawn is pinned, so g6 is undefended after 17... h5.

Shouldn't be too hard, right? Especially if, like me, you'd played through this game and its notes about a week before on a bus to work? Except I failed to get anywhere, even missing, for instance, that after 18.g4 hxg4 19.h5 Nxh5 the simple 20.Bxg7 wins trivially.

Have a go. Let us know how you got on, and what you think of this method. Oh, and most of Fischer's analysis of this game has also been posted to this chessgames page, if you want to see how it should be done. "Sac, sac . . . mate!"


Will said...

I think this idea has be recycled by a number of authors and teachers. Heisman would call this a stoyko exercise, Rowson mentions something similar and as you say so does Kotov.
Indeed, GM ram is book that is mainly based upon this idea though it includes no annotations to the positions.

Jonathan B said...

Personally I think 30 minutes would be way too long to analyse without moving the pieces or writing things down.

In games I try not to think longer than 10 minutes if at all possible. I just get confused and tend to end up making a mistake.

Anonymous said...

Timman's book Art of Analysis is a great source for positions of this sort. Plus the games are not so famous.

To Jonathan B:
I think your comment is an argument in *favor* of this exercise. Practice makes perfect. You could still set yourself a maximum of 10 minutes over-the-board for any one move, if you want. (I wouldn't.) But having practiced like this before-hand, you would see more, with fewer mistakes, and with less fatigue.