Blogs
The IBM Player Story

The IBM Player Story

Avatar of danheisman
| 5

About 10 years ago I had a student who was an engineer for IBM in Delaware. His vocation is relevant because I want you to get the impression he was an intelligent, wise adult, very typical of my normal student. But learning chess and understanding what it takes to be a good player sometimes requires insights that are not readily apparent to all, despite their age or wisdom in matters that are not chess-related.

This student wanted to take 3-4 lessons just to get started. Good enough. So we had those lessons and then he said to me something like...

"Thanks Dan! That's just what I wanted. I think I have enough now to go out and play."

So that's what he did. He played regularly online for several months. I checked on him about six months later and his online rating was ~800 (maybe equivalent to about 650 USCF/FIDE)! That's a rating more like you would see an 8-year-old after a few months, not a typical adult. I knew something was wrong, but what was it?

So I looked at his game history and noticed something glaring: whenever he got ahead significant material (at least a pawn, but certainly the exchange or more) he would not try to force trades of pieces; in fact, if his opponent was unwise enough to offer a trade he would almost always turn it down! This flouted one of the most important principles in chess:

When you're ahead pieces, make fair trade of pieces (but not necessarily pawns) [Corollary: When you're behind pieces don't trade pieces, but do try to get rid of all the opponent's pawns].

So I contacted him and told him what I found. He replied,

"Yes, Dan you're right, but it's a lot more fun playing when all the pieces are on the board. So I don't like to trade."

My answer was to the point. "Well, you're right that chess should be fun and anything that makes it more fun is probably good. However, it's not that much fun to not win when you're winning and most games are won by attrition by trading when ahead, as opposed to checkmating the opponent when material is even (or behind). If you don't trade pieces when you are winning, that's very similar to a football coach telling the scorekeeper 'Stop the clock whenever we are ahead.' How many games would that football team win?"

"None!" was his accurate reply.

"Correct, and although the analogy is not perfect - you can win chess games without trading when ahead - it's not a bad analogy in that it makes it much more difficult to win, as we can see by your rating!"

I went on to explain "The idea that you like to keep the pieces on the board is a question of style, and that should come into play when you have a virtually equal choice of making a trade or not. For example, if you can trade queens or not, and either way has about the same evaluation (equal, slight advantage, etc), then you should go with your preference - in this case you should not trade.

However, when you are winning or losing, then you should strongly follow the principle: When you're winning, everything being equal, make fair trades of pieces and that gets you closer to the win."

In my experience, many lower-rated players know that they should trade when ahead just as well as they know that, in general, doubled pawns are probably bad. However, it turns out that very few know that the principle about trading is MUCH more important (and prevalent) than implementing anything you know about doubled pawns! Having vs not having doubled pawns is usually a very tiny factor - perhaps measured around a tenth of a pawn or so (assuming other factors like king safety and isolation are not involved). But not trading pieces when winning or its cousin chess sin, unnecessarily trading pieces when losing are usually quite disastrous, by comparison.

I wrote about this widespread misconception (including the story about the IBM player) and the more general problem of learning how to comparatively value positives and negatives in Not All Bads are Equal. Almost everyone knows that a rook is usually worth more than a bishop or knight because they have "assigned" average values. However, I have found that these same players, when they have to address questions like "Is it worth giving up a pawn if I lose the right to castle?" (of course, it depends...) then often, even if the answer is abundantly clear to a strong player, they have no clue. Hopefully regular readers of my column are at least a little better prepared to face some of these, and many other, learning issues.