Cognition in Chess by J. Corey Butler (PhD and Class B chess player)

This is an abbreviated version of an article I intended for publication in the magazine, Chess Life several years ago. Sadly, my article was rejected, but perhaps it can find new life here on the internet instead of stagnating in my file cabinet.

What makes a strong chess player? One hypothesis is that people who become chess masters have exceptional intelligence or memory, but there is little evidence to support this view. Most researchers have found minimal correlations between measures of IQ and official chess ratings. On the other hand, many chess masters appear to have a phenomenal memory. They can recall games played years ago, move by move, and when shown an unfamiliar chess position for only a few seconds, they can reproduce it virtually without error on a new board and set. The catch, however, is that this feat is only possible when they are given positions taken from actual games. When the position is random, the master does only about as well as the amateur. General intelligence and memory by themselves do not appear to distinguish great chess players from ordinary ones.

Another hypothesis is that chess masters have amazing powers of visualization, and can think ahead many moves to see what the board will look like later in the game. In fact, chess players do not really “visualize” future positions in the sense of a detailed mental image, such having a picture of the board in one's head, but they are able to calculate long series of moves. Computer chess programs like Deep Blue and now Fritz owe their strength to an ever increasing ability to calculate and analyze future positions. Can we explain the skill of human grandmasters by their ability to calculate ahead? Probably to some extent, but it often surprises less skilled players how little masters actually calculate in many positions. In a quote variously attributed to Capablanca, Reti and other masters, the question was asked, “how many moves do you think ahead?” The surprising answer was “only one move, the right move.” In a classic study by de Groot at the University of Amsterdam, the findings showed that on average, masters calculated no deeper than weaker players, and often examined fewer variations. Nevertheless, they almost always selected superior moves.

So how do we explain the considerable ability of chess masters? Their memory is excellent, but only for meaningful chess positions. They have the ability to do long calculations in their head (especially in the end game), but they usually don’t do this any more than ordinary players. The view of many psychologists (e.g. Chase & Simon, 1973) is that the greatest difference in chess skill between masters and amateurs is in the realm of pattern recognition. Just as anyone with a driver’s license can glance at a stop sign and effortlessly realize what it means, highly skilled chess players only need to take a brief look at a chess position to assess it accurately. They can instantly see positional themes like pawn chains, weak squares, and open lines, as well as tactical possibilities like Knight forks. Patterns of pieces such as weakened King positions and Rook batteries are recognized and evaluated as the player decides what the best move is.

Chess masters can play under rapid time controls, sometimes as fast as 60 seconds for an entire game, with only minimal deterioration in the quality of their play. A study by Herbert Simon at Carnegie Mellon University examined nine simultaneous exhibitions given by world champion Garry Kasparov during the late 1980s and early 1990s against teams composed of masters and grandmasters. A simultaneous exhibition provides a good test of the pattern recognition hypothesis, because although there is some time for calculation, it is really a “look and move” situation. During this time, Kasparov’s rating was about 2750, and his performance ratings in the exhibitions reached a median level of 2646. This indicates that even under severe time constraints, Kasparov was able to perform within about 100 points of his normal, tournament playing strength. This is good evidence that rapid pattern recognition was the key to his success.

Lack of pattern recognition is why new players are often the victim of back rank mates, and other embarrassing losses. These defeats are embarrassing because the players know that the loss is obvious and don't need any explanation afterward. They simply fail to notice the danger when the King is sitting behind a wall of pawns while there are Rooks about. In contrast, an experienced player automatically sees the threat in this position and easily avoids it. Calculation is not necessary. The chess master has learned this and a multitude of other kinds of positions and can recognize them instantly. Stored themes and patterns may also be the basis of the chess master's superior calculation. Psychologists talk about the beneficial effects of "chunking" in working memory, and there is no doubt that familiar patterns of information are processed much more efficiently.

If rapid, accurate pattern recognition is the essential cognitive advantage that separates masters from patzers, can we use this knowledge to improve our skill at chess? Yes, but not without some difficulty. It's clear that chess patterns are stored in long-term memory as the result of countless hours playing and studying games, and there is simply no substitute for this experience. We can expect a typical learning curve here, with gradual progress over time, and an eventual plateau at some level of skill. One straightforward piece of advice that could be useful is to focus on books of combinations. I know that I was able to add about 200 points to my rating just by studying a few positions every day over the course of a few months. After a while the various tactical patterns became burned into my brain and I began to see them spontaneously during my games.

I hope this has been a useful introduction to the study of chess cognition. Now go study some mating patterns or endgame positions!