How many moves ahead do grandmasters think?

Adriaan de Groot's experiment

The Dutch psychologist and chess master Adriaan de Groot had five grandmasters and five somewhat weaker players think aloud while choosing a move. His book came out in 1946. The result was a surprise: the strongest players didn't look further ahead than the others. They searched just as deep, looked at as many positions, and considered as many moves. But they chose better moves, and found them faster.

Five plies isn't even three moves each. And players rarely looked at more than a hundred positions.

Later studies

Later, researchers did find some differences. Neil Charness found in 1981 that players calculated about half a ply further for every 200 rating points, among players between about 1300 and 2000. When Fernand Gobet repeated de Groot's experiment in Switzerland in 1998, the masters calculated 5.0 plies on average and the weakest players in the study 2.9.

But the longest line in Gobet's study, 23 plies, came from a player two levels below the masters. Being able to calculate far isn't what makes a master.

Patterns in memory

What sets them apart is memory for positions. In 1973, William Chase and Herbert Simon showed a master, a club player, and a beginner a position from a real game for five seconds. The master could set up about 16 pieces correctly, the club player 8, and the beginner 4. With the pieces scattered at random, the difference disappeared.

Later experiments have shown that strong players keep a small edge even with random positions, but a much smaller one. According to computer models of memory, a grandmaster needs to recognize at least 300,000 such patterns. That may explain why the right idea often comes to mind at once.

Patterns can be trained. That's what tactics puzzles are for: the same tricks, over and over, until you see them without looking.