Monday, 20 August 2012

De Groot, Thought & Chess

I've been doing a bit of reading over the past few weeks about thought processes and thinking in chess. The 'classic' study of chess players is Adriaan de Groot's university thesis, which went on to become the book Thought and Choice in Chess. Similar thought process work has been done by American Dan Heisman in his various coaching roles. I've also began reading a book by German GM Helmut Pfleger called Chess: The Mechanics of the Mind which covers much of this ground, as well as explaining some of the reasons behind the findings of de Groot & others, from the perspective of two strong chess players (Pfleger's co-author is IM Treppner).
These various experiments came to a number of general conclusions about the thinking of chess players:
1) Good players remember positions in 'chunks' or pieces (eg: Pf2, Pg3, Ph2, Bg2, Rf1, Kg1 as 'kingside castled fianchettoed) rather than an individual pieces on individual squares.
2) Stronger players are generally able to recall 'normal' chess positions faster & more accurately than weaker or less experienced players.
3) In a random position, stronger players have no particular advantage in trying to recall the position than weaker players.
4) Stronger players are able to think more efficiently about a position than weaker players.
5) Stronger players have a general 'memory' of around 10,000 chess positions, or partial positions to draw on, while weaker or less experienced players do not have this sort of memory available to them.

I had been hoping to conduct some related experiments on some players in the invitational chess event last weekend, but only managed to get a small amount of thought-related material done for a variety of reasons. In this case, I gave one of the players 5 seconds to look at a position I had printed on a sheet of paper & he was then asked to reconstruct the position as accurately & quickly as he could. In both cases, mistakes were made, although one position seemed easier than the other (both were from my games about 10-15 moves in, with one being a 'Broken Arrow' line of the modern, while the other was a standard Ruy Lopez position). Although the position was not recognised as being from a Ruy Lopez, many of the pieces were placed on the correct squares initially, with only a few minor mistakes in some piece placement. The other position was much tougher for the player to reconstruct, probably as it was not the sort of position he was used to playing in his own games.

If anyone reading is interested in being part of such experiments, feel free to let me know - I'm more than happy to help volunteers, if a suitable time & place can be found.

1 comment:

  1. Decent doco showing point 2 with Susan Polgar: