Saturday, February 26, 2005

Chess visualisation training

Some time ago a chessfriend and I were talking about how nice it would be if you had a little chessset in your head, where you could work out all variations while playing an OTB-game. So I started a little investigation on visualisation. On a certain moment I landed on the website of Jan Matthies He has a real nice chess visualisation program that you can use online.
I exercised for quite some time and found many interesting things about the human mind.
I will try to explain. I hope you are willing to do a few experiments.

Try the following:
Close your eyes and try to visualize a rook on a chessboard which moves from d1 to a7.

That it easy isn't it? It's easy to spot the file, rank and crosspoint.
Now try to visualize a bishop on c1 heading to a7.
(Now you can open your eyes.:)

That is far more difficult isn't it?
It's nearly imposible to get a good picture of the crosspoint.
It is much easier to see ranks and files than diagonals.

It turns out that the mind is extremely efficient and economical.
From the total geometry of a chessboard are only the most used fractions stored.
Ranks and files we have used all our life with tables, columns and the like.
Diagonals are things we seldom use.

Try another experiment.
Imagine a chessboard and visualize the diagonals from edge to edge.
That is simple. Try now to visualize the diagonal from c1 to h6.
That's very hard to see.
This is because we have often drawn diagonals in squares, but probably never drew other diagonals.
The mind is economical. What is not used is not stored in memory.
If we need it, it can be reconstructed out of the things that are stored in memory.

This is the reason why everything we try to visualize in the mind is so fuzzy and distorted.
From all the physical bodies around us we have only stored the most used geometric properties.
The same holds true for our dreams.

So if you want to visualize the board, you have to store far more geometrical properties of it in the mind. The program of Jan Matthies helps you to do this.
I did his exercises for quite some time.
I started to dream of beautiful chessboards in 3D and technicolor.
From time to time I saw the board and pieces so clear in my mind that I could handle the pieces and follow long variations, just like using a physical chessboard.

Now the most important question: did it help to improve my rating?
The answer is simple: No.

The reason for this is that scientific research (prof. Adriaan de Groot cs) have found that a blindplaying grandmaster doesn't see the board in his mind, but reconstruct the game from the positions he has stored in his memory.
So the grandmaster has stored over 50,000 elementary geometrical patterns of chesspositions in his memory. Patterns that he hasn't stored are for him fuzzy and distorted too. The performance of a grandmaster degrades immensily when he has to work with unknown random chesspositions in stead of normal positions from actual games.
So with the program of Jan Matthies you store the wrong geometrical data in your mind.
Good for boardbuilding, but useless for better playing chess.
Allthough his program is great fun and very instructive.


  1. Interesting and provocative! I never pinpointed that Bishop visualization difficulty before. As I was reading it I came to the conclusion that although it was extremely cumbersome to visualize the e3 square, it seemed slightly easier to target the square by imagining a Black pawn sitting at a7. Unfortunately I still had to figure out what the heck that particular square was.

    As far as usefulness, you may be too highly rated to see any difference. Just a guess. I haven't gone through that exercise, but I attempt to play 1 visualization game once a week where the board is two half moves behind. It does seem to be helping me, and my Pogo rating hit 1700.

  2. That is some great research there. A very interesting explaination on how the mind and memory works.


  3. Oh, also forgot to mention that your post also supports why DLM's program works, and why studying tactics every day is so important to improving in chess. We are starting to build our toolbox of 50,000 positions by starting on the first 1000 in CT Art. That combined with the positions we encounter in our games should give us a good headstart. I hope DG links your post.