Sunday, March 08, 2009

Curiosity killed the cat

From elsewhere on the web:

George Koltanowski, probably the best blindfold player ever, gives the following hints to playing blindfold chess:
  1. Cut up a (paper) chess board into four parts ... and memorize the squares.
    (Pay special attention to the color of each squares and diagonals.)

  2. Take an empty board and try to play over very short games of chess.

  3. Set up positions ... and work very hard at VISUALIZING the next 2-3 moves. |
    (DON'T move the pieces!)

  4. Practice your new craft whenever you get the chance! (Three times a week.)

  5. Don't be afraid to lose.

Of course I couldn't restrain my curiosity to have a look at what would be my next challenge in case I manage to obtain perfect board vision. So I simulated perfect board vision by using an empty board and I solved the following exercise from papa Polgars first brick.

diagram 1

White to move. Mate in two.

Solution: [1.Qd4]

While I was working my way through the brick the first time, I developed the good habit to write down the time I needed to solve the problems. This one took me 15 minutes the first time. Now, with an empty board, I needed about 20 minutes.

I had no problem at all to remember where the pieces are. Which means that I store the position in my long term memory. I can prove that with the following story (which I have told before): once I was playing blindfold chess against the computer while the telephone rang. After an hour I resumed playing without the need to lookup the actual board position.

The one thing that I had to repeat continuously though, is which squares were protected by which pieces. I already suspected I was bad at that here. Well, conscious repetition to prevent short term memory wipe out is a clear sign that there is a task that needs transfer into procedural memory. So I will have to devise a training for that.

(BTW have a close look at the first picture until you understand what you see.)


  1. When you started making moves on that problem what happened to the board image, was it a cloud or a square?

  2. I used a real visible empty board without pieces and that basically remained, well, a real visible empty board:) With virtual pieces, though.

    I was simulating perfect board vision, remember.

  3. I realize this stuff isn't an easy read. So please bear with me.

    The reason I started with board vision is solely to get the hang of it. Of how transfer of a conscious task to procedural memory should be trained. I think I know now. So it is time to move on. Since board vision in itself is only interesting if you want to play without a board. Which I don't. In a real game there is a real board, so there is no need for perfectboard vision. (allthough it is fun in itself)

    In this post I identified a conscious task that impedes me during calculation. That are the potential subjects for training that I'm after. I will focus on this one.

  4. Oops, guess I read the post with virtual procedural unconscious subwoofer on... ;-) I just hurried to solve the problem, thought that my "image" was somewhat fuzzy and wondered if exercises you did were helping with that.

  5. Dagnabbit! Two minutes staring at the board and I couldn't spot a mate-in-two. I examined Ng7+ and Qf3+, which go nowhere. But, I could win this position!

  6. Nice puzzle.

    I saw 1. Qd4 and got it.

    To understand the nature of this type of problems is to identify the escape squares for the King.

    What I do is picture the squares around the enemy king that it cannot move to then go on to see what other moves it can make and find ways to stop those moves.

  7. Your problem with "seeing" the covered or attacked squares, may well be something that is tackled up to a degree by MdlM micro drills.

  8. Phaedrus,
    Reasoning tells me lately that it is impossible to transfer something from the conscious part of the brain in an unconscious manner. In fact my experience from the past years forms overwhelming evidence that such transfer can't be done in an automatic way.

    When do I learn the most? When I start to do something new and I am not yet able to do it automaticly. Once I start to do it automaticly, I don't gain anything anymore.

    So the goal is to do something automaticly, but in order to train that, consciousness is needed. Automatic training is useless.

  9. The micro drills described by MdlM seem to be a very conscious exercise to me. The purpose is very much to make you aware of the relationship between squares and pieces.

  10. You underestimate my ability to do a conscious exercise in an automatic way.

  11. At first they are very conscious, the MDLM exercises. Lots of attention needed. Then they become automatic. That's the whole point.

    MDLM, remember, had to do them once a week to stay fresh. I wonder if we underestimate the power of such exercises.

  12. At first glance when I looked at the picture I thought it was two people mirroring each other. As I looked closer I notice that the "person" on the left is a chalk drawing of the person on the right. I looked at all the other pictures on your recent set of posts and noticed the same three dimensional look of these chalk drawings. Very cool stuff.