Wednesday, June 22, 2005

Intermediate results

I have played a lot of games against my computer at 6 ply deep G/15.
Nearly all the games I loose because of common tactical shots.
With patterns I learned long ago.
So the assimilated patterns have to be transferred to OTB play.
I am trying to find a scanmethod to help me to do that.

To that end I wrote down all the combinations which made me loose against the computer.
From those combinations I distilled a checklist.
To prevent all common combinations which can appear in the first 6 ply, it is necessary to check 207 points.
To check all 207 points at a pace of 7 seconds per check, I needed 25 minutes per move.
Combinations that loose a pawn are not even taken into account!

These figures make a lot of things clear that were incomprehensible first.
It's actually a miracle I blunder only 3 times per game!
To play blunder free you can do 4 moves in 2 hours. . .
The simple X-ray scan as invented by mousetrapper covers only a tiny part of these 207 points to check.

The Holy Grail I'm after now is to simplify this checklist so that the main part of it can be done in about 5 minutes. Untrained that is.
I hope this is doable and workable.
Let's see how far I can come.

**** KNIGHT ALERT *****
Things start to become alarming: Scitcat is kidnapped too!
Is this jealousy? Are we becoming too dangerous?
So other Knights, keep doors and windows closed and be careful!!
Maybe you are next.

3 comments:

  1. 5 minutes per check seems to be very long. When I play a 120 min/40 moves game, I benchmark every 5th move on my scoresheet and force myself to play faster than 15 minutes per mark, this is 3 minutes per move. If you need 5 minutes for safety check alone, then you are in time trouble very soon, if you have no mate in the middlegame.

    ReplyDelete
  2. I'm interested in seeing your checklist. I do think that the value of doing tactical problems is to reduce the need for a checklist. By looking at a postion
    saying what do I recognize here.
    ie: Open file to opponents King (look to capture a piece with a checking fork) bishop and knight a
    space apart (can I do a pawn fork)
    Pawn and Bishop a space apart (can I do the fork trick). This way we could look at a position and say here is what I recognize here and then investigate those tactics. This is what I am striving to do since a checklist for me is to cumbersome and I don't think I can regiment myself to use one consistantly. I'm enjoying yours and mousetrappers posts of late. Good food for thought.




    I think it underscores the value of doing pattern recognition problems to reduce the time. If we could instantly

    ReplyDelete
  3. I wonder why Scitcat took down his blog. I know he stopped his program, but he could have left his blog out there.

    ReplyDelete