Friday, February 20, 2009

Exercises a la Tempo

There are three main "inventions" concerning exercisesI have discovered the past few years. So far the very discoveries itself has costed me so much energy and time that I have never made a serious effort to bring these techniques into practice myself. Due to my latest search for a method that is less taxing for my short term memory I decided to live up to two of these ideas: backward thinking and formulating narratives. But in stead of doing this after I have looked up the solution, I start with it immediately. That means, I skip totally the phase of trial and error.

I start immediately to formulate narratives that describe what is going on in the position. I write them down and only after this I try to find candidate moves that make use of the findings in the narratives. After this I look up the solution in order to see what characteristics I didn't catch in my narratives. The first attempts to solve complex exercises this way look quite promising.

Let me give an example:

diagram 1

White to move.

I formulated the following narratives:
  • The black king has no space
  • Pawnphork on d6
  • Mating square c8 is protected 3 times
  • It seems that the d-file must be opened to make progress
  • The black bisshop is protected 4 times
  • Black bishop is pinned potentially
I missed the following:
  • The rook at c7 is trapped
  • d6 frees e6 and hence blacks minor pieces
  • with the d-line closed it will be difficult to convert the win of the exchange
  • A knight on b5 cannot be taken (=solution, this is better than Nf5+ which leaves the d-file closed and gives up e6)
The next week I will investigate if it works for more calculation intensive problems either. The advantages of this method are:
  • Due to the flexibility of narratives the method can be used for all kinds of problems, not only tactical
  • No spilling of time due to trial and error
  • Less taxing for the STM
  • Learning what characteristics of the position I miss every problem


  1. Nb5 is very nice. I see Nf5+, but of course I'm using a cruder method--check, capture, check, capture!

    Rybka and Fritz say the best move by far is Nxb7, clearing the path for d6+ or if ...Rxb7 then dxc6 Bxc6 Rxd8 with the exchange, a pawn, and a superior position.

  2. It is astonishing that Khmelnitsky didn't check his exam with a computer (this is position #31). Besides that, it means that I missed the potential phork on c6 too. But I'm not alone in that. Grandmasters have done this test too without seeing that. It doesn't alter the idea of the exercise though.