Thursday, June 13, 2013


1. In 70’s, when chess programmer manage to come up with 4 pieces tablebase, it shed a new light on how chess players looks at endgame.

Example is on Queen vs Rook ending in which previously thought as easy win for superior side. Not true, win, yes…easy win, no as what GM Walter Browne discovered when he failed to win in Q vs R against computer’s perfect defense in 70’s.

Makes me wonder, should I spend hours and hours studying Q vs R ending with good chances of drawing the inferior side since GM Walter Browne is having difficulty winning…good chances Malaysian player is having difficulty winning it too.

2. The progress in this field grow steadily…in 80’s we have 5 pieces tablebase, in 90’s we have 6 pieces tablebase and recently we have 7 pieces tablebase.

Makes me wonder, when will programmer could come up with 32 pieces tablebase? Which also mean solving chess…

3. The effect in this ever increasing power of tablebase is both amazing and scary. When 6 pieces tablebase first went out, some position that is thought to be draw…ET informed that actually there is a win in 150 moves!

and when 7 pieces tablebase went out recently, it has been discovered that certain position in which believed to be draw, there is a hidden win in 540 moves!

Makes me wonder, will computer (via say 8 or pieces tablebase), find a win in 1000 moves?

Makes me wonder too, say in tournament with time control of 2 hours and 30 seconds increment, a 540 moves game that started at 9 o’clock in the morning will finished at about 10 o’clock at night!

There goes the next round schedule at 3 pm!

4. Question: There is CD for millions of chess games (Chessbase)…there is CD for chess programs with playing strength of 3000 ELO…How come I never saw any players or shop having this tablebase CD?

Answer:Are you crazy to have all this crammed into CD?
5 pieces TB required about 7GBs of memory,
6 pieces TB required 1 terabyte (1000 GBs of memory)
7 pieces TB required about 50 to 200 Terabyte of memory (or about 200,000 GBs of memory) that's huge!

Makes me wonder, The memory of 8 pieces tablebase once it is completed…is it like 50,000,000,000  GBs ?

how about the memory for 32 pieces tablebase then?


  1. 32 pieces TB will have to wait for quantum computing to become reality ....

  2. Don't think too much Ilham...probably in a few years time a programmer will invent a vacuum machine that can announce mate in 765 moves after Carlsen play 1.e4 and Fischer grave will shake because he once famously scribbled "best by test" on one of his score sheets after playing 1. e4 !