A simple computer chess programming puzzle

[rjgibert]

so I'll change my program and provide my new numbers as soon as this is done

New results, now differentiating knight and bishop (still without allowing promotions, that one will follow later):

Code: Select all

total:             236196
even:                2266
within one pawn:     4256

Here "within one pawn" does not include "even".

Sven

I'm going to make an argument to show that your numbers 2266 must be too low.

The total value of Whites material is 1*8 + 3*2 + 3*2 + 5*2 + 1*9 = 39 excluding the King of course. Naturally it is 39 for black too. This means there are 39 + 39 +1 = 79 possible differences of material.

The average number of positions corresponding to each is 236196/79 > 2266 even though the number of even positions is expected to be the largest category. Not that I have split "within one pawn" into 2, +1 pawn and - 1 pawn into 2 categories when I say that the even category is expected to be the largest category. Therefore, your 2266 number is too low.

[Sven]

I'm sorry, I only gave the values for 100-325-325-500-975. Here are my values for 1-3-3-5-9:

Code: Select all

total:             236196
even:                8690
within one pawn:    17266

Does that sound correct for you now?
Sven

[rjgibert]

I'm sorry, I only gave the values for 100-325-325-500-975. Here are my values for 1-3-3-5-9:

Code: Select all

total:             236196
even:                8690
within one pawn:    17266

Does that sound correct for you now?
Sven

Your result is identical to my result now