2

Has anyone calculated the average Elo rating of a chess computer as a function of the number of half moves or plys ahead it can look. I realise that the Elo is also a function of the choice of move then taken which may differ from engine to engine.

  • 4
    ELO = Electric Light Orchestra. Elo = surname of Dr. Arpad Elo who created what we now call the Elo system. – kahen Oct 9 '14 at 10:59
  • This is likely (if not guaranteed) to be very different for each major chess engine. – Henry Keiter Oct 9 '14 at 13:52
  • 1
    Yes I mention that in the question. I am really just looking for an average estimate. – Dom Oct 9 '14 at 15:36
3

This is an area that academic can write papers on it. The relationships is complicated and might not be quantified.

Essentially, it's a non-linear relationship where the improvement is most obvious in shallow depth. The benefit diminishes as the engine go deeper and deeper in the search.

Reference: https://chessprogramming.wikispaces.com/Depth

Look for the "Diminishing Returns" section and you'll know more.

| improve this answer | |
  • Thanks for that reference. It provides some useful relative advantages between a n-ply and an n+1-ply engine. I was looking for an absolute Elo but I accept now that that may not be possible. – Dom Oct 10 '14 at 14:26
  • Absolute Elo is not possible because it varies from one engine to another. There is simply no formula. – SmallChess Oct 10 '14 at 14:52
3

There is a paper, which examines this relationship for Houdini 1.5.

http://web.ist.utl.pt/diogo.ferreira/papers/ferreira13impact.pdf

On page 77 you get the relevant table:

depth: elo:

20    2894

19    2828

18    2761

17    2695

16    2629

15    2563

…

8     2099

7     2033

6     1966
| improve this answer | |

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.