EPR, even better than LMR!

Discussion of chess software programming and technical issues.

Moderators: hgm, Dann Corbit, Harvey Williamson

Daniel Shawul
Posts: 4185
Joined: Tue Mar 14, 2006 11:34 am
Location: Ethiopia

Re: EPR, even better than LMR!

Post by Daniel Shawul »

I will try to be clearer.
My restriction is to not reduce heavily below 5 plies. So it doesn't directly go to qsearch just by heavy LMR. Null move can take it to qsearch but that is another issue. Your way of limiting the max reduction in a branch is just one way of doing it. You can also use my way of restriction or do one time non-recursive reduction of 4 plies to get the same effect.The 5 plies margin I have at the end is just to avoid horizon effect. Infact other much stronger engines do not use this criteria at all. What i do in psuedo-code

Code: Select all

if(n_moves > 36) {
  depth -= 4;
  if&#40;depth <= 5&#41; depth = 5; //others skip this check
&#125;
regards,
Daniel
JVMerlino
Posts: 1352
Joined: Wed Mar 08, 2006 10:15 pm
Location: San Francisco, California

Re: EPR, even better than LMR!

Post by JVMerlino »

Daniel Shawul wrote:I will try to be clearer.
My restriction is to not reduce heavily below 5 plies. So it doesn't directly go to qsearch just by heavy LMR. Null move can take it to qsearch but that is another issue. Your way of limiting the max reduction in a branch is just one way of doing it. You can also use my way of restriction or do one time non-recursive reduction of 4 plies to get the same effect.The 5 plies margin I have at the end is just to avoid horizon effect. Infact other much stronger engines do not use this criteria at all. What i do in psuedo-code

Code: Select all

if&#40;n_moves > 36&#41; &#123;
  depth -= 4;
  if&#40;depth <= 5&#41; depth = 5; //others skip this check
&#125;
Got it. Very sorry to make you explain it twice. :)

Thanks very much!
jm