Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: My pawn evaluation code (Bugfix)

Author: Dann Corbit

Date: 18:42:26 11/26/01

Go up one level in this thread


On November 26, 2001 at 17:55:23, Michel Langeveld wrote:

>On November 25, 2001 at 15:07:16, Tony Werten wrote:
>
>>On November 25, 2001 at 11:30:51, Michel Langeveld wrote:
>>
>>>Fixed some errors in the passer code...
>>>
>>>   //grouping rule
>>>   //isolated pawns
>>>   //passers
>>>   //double and triple pawns
>>>   bool whiteGroup = false;
>>>   bool blackGroup = false;
>>>   int whiteGroups = 0;
>>>   int blackGroups = 0;
>>>   bool whiteIsolated = false;
>>>   bool blackIsolated = false;
>>>
>>>   for (int i=0; i<8; i++)
>>>   {
>>>      if (evalStructure.whitePawnsPlace[i] > 0)
>>>      {
>>>
>>>         //double pawns
>>>         if (evalStructure.whitePawnsPlace[i] == 2)
>>>         {
>>>              score -= PAWNDOUBLE_PENALTY;
>>>         }
>>>         if (evalStructure.whitePawnsPlace[i] == 3)
>>>         {
>>>            score -= PAWNTRIPLE_PENALTY;
>>>         }
>>>
>>>         //passer?
>>>         if (evalStructure.blackPawnsPlace[i] == 0)
>>>         {
>>>            if (
>>>                 (i != 0 &&
>>>                  evalStructure.blackPawnsPlace[i-1] == 0 &&
>>
>>connected passed pawns don't pass this test
>>
>>>                  evalStructure.whitePawnsPlace[i-1] > 0 &&
>>>                  evalStructure.blackPawnsPlace[i-2] == 0) ||
>>
>>How are passed pawns and not having a pawn 2 files away related ? BTW i-2 can
>>become -1 here, hope your array runs from -1 to 8.
>
>I also evaluate pawns 2 further than the current pawn because I had big problems
>with evaluation the following pawn structure.
>
>[D] 8/pppp4/8/8/8/8/P1PP4/8 w - - 0 1
>
>How many passed pawns does white have....?
>
>p.s.1 yes there's a bounce problem in my code. I programmed it in by purpose to
>gain some extra speed. But fixed it later again.
>p.s.2 I'm not sure what you meant by the connected passed pawns comment. Can you
>give an example?

I am curious as to exactly how you evaluate this position.
I think (for instance) both of these probably lose:
[D]7k/pppp4/8/8/8/8/P1PP4/7K w - -
[D]7k/pppp4/8/8/8/5K2/P1PP4/8 w - - c0 "a bit more questionable...";

And this probably wins:
[D]7k/pppp4/5K2/8/8/8/P1PP4/8 w - -

So it seems to me that tempi dominate a position like this.



This page took 0 seconds to execute

Last modified: Thu, 15 Apr 21 08:11:13 -0700

Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.