Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: The Code for the Rybka-Mate-Bug

Author: Vasik Rajlich

Date: 05:03:57 12/14/05

Go up one level in this thread


On December 14, 2005 at 06:30:02, Chan Rasjid wrote:

>On December 14, 2005 at 04:44:15, Steve Maughan wrote:
>
>>Vasik,
>>
>>>The problem is when the hash entry ages.
>>
>>This is easy to get around - just store the distance to mate from the *position*
>>(i.e. not the root) in the hash table.
>>
>>In pseudo code:
>>
>>procedure push_hash(position)
>>begin
>>	if winning_mate_value_in_hash then
>>		hash_draft_depth += position's_ply_from_root
>>	else if losing_mate_value_in_hash then
>>		hash_draft_depth -= position's_ply_from_root
>>	store(position)
>>end;
>>
>
>procedure probe_hash(position)
>>begin
>>	if found_in_table(poistion) then
>>	begin
>>		get_hash(position)
>>		if winning_mate_value_in_hash then
>>			hash_draft_depth -= position's_ply_from_root
>>		else if losing_mate_value_in_hash then
>>			hash_draft_depth += position's_ply_from_root
>>	end
>>end;
>>
>>I hope this helps!
>>
>>Regards,
>>
>>Steve
>
>I think this is what Dr. Hyatt meant - store exact-score +- ply?
>It is as in your code.
>
>Could Vasik mean something need to be done in some cases for mate entries
>in TT. I have no clue anything need to be done even if we do not clean TT after
>a game.
>
>In a normal search, Snailchess always return whenever a mate score is found.
>Getting the shortest mate here seems illogical as long as it is a mate-score
>for a position and hashed. Only the general search probe may use them for
>cut-off and they need not be the shortest-dist-to-mate. Comparing 2 mate-scores
>may not be accurate (shortest) but it should not be a problem. When a plus-mate
>is confirmed at root, we call root_mate_search() and this is one of the cleanest
>and easy call that can eliminate all search intricacies with not the slightest
>drawback. I think this should be applicable even for very top programs. Weak
>snailchess never ever fail to complete the game with checkmate with this style.
>
>I don't know the case of Rykba.
>
>Rasjid

It's a question of searching at the root with either

search_root (-shortest_mate, +shortest_mate, ...)

or

search_root (-longest_mate, +longest_mate, ...)

The latter is very slightly more efficient, but the former has better mating
technique which some users apparently like :)

Vas



This page took 0.01 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.