Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Question about Crafty output (Bob??)

Author: Les Fernandez

Date: 07:07:26 04/20/05

Go up one level in this thread


On April 20, 2005 at 08:40:11, Robert Hyatt wrote:

>On April 20, 2005 at 01:57:21, Jeremiah Penery wrote:
>
>>On April 19, 2005 at 21:44:53, Les Fernandez wrote:
>>
>>>On April 19, 2005 at 21:20:27, Jeremiah Penery wrote:
>>>
>>>>On April 19, 2005 at 01:43:24, Les Fernandez wrote:
>>>>
>>>>>On April 19, 2005 at 00:41:36, Les Fernandez wrote:
>>>>>
>>>>>>On April 18, 2005 at 22:39:01, Robert Hyatt wrote:
>>>>>>
>>>>>>>On April 18, 2005 at 16:43:58, Les Fernandez wrote:
>>>>>>>
>>>>>>>>Can someone explain to me the differences in the following outputs by Crafty.
>>>>>>>>These are some of the scenarios I came across.
>>>>>>>>
>>>>>>>>ce 1; pv;
>>>>>>>>ce 0; pv;
>>>>>>>>pv;
>>>>>>>>ce 0;
>>>>>>>>ce 1;
>>>>>>>>
>>>>>>>>Thanks,
>>>>>>>>
>>>>>>>>Les
>>>>>>>
>>>>>>>I don't know what that is.  the ce 1 usually means that one side is material
>>>>>>>ahead, but the position is dead drawn.  ce 0 usually means equal material
>>>>>>>draw...
>>>>>>
>>>>>>Hi Bob,
>>>>>>
>>>>>>Maybe I will post one position for each of the above outputs to see if you can
>>>>>>qualify it Tuesday evening.  Thanks for your input.
>>>>>>
>>>>>>Les
>>>>>
>>>>>Decided to get them posted tonight <S>
>>>>>
>>>>>Here they are (actual output)
>>>>>
>>>>>k1K5/8/8/1R6/8/8/8/8 b - - acd 1; acn 1; acs 1; ce 0; pv Ka7;
>>>>>k1K5/6R1/8/8/8/8/8/8 b - - acd 1; acn 1; acs 1; ce 1; pv;
>>>>>k1K5/8/8/8/8/8/8/R7 b - - acd 1; acn 1; acs 1; ce -32767; pv;
>>
>>Ok, I looked more closely at these positions.  The second position has no legal
>>moves, so it's a stalemate with no possible PV.  I don't know why the ce would
>>be 1, however.  In the third position, black is already mated, so has no legal
>>moves.  I'm looking through the source (I assume you're using Crafty compiled
>>with the EPD support) to try and figure out why the first position isn't being
>>analyzed properly (I get the same output as you did on my machine).
>
>
>I responded to this yesterday, but somehow it didn't make it due to some sort of
>network glitch I apparently overlooked.
>
>Remember that crafty scores actual draws at -1, 0 and +1.

Hi Bob,

That explains why I did not see your response.  Listen so that I understand the
nomenclature if it shows a -1 that means "that the position is a dead draw, but
black is ahead in material" whereas 0 means "that the position is a dead draw,
but no one is ahead in material and that +1 means "that the position is a dead
draw, but white is ahead in material.  As we have discussed in the past I am
trying to develope the necessary logic so that I can fill in the missing pieces
in the output of crafty after I run crafty through all the krk positions.  I
remember that one condition where a value of ce 0 is outputed if there is only 1
possible move to be made with the side to move.  I hope after being able to
rectify most of these discussed variations I will reduce the set of epd's that
need my direct attention.  Can I also conclude that anytime a pv is listed as
being blank that the position is in fact a Stalemate and if so why is ce = 1
(does this mean white is ahead of material)?  And can I assume that anytime
other time pv is null that the current position is alrady a mate?  BTW as
Jeremiah talks about in the following thread about eliminatin a small code
section to correct one of these concerns have you received any feedback as to
whether the developer of the code has ever fixed this issue?

Thanks Bob,

Les

>  +1 means that the
>position is a dead draw, but white is ahead in material.  This is an artifact of
>"swindle mode" since given the choice of a draw with material down, or a draw
>with material up, I want to be material up hoping my opponent will make a
>mistake if he doesn't have endgame tables...
>
>so the ce 1 just means that white is up material, in a dead drawn position.



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.