Author: Gordon Rattray
Date: 08:21:04 11/22/01
Go up one level in this thread
On November 22, 2001 at 10:48:01, Victor Zakharov wrote:
>On November 22, 2001 at 09:42:36, Robert Hyatt wrote:
>
>>On November 22, 2001 at 05:27:05, Gordon Rattray wrote:
>>
>>>The Fritz GUI analyses games ("Full Analysis") by starting at the end of the
>>>game and retracting moves. How does this compare to going forwards? Does it
>>>produce better results?
>>>
>>>I think this issue has been discussed before, but my search has failed to find
>>>anything. Please feel free to forward me a past link if appropriate.
>>>
>>>Gordon
>>
>>
>>Here is the idea...
>>
>>If you start at the end of the game, you load the hash table with stuff
>>that will help as you search at earlier moves... with the "idea" that
>>earlier analysis will be more accurate since it will have access to these
>>scores.
>>
>>It doesn't work however.
>
>
>It worked Ok in a lot of situations I have met.
>
>
>>IE pick three points in the game, (a) where a key mistake is made, (b) a
>>position further into the game, and (c) a position near the end where the
>>program can see that it is lost. As you search backward, when you reach
>>(b) the search might well _still_ see that it is lost, because of the persistent
>>hash entries that help. But when you back up past (b) eventually the
>>hash entries get replaced, and you "lose the key scores". You don't find the
>>_real_ place where you screwed up (a), instead the score seems to drop at
>>(b) which is the wrong place.
>
>
>I don't understand why engine must lose "key scores". It depends on the program
>that analyzes position. Tiger14/Gambit-II (at least in CA interface) remembers
>score quite Ok in similar analysis. Don't forget that not only "hash table"
>keeps the history. Most engines remember what moves lead to good/bad evaluation
>and tried to use them first/last in a search. Hiarcs is very good in similar
>learning. (However it loses hash tables values quickly). Tiger is more perfect
>here.
>
>The real problem is when the program thinks that it found a better move with
>better score, but really it is mistaken. So experienced chess players make the
>new move and continue the line. Sometimes a program is able to find that it is
>wrong and finally bring the score to the point (a).
>
>Sometimes it fails. But method really works and people that use it smartly can
>get large advantage.
When you say "smartly", is this only something that a human can do? If not, why
doesn't the program itself "step into" its mainline and then retract? Is it
because of time issues? Or because it doesn't give a benefit often enough?
I'm just a bit puzzled as to why "computer + human" can be better than just
"computer", especially if we're not assuming the human has any great chess
playing ability.
Gordon
>
>>Since neither way finds the actual mistake, I don't like the back-to-front
>>approach because if you do search front to back you will find the "mistake"
>>at a different place, which is nothing more than confusing.
>
>I don'y state that usual front-to-back analysis is bad. Both have advantages and
>lacks. And nothing bad that two approaches can give different positions where
>mistake took place. Just analyze both positions more deeper (and interval
>between them). And you will find the truth.
>
>Victor
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.