Author: Vasik Rajlich
Date: 09:29:29 05/18/05
Go up one level in this thread
On May 17, 2005 at 21:52:26, Robin Smith wrote:
>On May 17, 2005 at 16:57:12, Dieter Buerssner wrote:
>
>>On May 17, 2005 at 14:29:19, Robin Smith wrote:
>>
>>>On May 13, 2005 at 23:49:26, Komputer Korner wrote:
>>>
>>>"Komputer Korner",
>>>
>>>Your post is full of errors. See below for a point by point rebuttal.
>>>
>>>>On May 12, 2005 at 18:31:52, Dann Corbit wrote:
>>>>
>>>>>Bookups backsolving is basically a minimax algorithm + refutations.
>>>>>
>>>>>ChessAssistant does the same thing.
>>>>>
>>>>>It is a very good idea.
>>>>>
>>>>>IMO-YMMV.
>>>>The following explanation will prove that I am the real KK. Compare my answer
>>>>with logical opening theory articles and the 10 kommandments I did 5 years ago.
>>>>The problem is that minimax is only useful to a computer.
>>>
>>>Wrong. Minimax is useful to anyone doing tree searching. Chess masters use
>>>minimax all the time; they just don't call it that. Instead they say things like
>>>"16.Re1 initially looked appealing, but then I saw the killer reply Bxh7+". In
>>>essense this is a verbal description of either the masters thinking or the
>>>minimax algorithm.
>>
>>Robin, I think, you are arguing for alpha-beta search here, not for minimax.
>>Minimax will always look at all moves, even when you already found a refutation
>>("beta cutoff") for you opponents last move. A similar term "brute force"
>>searching is often used for both, pure minimax and for alpha-beta (and as a
>>reader, it is not always clear, which one is meant).
>>
>>A couple of years ago, I read some game analysis (description of his thinking
>>processes during the game) by Robert Hübner in a chess magazine. The way he
>>argued, I got almost convinced, that he used alpha-beta search. I will not be
>>able to give a correct citation to the article (it was in Rochade Europa). From
>>my memory he used phrases like: "there was no need to look for other
>>alternatives in this line, after I had analysed ... [he got a beta cutoff]" and
>>more such things. At that time I already understood alpha-beta search. It
>>seemed, he explained it in non technical words.
>>
>>Regards,
>>Dieter
>
>Hi Dieter,
>
>You are, of course, correct. I would add that 1) as you know minimax and
>alpha-beta produce identical final results, alpha-beta just gets the minimax
>result faster and 2) No-one (that I know of) builds a Bookup database containing
>every move in every position (brute force); obviously bad moves are left out of
>the database. "Komputer Korner" was in effect talking about a pseudo alpha-beta
>as well. And his point that "minimax [and therefore by extension alpha-beta] is
>only useful to a computer" is just plain not correct.
>
>Regards,
>Robin
Actually, there are two separate issues here:
1) how you select the moves to search
2) how you propagate the scores up the tree
I took the original statement by Robin to mean simply that in some cases, a
human will propagate scores up the tree in a minimax manner - which is of course
true.
Note that there are many ways to propagate scores in a non-minimax manner as
well. One example is taking into account second-best moves. Maybe at some point
in the tree, one side can either force a perpetual check, or continue playing in
an unclear position. These two possibilities together should be equivalent to
having some advantage - but such reasoning is outside the scope of minimax,
which can only take into account the score from a single branch.
Similar arguments can be used in the opening. Maybe you like to play 1. e4, but
right at this moment the defense 1. .. c6 is annoying you and you see no
advantage. According to a strict minimax, you should stop playing 1. e4, but
practically speaking you might play it anyway, since not everyone will play 1.
.. c6.
Note also that not all computer chess programs use minimax propagation. Two
alternatives are B* (Berliner, etc) & BPIP (Smith & Baum), and probably there
are many more.
I hope that this is at least partially clear (somehow I suspect not :)) ..
Vas
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.