[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [computer-go] Pattern matching - example play
----- Original Message -----
From: "Mark Boon" <tesujisoftware@xxxxxxxxxxxxxxxxx>
> > Isn't half a million games "an "accumulation" of all previous Go knowledge
> > condensed into those games?
> Well, maybe this is the focal point of the discussion. Does half a million
> games contain all Go knwoledge? I think it seems obvious that it does not.
> And also not 2 million, or any other 'practical' number of games. And I even
> question (as explained in my previous post) that it contains all
> 'fundamental' knowledge. Defining 'fundamental' is maybe not so obvious, but
> the pattern in my previous post is an example of what I mean.
Mark, are you saying that a game (or a million games) contains not only the
knowledge that is seen in what was played, but also the one that is revealed in
what (and why) was *not* played? And the latter is not easily (if at all)
extractable from only pattern matching.
Having one million games lets more of that invisible knowledge sip through
(because there are no two identical games, so there are bound to be played many
variations of the same position). Is it enough of it sipping through? We'll see
that as Frank continues his hard work.
<sidenote>
I think that the reticence to accept that only the moves played are enough (for
a pro level player) is because of games (I remember seeing one of Shusaku's)
where apparently everything is quiet and there's no fighting and one side wins
with 3.5 points. But the analysis shows some very fierce fighting that would
have taken place if any one of the players would have played a little more
aggresively.
Is the result of that fight stored in the peaceful variant? I think not, as the
result might or might not give White the victory. One has to analyze to see
that. What we can say is that White didn't have the confidence to read 80 moves
ahead and base his play upon that reading, so he choose peace.
This is however just a feeling we have, and I'd venture to guess that this
feeling is stronger for stronger players, because they are more "indoctrinated"
with the game. It may be right, or it might be wrong. We will be able to tell
only after trying.
</sidenote>
best regards,
Vlad
_______________________________________________
computer-go mailing list
computer-go@xxxxxxxxxxxxxxxxx
http://www.computer-go.org/mailman/listinfo/computer-go/