Patterns (was: Re: Introduction)

[mueller@xxxxxxxxx (Martin Mueller)]

>David Fotland <fotland@xxxxxxxxxxxxxxxxx>:
>I think the current approach can get to amateur 1 dan level, since there
>really aren't that many patterns a 1 dan knows, but I doubt it will get to
>professional strength.
>
I think it is not only the quantity, but also the quality of patterns. I'm
not a pro, but the patterns I use are much more sophisticated than those of
my program. I think that above the current
raw-board-plus-some-basic-checks, we need to define a hierarchy of more
abstract and more powerful patterns, involving higher level concepts and
encoding the context of a pattern. E.g. since (some reasoning about group
strength here), use this low-level pattern to cut here, force him to run
there, then use another low level pattern to set up a double threat, etc.
All this has been thought and written about at least since Reitman/Wilcox,
but afaik there is still no working implementation.

>Another option is to have the computer learn interesting patterns itself.  I
>think this also needs much more compute power than we have today, but it will
>come.

I believe that with sophisticated patterns you could use better ways of
learning than with primitive patterns which need statistics to weed out the
(usually) good from the (usually) bad.

        Martin