RE: [computer-go] Modern brute force search in go

["David Fotland" <fotland@xxxxxxxxxxxxxxx>]

> 
> For go programs it will be very interesting to know what the 
> tactical barrier depth in go is.
> 
> As i'm a layman there perhaps others can put a light on this?
> 

Pick up any go book or magazine with commentary.  You will see lots of
sequences over 20 ply.  I'm not a very strong go player (3 dan amateur),
and I was teaching go to a 2100 chess player.  He was shocked when
I played a 15 ply tactic that was obvious to me.  I've seen commentary and
evidence 
that top pros can work out the entire endgame from 50-100 ply from the end.

Trivial tactics involving blocks with 3 liberties that any beginner knows
are over 
10 ply deep.

When I see a tactical mistake by a pro with commentary it seems it is often
because the
correct move looks strange, and one pro missed it at about 5 or 10 ply deep
in the tree.

I think to beat top pros you need at least 30 ply lookahead in difficult
positions where
the correct moves are not obvious at my level of play.

David


_______________________________________________
computer-go mailing list
computer-go@xxxxxxxxxxxxxxxxx
http://www.computer-go.org/mailman/listinfo/computer-go/