Re: computer-go: Go Devil

[Compgo123@xxxxxxx]

Don Dailey said:
.................

100 ELO points per stone may not be accurate. Actually rating in chess may 
not correlate with Go that much at all. Of course on the other hand it may 
correlate well. This NEED to be found out. Before we find it out, the merit 
of arguement based on this assumption is really hard to evaluate. 

You think even a top Go player has two stone handicap there is still may ways 
for his opponent (with ability similar to God) to win the game. Before one 
can make this statement, one need to know the degree of the top Go players 
mastering the game. For example, if a top Go player can enumerate all 
possible meaningful ways the game plays out given a two stone handicap and he 
finds no way to win the game. Then he can play God and win with two stone 
handicap. Of course the question then becomes that do top Go players master 
the game that well? Well it seems they do. At least this is what they feel.

To counter above arguement, one may say if they master the Go game that well 
then why the outcome of a game is often so uncertain. The answer of this 
question lies in the design of the game. For example two swimmer got the same 
time in a 50 meter race. This is because the accuracy of the clock. If the 
distance is measured to the accuracy of 1 nm, then the possibility of two 
swimmer get the same time is greatly reduced. In one sense, the difference in 
score is amplified.

Dan Liu