Re: [gnugo-devel] Recent games

[Daniel Bump]

> > Could you give some examples of the phenomenon you are describing?
> > I think they may be frequent in games where GNU Go gives a 9 stone
> > handicap. But generally I think GNU Go is better now at giving
> > big handicaps than 3.0.0.
> 
> Maybe the phenomenon is not as frequent as I had imagined. nngs1:2 is
> a bit of an example, also the 3-3 invasions that your patch took care of
> by an increased negative shape value.
> The best example I have seen is nngs:900. It is from a 9stone game, but
> the mistake is extremely urgent.

I don't agree about nngs1:2. The tenuki move is only valued around
20 and the problem is that GNU doesn't understand that the hane
threatens to save the W stones at the top.

About nngs:900, you are absolutely right, this is very urgent
and we need to take the time to fix this. The move to defend
the W dragon at the top, though valued at 23 isn't even in
the top ten.

I looked at trevorb:180 and I agree, this is a problem we
need to fix. I hope we can do it without losing the
improvement we got from the experimental influence.

Dan