Re: [gnugo-devel] 3.5.3 vs 3.5.4

[Arend Bayer]

On Fri, 20 Feb 2004, Paul Pogonyshev wrote:

> max-d wrote:
> > B- SP Lee pointed  out some 3-5-4's mistakes
> >
> > C- I had run a lot of games which showed that 3-5-3 did better than    3.4
> > vs Handtalk
> >
> > I tried to verify things with an other match (size 13 level 10)
> >
> > = gnugo353 91
> > = gnugo_3_5_4 67
> >
> > almost 57%
> > Seems that something was broken in gnugo_3_5_4
>
> List of patches that can be theoritically responsible for something
> bad between 3.5.3 and 3.5.4:

I think the best way to start analyzing is this is to run
matcher_check on a big number of games and manually look through the
deviations. I may do this sometime but would welcome anyone to beat
me on this.
57% is a big enough winning percentage to give some worries I would
say (assuming colors were alternating).

Arend