Re: [Bug-gnubg] Problems with latest Windows Build

[Joseph Heled]

Definitely. If you don't get something like less than 1% difference in move 
selection (for "interesting" moves) then something is still wrong.
I am sorry about the bug, but I could not solve it until I got a good example 
...

-Joseph

-Joseph

Jim Segrave wrote:
> On Fri 05 Nov 2004 (14:11 +1300), Joseph Heled wrote:
>
> > This is the way to report problems - with an explicit example.
> >
> > Good news found a bug (order of crashed and race nets was reversed) - 
> > making crashed /race prune results complete nonsense.
> >
> > I am submitting a fix now.
>
>
> Shall I rerun my match database through this version to get a new set
> of stats on differences with the pre-pruning gnubg? It takes about 36
> hours to re-run the analyses, but that's not a problem.