Re: [gnugo-devel] Level 15

[Heikki Levanto]

On Mon, Feb 03, 2003 at 07:11:44AM -0800, address@hidden wrote:
> Certainly GNU Go is tuned to do well on these regressions at
> level 10. This probably helps keep the program strong but it
> be that the regressions are a very poor measure of actually
> strength.

Agreed! The regressions only test things we have found worthwhile testing,
and many of these we have solved in later versions. They are valuable in
checking that a fix doesn't break something unexpected, but for strength
comparisions actual games would be much to prefer. Even better would be to
force-feed existing games into the two versions, let them both choose a
move, and have a human arbiter tell which move was better. Of course this is
hardly practical.

-H

-- 
Heikki Levanto  LSD - Levanto Software Development   <address@hidden>