[Computer-go] CrazyStone in the 5-dan footsteps of Zen

["Ingo Althöfer"]

Steve uurtamo <uurtamo at gmail.com> wrote:
> keep up the great work -- the bigger the pool of strong machine
> players, the better for everyone.

Right. Having a pool of several bots with similar strength in
the top typically results in faster progress.

And one day (maybe in two years or so) we need someone like Fabien Letouzy
in chess: making the source of his Fruit public in 2005 was the starting
point for an explosion in performance. (By the way: congratulations to
Don Dailey for his strong new chess program "Komodo 4"! He made it public
a few days before Christmas. It jumped immediately to rank 2 in one of the
most serious rating lists, see at
http://www.inwoba.de/
"But" one day after Christmas another new program (Critter 1.4) entered the
scene and surpassed Komodo by 2 Elo points. So, Don's baby is now on a
very strong third rank in chess.)

***********************************
One point I forgot in the original message:
Congratulations to Remi for the fine performance of his bot!

Remi Coulom wrote:
> > The big improvement I got recently came from tuning one parameter that I
> > had forgotten to tune for a few years. 

That is one of the errors one can make. Two others are:
(i) forgetting to introduce new parameters;
(ii) forgetting to kick out some old parameters, when new ones come in.


> > Improving the KGS rank at 5d is very difficult. The current experimental
> > version of Crazy Stone wins 78% against Crazy Stone 2011. That's less than
> > one stone difference on KGS. 

One question out of interest: At which time controls did you run these
selfplay games?

Ingo.




-- 
NEU: FreePhone - 0ct/min Handyspartarif mit Geld-zurück-Garantie!		
Jetzt informieren: http://www.gmx.net/de/go/freephone