Re: [Bug-gnubg] Measuring performance levels 2

[Joern Thyssen]

On Mon, Oct 28, 2002 at 08:18:20AM +0100, Morten Wang wrote
> * Douglas Zare
> > I look forward to seeing gnu 2-ply rollout data, to see if gnu 2-ply
> > plays this one better than Snowie 3.
> 
> seems like it doesn't.  my 21600 game rollout isn't done yet (still
> 21% to go) but the results so far are:
> 
> Win    W(g)    Win(bg)    L(g)   Lose(bg)       Cubeless
> 16.18% 0.00%   0.00%      6.46%  1.23%          -0.7533
>  0.12% 0.00%   0.00%      0.06%  0.02%           0.0025
> 
> (21600 games, cubeless settings, no truncation, variance reduction,
> rotate rolls, no jacoby, world class (2ply, huge (8 cand, .160
> tolerance)) running a fresh build grabbed from CVS yesterday evening)

It looks like gnubg is trying some fancy trap-play. 

In my 0-ply and 1-ply rollouts I didn't see any gammon loses as at all,
but 2-ply loses 6.5% gammons and some backgammons as well.

Jørn

-- 
Joern Thyssen, PhD
Vendsysselgade 3, 3., DK-9000 Aalborg, Denmark
+45 9813 2791 (private) / +45 2818 0183 (mobile) / +45 9633 7036 (work)
Note: new mobile number!