[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: Réf. : [Bug-gnu bg] Curren t categorization of doubling mistakes (ne
Re: Réf. : [Bug-gnu bg] Curren t categorization of doubling mistakes (new thread)
Tue, 7 Oct 2003 08:52:18 +0000
On Tue, Oct 07, 2003 at 09:53:09AM +0200, address@hidden wrote
> >In the position referred to, my Doubling Point was 46.3% and my CP 68.4% :
> >have a guess, in 90% of cases, which one I will be closer to when I
> You should have reduobled as soon has your W% was above 46.3% : that's the
> optimal play accoriding to the underlying model.
We should be careful here.
We can calculate double points and take points in both dead cube and
fully live cube models. I'm probably using another MET than you, but
here are my numbers:
Dead cube DP = 47%
Dead cube CP = 63%
Live cube DP := Live cube CP = 69%
This means that you can double from 47%, but it does not mean that you
should! Your opponent can take up to 63%-69%. Which of the two
numbers depend on the cube efficiency. If the cube efficiency is 0 (dead
cube) the opponent can take at 63%, but if the cube efficiency is 1
(fully live cube) he can take at 69%. In reality somewhere inbetween
(gnubg uses a typical value of 2/3 yeilding a take point of approx.
You want to double as close as possible to, say, 66%, but you do not
want to miss your market. I think this is where the early-late ratio
enters: what's the price of doubling early versus missing your market.
I think the newest categories work fine. With the old code (or the
original code) you always have to problem of defining a radius.
Description: PGP signature