[Bug-gnubg] Illogical doubling/resigning with Jacoby rule

[Viehoff Ivan]

Dear gnubg
With Jacoby rule and cube centred, GnuBg sometimes thinks the position
is "too good to double." In the same situation, it likewise will refuse
to accept the resignation of a single game.  I use GnuBG at expert
(0-ply) level.  However even the 3-ply cubeful evaluation also thinks it
is "too good" to double.

With the tables turned, when I have been massively winning with cube
centred (and it advised me I was too good to double), I have played on
to see what happened.  Sufficiently late in the game that the gammon was
certain, it then resigned a single game (rather than a gammon) to me.
So it did realise eventually, but it took a long time.

Best wishes and many thanks for your excellent program

--
Booz Allen Hamilton
7 Savoy Court
The Strand
London WC2R 0EZ

Direct Telephone: +44 20 7393 3739
Fax: +44 20 7393 3305

Email:  address@hidden
Web:  http://www.bah.com