Re: [Bug-gnubg] 1-chequer hypergammon

[Ned Cross]

Hmm, I've never played 1-checker hyper, but it seems to me gammons would
always be equal to wins for both sides by definition.  So the cube would
seem to be an attempt to activate gammons that would be lost over the next
exchange in favarable scenarios for X.  I am not sure about the beaver.

That either side has a nonzero probability of winning backgammon is simply
wrong, since contact is broken and both sides have cleared the enemy home
board. Removing the BG probabilities should slightly alter the equities.


----- Original Message -----
From: "Ian Shaw" <address@hidden>
To: "GnuBg Bug (E-mail)" <address@hidden>
Sent: Thursday, April 10, 2003 9:10 AM
Subject: [Bug-gnubg] 1-chequer hypergammon


How can Double/Beaver be correct? Is it expecting Double - Beaver - Drop 2
points? If so, then Double - Take ought to be better.
And how can the gammons be the same as the wins for both sides?

    GNU Backgammon  Position ID: AAIAAAAIAAAAAA
                    Match ID   : MBkAAIAAKAAA
    +-1--2--3--4--5--6-------7--8--9-10-11-12-+  O: Leader
    |                  |   |                  |  8 points
    |                  |   |                  |
    |                  |   |                  |
    |                  |   |                  |
    |                  |   |                  |
    |                  |BAR|                  |^
    |                  |   |                  |
    |                  |   |                  |
    |                  |   |                  |
    |                  |   |                  |  Cube offered at 2
    |                  |   | O        X       |  5 points
    +24-23-22-21-20-19------18-17-16-15-14-13-+  X: Trailer


Cube analysis
2-ply cubeless equity  +0.025
  0.498 0.498 0.038 - 0.502 0.502 0.006
Cubeful equities:
1. Double, take         -0.035
2. Double, pass         +1.000  ( +1.035)
3. No double            -0.049  ( -0.014)
Proper cube action: Double, beaver

--Ian


_______________________________________________
Bug-gnubg mailing list
address@hidden
http://mail.gnu.org/mailman/listinfo/bug-gnubg