Re: [Bug-gnubg] Use of Two-Sided DB in matches

[Neil Robins]

XG Money play:

Analyzed in 1 ply
Player Winning Chances:    73.53% (G:  0.00% B:  0.00%)
Opponent Winning Chances:  26.47% (G:  0.00% B:  0.00%)

Cubeless Equities: No Double=+0.471, Double=+0.941

Cubeful Equities:
      No Double:   +0.524 (-0.372)
      Double/Take: +0.896
      Double/Drop: +1.000 (+0.104)

Best Cube action: Double / Take

XGID=--A-AAA------------b----b-:0:0:1:00:0:0:3:0:8

Score is X:0 O:0.
 +13-14-15-16-17-18------19-20-21-22-23-24-+
 |                  |   | O              O |
 |                  |   | O              O |
 |                  |   |                  |
 |                  |   |                  |
 |                  |   |                  |
 |                  |BAR|                  |
 |                  |   |                  |
 |                  |   |                  |
 |                  |   |                  |
 |                  |   |                  |
 |                  |   | X  X  X     X    |
 +12-11-10--9--8--7-------6--5--4--3--2--1-+
Pipcount  X: 17  O: 14 X-O: 0-0
Cube  : 1
X on roll, cube action

Analyzed in 2 ply
Player Winning Chances:    73.54% (G:  0.00% B:  0.00%)
Opponent Winning Chances:  26.46% (G:  0.00% B:  0.00%)

Cubeless Equities: No Double=+0.471, Double=+0.941

Cubeful Equities:
      No Double:   +0.548 (-0.304)
      Double/Take: +0.851
      Double/Drop: +1.000 (+0.149)

Best Cube action: Double / Take

----- Original Message -----

From: Michael Petch

To: Michael Petch ; Neil Robins ; address@hidden

Sent: Saturday, September 05, 2009 8:01 AM

Subject: Re: [Bug-gnubg] Use of Two-Sided DB in matches

I’m getting tired. All the lookaheads in this case will come right from the two sided database and have janowski applied to them (I forgot that we aren’t in a contact position for a moment – brain phart). All the evals should come really fast because of this (even on larger plies). On each lookahead the static evals from the database are reapplied with janowski up to the given ply.

One other thing I’d like to know. If you take the same position as a money game in XBG can you also post the results. I’d be curious to see how the values between a XBG’s money eval and Match eval differ.

On 05/09/09 12:38 AM, "Michael Petch" <address@hidden> wrote:

Not necessarily. Interestingly enough on one ply It says double/take but that based on the potential 21 rolls and their resultant equities (each that will also have janowski applied). It also assumes the cube is seen within the plies that are given.

I have a question for you. On simple 0 ply what are XBG’s results? Would you post them. I’d be curious.

On 05/09/09 12:11 AM, "Neil Robins" <address@hidden> wrote:

Surely, therefore, at least 1-ply or higher should see the recube and the then dead cube situation and go to the DB for an exact valid value.

---

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