FIBS Board backgammon forum

If we play 18/12, a roll of 6-5 or 6-4 will force us to leave a shot.  I don't see any such rolls with the other play.

That's enough for me.   Keeping the spare on the five might be helpful later, but that's speculative.

6-4 still plays safe (12/2) after 18/12 now, so it's just the 6-5 roll that'd leave a blot. A 2/36 * 11/36 = 22/1296 = 11/648 = ~0.2% parlay.

One of the reasons that some rolls will leave blots from here, is that there aren't spares on the high points, just one checker on the 5pt that can play a one, two or three; the rest is stripped. By playing 18/14 5/3, we make that problem worse: now all high points are stripped and many of our next moves will have forced plays because of that. It is safe for the next turn, but probably quite unsafe for the rest of the bearoff. It likely wins fewer gammons too, for pretty much the same reasons. Leaving the outfield checker on 14 also feels worse than leaving it on 12, in direct range for the bear-in.

So even if 18/12 may leave a shot next turn, it looks clearly better for the rest of the bear-in/off, and against the long-lasting acepoint game this looks most important. You don't want your spares on the lowest points against an acepoint game.

18/12 for me.

Quote from: Zorba on August 04, 2010, 06:21:21 PM

One of the reasons that some rolls will leave blots from here, is that there aren't spares on the high points...

So even if 18/12 may leave a shot next turn, it looks clearly better for the rest of the bear-in/off, and against the long-lasting acepoint game this looks most important. You don't want your spares on the lowest points against an acepoint game.

This is one of those pay-now or pay-later scenarios, and  of the sort that I struggle with.  For instance, the spare on the 5 is a double-edged sword - it saves you from leaving a blot sometimes and forces you to leave one sometimes.

Agree that the chances of us leaving a shot and him hitting it are slim, so the pay-now price is very small.  I'll trust your judgment on the value of the spare on the 5 - if it's good at all, it's got to be worth more than the miniscule pay-now price.

Changing my vote....

gnubg 2-ply evaluation has these two plays dead even. 0-ply rollout favors the pay-now option.

Keeping a spare on the 5 is worth more than the limited extra safety of the pay-later option.




    1. Rollout          18/12                        Eq.:  +0.950
       0.909 0.123 0.003 - 0.091 0.001 0.000 CL  +0.989 CF  +0.950
      [0.001 0.002 0.000 - 0.001 0.000 0.000 CL   0.003 CF   0.004]

    2. Rollout          18/14 5/3                    Eq.:  +0.923 ( -0.027)
       0.901 0.118 0.003 - 0.099 0.001 0.000 CL  +0.965 CF  +0.923
      [0.001 0.002 0.000 - 0.001 0.000 0.000 CL   0.003 CF   0.004]
        Full cubeful rollout with var.redn.
        1296 games, Mersenne Twister dice gen. with seed 896621922 and quasi-random dice
        Play: 0-ply cubeful prune [expert]
        Cube: 0-ply cubeful prune [expert]