what's your play?

[badgolferman]

GNU Backgammon Position ID: 4t42ABDsth0AQA
Match ID : QQnxAAAAAAAE
+13-14-15-16-17-18------19-20-21-22-23-24-+ O: gnubg (Cube: 2)
| O O O | | O O O | 0 points
| O O O | | O O |
| | | O O |
| | | O |
| | | |
v| |BAR| | 7 point match
| | | |
| | | |
| X | | X |
| X X | | X X X X | Rolled 24
| X X | X | X X X X O | 0 points
+12-11-10--9--8--7-------6--5--4--3--2--1-+ X: badgolferman

[Bradley K. Sherman]

badgolferman <REMOVETHISb...@gmail.com> wrote:
> +13-14-15-16-17-18------19-20-21-22-23-24-+ O: gnubg (Cube: 2)
> | O O O | | O O O | 0 points
> | O O O | | O O |
> | | | O O |
> | | | O |
> | | | |
>v| |BAR| | 7 point match
> | | | |
> | | | |
> | X | | X |
> | X X | | X X X X | Rolled 24
> | X X | X | X X X X O | 0 points
> +12-11-10--9--8--7-------6--5--4--3--2--1-+ X: badgolferman

Bar/21,8/6. I want O moving checkers next roll,
not sitting on the bar.

--bks

[Michael Petch]

Well, I'd definitely not hit O on the 23pt. I'd want to get to the edge
of the prime so that getting out only requires a 6. If I hit on the 23
pt I'm likely going to crunch first before escaping. Hitting gives O
chances to dance and maintain his side of the board while I have to get
lucky to roll a 2 followed by a 6.

As for playing the 2 I'd slot the 2pt and put O on the bar. It seems
that should yield more gammons, and still give me a chance to roll my 6
to escape O's side in time.

[Michael Petch]

On 2012-06-01 10:42, Michael Petch wrote:
> It seems
> that should yield more gammon

On second thought this wouldn't be true. Hitting on the 2 point wouldn't
do much for gammons, but getting more wins seems plausible.

[Walt]

Seems to me the choice is whether to concentrate on just winning or
trying to get a gammon by picking up a second checker.

Entering on the 21 seems mandatory, so that the runner can see daylight
and leap with a 6. Hitting loose on the 2 gives me a shot at the blot
on the 23, but at the expense of possibly allowing O to sit on the bar
while my board crunches or being hit and having a second checker stuck
behind the 5 prime while my board crunches.

If I play 8/6, any six forces O to reduce the blockade. I think I like
the odds of either of us shaking a six in the next four rolls better
than me shaking a six in the next two rolls.

I'll play bar/21 8/6 and hang on to the fairly sure single point win.

//Walt

[Walt]

Interesting. I was so sure of the value of stepping up to the edge of
the prime that I didn't even consider hitting on the 23. And if I won't
hit now, I can't see trying to pick up the second checker - if I wanted
it, I'd hit it now.

bar/21 8/6 is even more clear now.

//Walt

[Tim Chow]

On Jun 1, 12:19 pm, "badgolferman" <REMOVETHISbadgolfer...@gmail.com>
wrote:

Bar/21 for sure and not bar/23*. It's far more important to reach the
edge of O's prime than to hit another checker.

The choice between 8/6 and 4/2* is trickier. If we play 8/6, then
with most of her rolls that don't involve a 6, I think O will point on
us. Though we often have strong replies, we could also get counter-
primed and be forced to break our own six-prime. So 8/6 is not
without its own risks.

4/2* risks getting hit back, and then we'll have two checkers needing
to escape. Even if O dances, our 6's are duplicated to escape and
cover, so our blot jeopardy could persist for a few rolls. On the
other hand, the more likely outcome is that O won't hit and we'll have
several rolls to escape our rear checker.

Not an easy choice but I will try 4/2*.

---
Tim Chow

[badgolferman]

Obviously entering on the 21-point is the only rational choice for the
4. The question is what to do with the 2, 4/2* or 8/6. I have
included the rollout for only those two plays.

I played the 8/6, using the same reasoning as Bradley. I want O moving
checkers on the next play and giving me a better chance to escape my
back checker. gnubg prefers hitting on the 2-point for some reason.
To me that only creates a bigger risk of crunching my prime before O is
forced to do the same, not to mention having two of my checkers on the
other side now. I really don't see the advantage to that play.

Playing the 8/6 worked out in that game. I managed to escape the prime
after O was forced to crunch first. I'm pretty sure I would still make
the same play again despite what gnubg says.


1. Rollout bar/21 4/2* MWC: 57.76%
0.763 0.201 0.004 - 0.237 0.044 0.004 CL 58.80% CF 57.76%
[0.004 0.004 0.001 - 0.004 0.002 0.002 CL 0.13% CF 0.17%]
Full cubeful rollout with var.redn.
492 games, Mersenne Twister dice gen. with seed 920907459 and
quasi-random dice
Play: 0-ply cubeful prune [expert]
Cube: 0-ply cubeful prune [expert]

2. Rollout bar/21 8/6 MWC: 56.89% (
-0.88%)
0.725 0.237 0.010 - 0.275 0.057 0.006 CL 58.24% CF 56.89%
[0.003 0.002 0.001 - 0.003 0.001 0.001 CL 0.10% CF 0.12%]
Full cubeful rollout with var.redn.
1296 games, Mersenne Twister dice gen. with seed 920907459 and
quasi-random dice
Play: 0-ply cubeful prune [expert]
Cube: 0-ply cubeful prune [expert]

[Tim Chow]

On Jun 2, 8:24 am, "badgolferman" <REMOVETHISbadgolfer...@gmail.com>
wrote:

> I played the 8/6, using the same reasoning as Bradley.  I want O moving
> checkers on the next play and giving me a better chance to escape my
> back checker.  gnubg prefers hitting on the 2-point for some reason.
> To me that only creates a bigger risk of crunching my prime before O is
> forced to do the same, not to mention having two of my checkers on the
> other side now.  I really don't see the advantage to that play.

I think the main point is that O has more counterplay after 8/6 than
you think she has. Do you know how to use GNU's temperature map
feature? It will show you how GNU thinks O ought to play each of her
next rolls. I'll bet that some of those plays will surprise you. If
you play the position out against GNU several times, I think you'll be
attacked and successfully counter-primed more often than you expect.

---
Tim Chow