Money game: X to play 4-1
tc...@lsa.umich.edu
Match ID : cAkGAEAAMAAA
+13-14-15-16-17-18------19-20-21-22-23-24-+ O: gnubg
| O O O O | | O O O | 4 points
| O O | | O O O |
| | | |
| | | |
| | | |
v| |BAR| | (Cube: 1)
| X | | |
| X | | X |
| X | | X X |
| X | | X X O X | Rolled 41
| O X | X | X X O X | 6 points
+12-11-10--9--8--7-------6--5--4--3--2--1-+ X: tchow
As one of my colleagues likes to say, "Do something that lets you take a
double!"
--
Tim Chow tchow-at-alum-dot-mit-dot-edu
The range of our projectiles---even ... the artillery---however great, will
never exceed four of those miles of which as many thousand separate us from
the center of the earth. ---Galileo, Dialogues Concerning Two New Sciences
tc...@lsa.umich.edu
> GNU Backgammon Position ID: zNZMAQzHPB8AQA
> Match ID : cAkGAEAAMAAA
> +13-14-15-16-17-18------19-20-21-22-23-24-+ O: gnubg
> | O O O O | | O O O | 4 points
> | O O | | O O O |
> | | | |
> | | | |
> | | | |
> v| |BAR| | (Cube: 1)
> | X | | |
> | X | | X |
> | X | | X X |
> | X | | X X O X | Rolled 41
> | O X | X | X X O X | 6 points
> +12-11-10--9--8--7-------6--5--4--3--2--1-+ X: tchow
No responses in two days...I figure I might as well post the rollout.
Over the board, I played bar/24 8/4, figuring that I didn't want a second
checker sent back. However, this is the wrong plan. After bar/24 8/4,
O is guaranteed to (at worst) make the bar point next turn, making it
very tough for X's back checker to escape. And X is likely to have to
expose another blot soon anyway. Bar/21 8/7 may look suicidal, but O's
4's are duplicated and O is not guaranteed to hit. Even if O does hit,
it may well be a loose hit that gives X a return shot. X needs to try
to escape the back checker and now is likely to be the best chance he's
going to get. If X is not hit then the 8/7 slot may let him make a
badly-needed new point, plus X has good chances of escaping immediately.
Bar/21 8/7 is right by a long shot, and gives X a clear take.
I rolled this out as money play with Jacoby, but we see that the equities
of bar/21 6/5 and bar/24 6/2 are less than -1.0, which doesn't make sense.
It means that GNU is overestimating X's chances and having X take the cube
when X should drop. GNU says that bar/24 8/4 is a bare take but given
that GNU is overestimating X's chances, the truth might actually be a drop.
1. Rollout bar/21 8/7 Eq.: -0.808
0.290 0.098 0.003 - 0.710 0.185 0.006 CL -0.510 CF -0.808
[0.002 0.001 0.000 - 0.002 0.003 0.000 CL 0.004 CF 0.013]
2. Rollout bar/24 8/4 Eq.: -0.954 ( -0.146)
0.250 0.076 0.001 - 0.750 0.136 0.005 CL -0.565 CF -0.954
[0.001 0.001 0.000 - 0.001 0.002 0.000 CL 0.004 CF 0.010]
3. Rollout bar/21 6/5 Eq.: -1.013 ( -0.205)
0.268 0.098 0.005 - 0.732 0.233 0.007 CL -0.602 CF -1.013
[0.002 0.001 0.000 - 0.002 0.003 0.000 CL 0.004 CF 0.013]
4. Rollout bar/24 6/2 Eq.: -1.027 ( -0.219)
0.243 0.078 0.001 - 0.757 0.156 0.006 CL -0.597 CF -1.027
[0.001 0.001 0.000 - 0.001 0.002 0.001 CL 0.004 CF 0.010]
Full cubeful rollout with var.redn.
1296 games, Mersenne Twister dice gen. with seed 843427186 and
quasi-random dice
Play: supremo 2-ply cubeful prune [world class]
keep the first 0 0-ply moves and up to 16 more moves within equity 0.32
Skip pruning for 1-ply moves.
Cube: 2-ply cubeful prune [world class]
David C. Ullrich
>In article <4b66185f$0$494$b45e...@senator-bedfellow.mit.edu>, I wrote:
>
>> GNU Backgammon Position ID: zNZMAQzHPB8AQA
>> Match ID : cAkGAEAAMAAA
>> +13-14-15-16-17-18------19-20-21-22-23-24-+ O: gnubg
>> | O O O O | | O O O | 4 points
>> | O O | | O O O |
>> | | | |
>> | | | |
>> | | | |
>> v| |BAR| | (Cube: 1)
>> | X | | |
>> | X | | X |
>> | X | | X X |
>> | X | | X X O X | Rolled 41
>> | O X | X | X X O X | 6 points
>> +12-11-10--9--8--7-------6--5--4--3--2--1-+ X: tchow
>
>
>No responses in two days...I figure I might as well post the rollout.
I for one haven't been reading r.g.b every day.
Saw the position just now. Probably wouldn't have said anything even
if I'd seen it while it was fresh, because what seemed right to me
seems a little crazy.
My spontaneous reaction was bar/21, 6/5. More or less as you suggest
below - looks very bad, but things are not good for X, this may be his
best chance to escape, and getting hit is not the same as losing the
game, quite (because he has at least one roll after O hits before O
can double).
Took me a second to see why 8/7 would be much better than 6/5.
Of course that's clear, I was just being slow, in an early-morning
sort of way.
Walt
> GNU Backgammon Position ID: zNZMAQzHPB8AQA
> Match ID : cAkGAEAAMAAA
> +13-14-15-16-17-18------19-20-21-22-23-24-+ O: gnubg
> | O O O O | | O O O | 4 points
> | O O | | O O O |
> | | | |
> | | | |
> | | | |
> v| |BAR| | (Cube: 1)
> | X | | |
> | X | | X |
> | X | | X X |
> | X | | X X O X | Rolled 41
> | O X | X | X X O X | 6 points
> +12-11-10--9--8--7-------6--5--4--3--2--1-+ X: tchow
(posted without reading your rollout... I've been away, so I didn't see
this problem until now)
The match score is 6-4, but you don't say what the match length is, or
whether it's Crawford. I think this matters. Normally, I'd play not to
be gammoned here, but if the match is to 7 gammons don't matter much,
OTOH, a doubled gammon *does* make a difference. And if the cube is out
of play (Crawford) that may change my checker play too. So without
knowing the match length I don't know what the priorities are. Avoid
gammons? Try to win? Try to avoid being doubled out?
Anyway, X is in bad shape - lots of contact with three checkers burried
and a ridiculous stack on the 8. I don't see many winning scenarios
other than a big roll to escape the runner. However, entering on the 21
to prepan escape leaves the runner utg, and leaves an awkward 1 to
play. Entering on the 24 is somewhat passive, but it allows fewer
gammons and allows almost as many escaping numbers.
So, enter on the 24 and unstack the eight. I think it's better to start
unstacking the biggest stack first, and I don't mind a spare on the 4.
bar/24 8/4.
//Walt
tc...@lsa.umich.edu
Walt <walt_...@SHOESyahoo.com> wrote:
>The match score is 6-4, but you don't say what the match length is, or
>whether it's Crawford.
The subject line says "Money game."
By the way, here's one reason I used to avoid posting the Match ID along
with the position ID. If I didn't post the Match ID then I felt free to
zero out the match score (for money games) with a text editor, to avoid
the kind of confusion that Walt exhibits here. But if I post the Match ID,
then I don't want the Match ID to contradict the human-readable match
score. And I don't like to edit the original GNU position because then
I risk accidentally overwriting my saved match file.
My only suggestion to r.g.b. readers is as follows: Assume that I have
not messed up. So far, I have never messed up, in the sense of posting
a position where the match score matters but failing to mention the
match score. If you find yourself thinking that I've omitted some
crucial piece of information about the match score, then (1) look harder,
and (2) if no match score is given, assume money-play with Jacoby.
Simon Woodhead
> The subject line says "Money game."
>
> By the way, here's one reason I used to avoid posting the Match ID along
> with the position ID. If I didn't post the Match ID then I felt free to
> zero out the match score (for money games) with a text editor, to avoid
> the kind of confusion that Walt exhibits here. But if I post the Match ID,
> then I don't want the Match ID to contradict the human-readable match
> score. And I don't like to edit the original GNU position because then
> I risk accidentally overwriting my saved match file.
Why not simply start a money session with gnubg ? Then the
matchid would be correct and you wouldn't have to "zero out"
the score.
> My only suggestion to r.g.b. readers is as follows: Assume that I have
> not messed up. So far, I have never messed up, in the sense of posting
> a position where the match score matters but failing to mention the
> match score. If you find yourself thinking that I've omitted some
> crucial piece of information about the match score, then (1) look harder,
> and (2) if no match score is given, assume money-play with Jacoby.
You always leave out the pip count which is exceptionally annoying.
Walt
> In article <p9Can.176173$er1....@en-nntp-07.dc1.easynews.com>,
> Walt <walt_...@SHOESyahoo.com> wrote:
>> The match score is 6-4, but you don't say what the match length is, or
>> whether it's Crawford.
>
> The subject line says "Money game."
Oh. You mean I have to read the directions?
(c:
//Walt
tc...@lsa.umich.edu
Simon Woodhead <simon.w...@uniq.com.au> wrote:
>Why not simply start a money session with gnubg ? Then the
>matchid would be correct and you wouldn't have to "zero out"
>the score.
These positions are typically taken from training sessions with GNU
where I play 5 or 10 consecutive money games at a time and save the
session. To do what you suggest I would need to restart the money
session after each game, which I find inconvenient.
>You always leave out the pip count which is exceptionally annoying.
I do this intentionally because all my backgammon experience with human
opponents is over the board where there is no pip count available.
Anyway, if you don't like the way I do things, why not post some positions
of your own for the rest of us to think about? It's been a while since
I've seen anybody but myself initiate any substantive backgammon content
on this newsgroup.
David C. Ullrich
wrote:
And don't forget to make sure the computer's plugged in.
That's very important here.
>(c:
>
>//Walt
David C. Ullrich
>In article <037b2d06$0$1303$c3e...@news.astraweb.com>,
>Simon Woodhead <simon.w...@uniq.com.au> wrote:
>>Why not simply start a money session with gnubg ? Then the
>>matchid would be correct and you wouldn't have to "zero out"
>>the score.
>
>These positions are typically taken from training sessions with GNU
>where I play 5 or 10 consecutive money games at a time and save the
>session. To do what you suggest I would need to restart the money
>session after each game, which I find inconvenient.
>
>>You always leave out the pip count which is exceptionally annoying.
>
>I do this intentionally because all my backgammon experience with human
>opponents is over the board where there is no pip count available.
I agree on that - seems to me people would be better off if programs
didn't display the pip count.
Walt
I'm pretty sure it is, but how can I tell for sure?
//Walt
tc...@lsa.umich.edu
Walt <walt_...@SHOESyahoo.com> wrote:
>I'm pretty sure it is, but how can I tell for sure?
For the low, low price of just a few months' salary, the Dogbert Computer
Consulting Company will gladly email you explicit instructions.
Grunty
> In article <037b2d06$0$1303$c3e8...@news.astraweb.com>,
> Simon Woodhead <simon.woodh...@uniq.com.au> whined:
>
> >You always leave out the pip count which is exceptionally annoying.
>
> of your own for the rest of us to think about?
Hey Tim, calm down, and please understand Simon.
When he runs out of fingers he has to use toes. Now that's
"exceptionally annoying."
David C. Ullrich
wrote:
From the headers it appears you're running Windows. So I don't
know how to tell, sorry. Linux comes with a nice little utility that
you can execute to check this.
>
>//Walt