Every roll is a cube decision
tc...@lsa.umich.edu
stopped to think about.
By posting it here, I'm giving you the advantage of knowing that it's
something that one should stop and think about. So, after thinking
about it, what is your answer? Ignore the score; this is a money game.
GNU Backgammon Position ID: drwNADCYt5EBMA
Match ID : cAkAAGAAAAAA
+24-23-22-21-20-19------18-17-16-15-14-13-+ O: gnubg
| X O O O | | O O X | 6 points
| X O O O | | O O X |
| O O | | |
| O | | |
| | | |
| |BAR| |v (Cube: 1)
| | | |
| X | | |
| X | | |
| O X X | | X X | On roll
| O X X | | X X X | 0 points
+-1--2--3--4--5--6-------7--8--9-10-11-12-+ X: tchow
--
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
Bradley K. Sherman
<tc...@lsa.umich.edu> wrote:
> ...
>about it, what is your answer? Ignore the score; this is a money game.
Does "money game" imply Jacoby rule in r.g.b. jargon?
--bks
tc...@lsa.umich.edu
That's the usual convention, but if you think that Jacoby makes a
difference in this position, then feel free to explain.
Bradley K. Sherman
<tc...@lsa.umich.edu> wrote:
>In article <hm9g3f$dgn$1...@reader2.panix.com>,
>Bradley K. Sherman <b...@panix.com> wrote:
>>In article <4b883e35$0$506$b45e...@senator-bedfellow.mit.edu>,
>> <tc...@lsa.umich.edu> wrote:
>>> ...
>>>about it, what is your answer? Ignore the score; this is a money game.
>>> ...
>>
>>Does "money game" imply Jacoby rule in r.g.b. jargon?
>
>That's the usual convention, but if you think that Jacoby makes a
>difference in this position, then feel free to explain.
No, I haven't even looked at the position yet. Just trying
to get my bearings here.
--bks
Bradley K. Sherman
> Match ID : cAkAAGAAAAAA
> +24-23-22-21-20-19------18-17-16-15-14-13-+ O: gnubg
> | X O O O | | O O X | 6 points
> | X O O O | | O O X |
> | O O | | |
> | O | | |
> | | | |
> | |BAR| |v (Cube: 1)
> | | | |
> | X | | |
> | X | | |
> | O X X | | X X | On roll
> | O X X | | X X X | 0 points
> +-1--2--3--4--5--6-------7--8--9-10-11-12-+ X: tchow
X's timing is excellent, O's is abysmal. The only
problem with doubling is that O is ahead in the race.
In a lot of cases X will lose his market for a double
if he doesn't double now. The third checker on 22 is
the swing vote for me. Double.
--bks
Peter Schneider
"Bradley K. Sherman" <b...@panix.com> wrote
> X's timing is excellent, O's is abysmal. The only
> problem with doubling is that O is ahead in the race.
> In a lot of cases X will lose his market for a double
> if he doesn't double now. The third checker on 22 is
> the swing vote for me. Double.
And since it's a close cube it's a big take by implication? Which would be
my choice. I'm not even sure about the cube -- X may never close their 5pt,
doesn't have much of a board, so O has plenty of time and opportunity to
escape. No need to contain X's back checker either since O is ahead. The
main danger for 0 is to leave more shots by rolling awkwardly.
But X will often catch an escaped checker, O has no connection so that
their back checkers are all on their own.
Does Jacoby play a roll? Possibly, but for both -- X too has some gammon
losses if things don't go smoothly. But true, in the scenario where O
leaves more shots (very likely) gammons become a greater threat for 0,
which speaks for the cube.
Best,
Peter aka the juggler
Bradley K. Sherman
Jacoby, in my antiquated backgammon model, implies more aggressive
doubling, and O is likely to end up with 3+ checkers on the 1 pt
behind a prime. It seems like this must be a take, and that's
why I didn't mention it, but now I am suspicious that the way the
question was asked implies this a drop!?
--bks
Stick
First let me try to drag another person to my forums since it seems
Bradley is new here and may be looking for a free online backgammon
resource.
http://www.bgonline.org/forums/
Clear cube, clear take. These are the sort of positions that are fun
to adjust and see what makes it a stronger cube and how you can
eventually push it into pass territory. You can do the inverse too of
course.
Stick
On Feb 27, 8:57 am, b...@panix.com (Bradley K. Sherman) wrote:
> In article <7us16cFpq...@mid.individual.net>,
Bradley K. Sherman
Stick <check...@yahoo.com> wrote:
>
>First let me try to drag another person to my forums since it seems
>Bradley is new here and may be looking for a free online backgammon
>resource.
>
>http://www.bgonline.org/forums/
Sorry, I'm a USENET booster. Tell them to come here.
--bks
Walt
> Below is a cube decision that, not so long ago, I would not even have
> stopped to think about.
>
> By posting it here, I'm giving you the advantage of knowing that it's
> something that one should stop and think about. So, after thinking
> about it, what is your answer? Ignore the score; this is a money game.
> GNU Backgammon Position ID: drwNADCYt5EBMA
> Match ID : cAkAAGAAAAAA
> +24-23-22-21-20-19------18-17-16-15-14-13-+ O: gnubg
> | X O O O | | O O X | 6 points
> | X O O O | | O O X |
> | O O | | |
> | O | | |
> | | | |
> | |BAR| |v (Cube: 1)
> | | | |
> | X | | |
> | X | | |
> | O X X | | X X | On roll
> | O X X | | X X X | 0 points
> +-1--2--3--4--5--6-------7--8--9-10-11-12-+ X: tchow
X is behind in the race. Both sides have their runners stuck on the ace
point. O has timing problems - unless he can activate his runners, his
board will crunch. And any split will put a blot utg.
Now that you've made me stop and think about this position, I see that
most sequences lead to a clear double for X - with most rolls, X can
easily strengthen his blockade and then O's position starts to
deteriorate. So a double seems sound - as O I'd have to think
carefully about whether to take, and Kit has a rule about that.. But I
only arrived at this through QF - over the board I would have just rolled.
//Walt
Grunty
> In article <54c6a46a-ee4b-4df2-854c-283496311...@b7g2000yqd.googlegroups.com>,
>
> Stick <checkmug...@yahoo.com> wrote:
>
> >First let me try to drag another person to my forums since it seems
> >Bradley is new here and may be looking for a free online backgammon
> >resource.
>
> >http://www.bgonline.org/forums/
>
> Sorry, I'm a USENET booster. Tell them to come here.
>
> --bks
YEAH!
(just my 2 cents)
Walt
>>> Does "money game" imply Jacoby rule in r.g.b. jargon?
>> That's the usual convention, but if you think that Jacoby makes a
>> difference in this position, then feel free to explain.
>
> No, I haven't even looked at the position yet. Just trying
> to get my bearings here.
I get confused about this myself sometime. But the convention is:
If no match length is given, then it's money play.
If it's money play, Jacoby is in effect.
If it's match play, Crawford is in effect.
HTH. Welcome to the newsgroup. Don't let the bad behavior of one or two
posters get you down.
//Walt
Bradley K. Sherman
> Match ID : cAkAAGAAAAAA
> +24-23-22-21-20-19------18-17-16-15-14-13-+ O: gnubg
> | X O O O | | O O X | 6 points
> | X O O O | | O O X |
> | O O | | |
> | O | | |
> | | | |
> | |BAR| |v (Cube: 1)
> | | | |
> | X | | |
> | X | | |
> | O X X | | X X | On roll
> | O X X | | X X X | 0 points
> +-1--2--3--4--5--6-------7--8--9-10-11-12-+ X: tchow
What's the answer?
--bks
Peter Schneider
<tc...@lsa.umich.edu> wrote
> Below is a cube decision that, not so long ago, I would not even have
> stopped to think about.
>
> By posting it here, I'm giving you the advantage of knowing that it's
> something that one should stop and think about. So, after thinking
> about it, what is your answer? Ignore the score; this is a money game.
>
>
> GNU Backgammon Position ID: drwNADCYt5EBMA
> Match ID : cAkAAGAAAAAA
> +24-23-22-21-20-19------18-17-16-15-14-13-+ O: gnubg
> | X O O O | | O O X | 6 points
> | X O O O | | O O X |
> | O O | | |
> | O | | |
> | | | |
> | |BAR| |v (Cube: 1)
> | | | |
> | X | | |
> | X | | |
> | O X X | | X X | On roll
> | O X X | | X X X | 0 points
> +-1--2--3--4--5--6-------7--8--9-10-11-12-+ X: tchow
At a second glance a beautiful double/take. O is very inflexible and likely
to crash, X has plenty of timing and good chances to catch an escaped
checker from O. X pointing on the own 5pt is such a classic position -- O
will think twice before splitting to X's 5pt (and may then refrain) -- but
that would be one of the best chances to anchor or run. Also 66s don't work
for O. Still a take because of the race and X's still imperfect block
(implying indirect shots, anchors or escapes for O) and X's back checkers
on the ace.
Best,
Peter aka the jugular
Walt
> In article <4b883e35$0$506$b45e...@senator-bedfellow.mit.edu>,
> <tc...@lsa.umich.edu> wrote:
>> GNU Backgammon Position ID: drwNADCYt5EBMA
>> Match ID : cAkAAGAAAAAA
>
> What's the answer?
Tim usually posts a rollout after a couple of days. If you're in a
hurry, you can always put the position and match id's into gnubg and do
your own.
//Walt
tc...@lsa.umich.edu
Bradley K. Sherman <b...@panix.com> wrote:
>What's the answer?
Yes, sorry for the delay in posting the rollout. I've been on a business
trip all week, and for logistical reasons I am not able to post the rollout
until I get back. I'll get to it this coming weekend.
tc...@lsa.umich.edu
> GNU Backgammon Position ID: drwNADCYt5EBMA
> Match ID : cAkAAGAAAAAA
> +24-23-22-21-20-19------18-17-16-15-14-13-+ O: gnubg
> | X O O O | | O O X | 6 points
> | X O O O | | O O X |
> | O O | | |
> | O | | |
> | | | |
> | |BAR| |v (Cube: 1)
> | | | |
> | X | | |
> | X | | |
> | O X X | | X X | On roll
> | O X X | | X X X | 0 points
> +-1--2--3--4--5--6-------7--8--9-10-11-12-+ X: tchow
As some have already mentioned, this position is a big double and a big take.
X's partially trapped back checkers and race deficit might make some people
think that X is far from being able to double. But in fact, X's race deficit
is a plus here; it means that X has a timing advantage. X has control of the
outfield, whereas O's rear checkers are cut off from the rest of O's army,
which is too far advanced and in danger of crashing. X's budding prime is
superior to O's. If X makes another blocking point or escapes his back
checkers then O will be in serious trouble. Now is the time to double,
while O can still take. If O escapes the back checkers with impunity while
X rolls poorly then O can certainly turn the game around.
This is a good example of a position where visualizing what is likely to
happen in the next few rolls is very important. A static assessment may
miss the crucial features of the position.
Rollout details:
Centered 1-cube:
0.688 0.209 0.009 - 0.312 0.102 0.009 CL +0.483 CF +0.695
[0.002 0.002 0.000 - 0.002 0.001 0.001 CL 0.004 CF 0.011]
Player gnubg owns 2-cube:
0.698 0.208 0.011 - 0.302 0.094 0.008 CL +1.027 CF +0.825
[0.002 0.002 0.001 - 0.002 0.001 0.001 CL 0.010 CF 0.017]
Full cubeful rollout with var.redn.
1296 games, Mersenne Twister dice gen. with seed 849725326 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]
tc...@lsa.umich.edu
>Rollout details:
>Centered 1-cube:
> 0.688 0.209 0.009 - 0.312 0.102 0.009 CL +0.483 CF +0.695
> [0.002 0.002 0.000 - 0.002 0.001 0.001 CL 0.004 CF 0.011]
>Player gnubg owns 2-cube:
> 0.698 0.208 0.011 - 0.302 0.094 0.008 CL +1.027 CF +0.825
> [0.002 0.002 0.001 - 0.002 0.001 0.001 CL 0.010 CF 0.017]
>Full cubeful rollout with var.redn.
>1296 games, Mersenne Twister dice gen. with seed 849725326 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]
By the way, since the Jacoby rule came up in this thread, I also rolled
the position out without Jacoby. See below. The results strike me as a
little fishy, since logically, the equity after "double/take" should be
the same regardless of whether the Jacoby rule is in effect, and it's
not, by a statistically significant amount. Possibly I did something
wrong, or possibly there's a bug in GNU, or possibly GNU is not handling
the cube correctly, or possibly if I rolled it out longer then the anomaly
would disappear, despite the seeming statistical significance. Anyway, I
doubt that the presence/absence of the Jacoby rule will change the verdict
of "double/take" for this position.
Rollout details:
Centered 1-cube:
0.685 0.206 0.008 - 0.315 0.103 0.008 CL +0.473 CF +0.681
[0.002 0.002 0.001 - 0.002 0.001 0.000 CL 0.004 CF 0.009]
Player gnubg owns 2-cube:
0.694 0.202 0.009 - 0.306 0.096 0.008 CL +0.991 CF +0.766
[0.002 0.003 0.001 - 0.002 0.001 0.000 CL 0.010 CF 0.014]
Full cubeful rollout with var.redn.
1296 games, Mersenne Twister dice gen. with seed 850057193 and quasi-random