A position
Robert-Jan Veldhuizen
Here is another position from my match against MMagnani. It was one of those
games in which neither player gets "lucky" rolls and I found almost every move
in the game a tough choice. Maybe some players can shed some light on this
position and, more generally, what sort of "long-term" goals one should try for
in this sort of position (if there are any...).
+-1--2--3--4--5--6--------7--8--9-10-11-12-+ O: MMagnani - score: 1
| O X O | | O O X |
| O X O | | O O X |
| O | | |
| O | | |
| O | | |
| |BAR| |v 3-point match
| | | |
| X | | X |
| X | | X |
| X O X | | X |
| O O X O X | | X X |
+24-23-22-21-20-19-------18-17-16-15-14-13-+ X: Zorba - score: 0
BAR: O-0 X-0 OFF: O-0 X-0 Cube: 1 turn: Zorba
It's your turn. Please roll or double
Pipcounts: You 143 MMagnani 151
You roll 4 and 3.
Please move 2 pieces.
> 16 19 19 23
So, I moved 16-19 19-23*. One hit, one blot on my 2pt.
How about 17-21 21-24* ? Leaves an extra blot/builder on the 16pt. Is this
better or worse ?
I guess the safe 16-19 17-21 is too cramped ?
And I didn't dare to think about breaking my midpoint here, I hope I was right
there.
Thanks and good luck,
--
<tsb>Zorba
Ron Karr
Robert-Jan Veldhuizen wrote:
>
> Here is another position from my match against MMagnani. It was one of those
> games in which neither player gets "lucky" rolls and I found almost every move
> in the game a tough choice. Maybe some players can shed some light on this
> position and, more generally, what sort of "long-term" goals one should try for
> in this sort of position (if there are any...).
>
> +-1--2--3--4--5--6--------7--8--9-10-11-12-+ O: MMagnani - score: 1
> | O X O | | O O X |
> | O X O | | O O X |
> | O | | |
> | O | | |
> | O | | |
> | |BAR| |v 3-point match
> | | | |
> | X | | X |
> | X | | X |
> | X O X | | X |
> | O O X O X | | X X |
> +24-23-22-21-20-19-------18-17-16-15-14-13-+ X: Zorba - score: 0
>
> BAR: O-0 X-0 OFF: O-0 X-0 Cube: 1 turn: Zorba
> It's your turn. Please roll or double
> Pipcounts: You 143 MMagnani 151
> You roll 4 and 3.
> Please move 2 pieces.
> > 16 19 19 23
This position qualifies as a "mutual holding game", where both sides
have advanced anchors. The leader in the race (pip count) is usually at
an advantage, IF he can manage to clear the anchor without being
attacked (usually by rolling doubles). In this position, you are at an
advantage, since you are ahead in the race. Also, O has 2 additional
men back, which means even if he catches up by rolling big doubles, he
can't escape everyone.
Your anchor is not as advanced as O's, since you're back on his 3 point,
which is a disadvantage for you. However, he won't be able to fill in
his 4 and 5 points easily, so it may not be too hard for you to escape
eventually. Also, you have control of the outfield. He will be trying
to run his spare back men out (keeping the 20 point), and you will have
the chance to hit him.
>
> So, I moved 16-19 19-23*. One hit, one blot on my 2pt.
>
> How about 17-21 21-24* ? Leaves an extra blot/builder on the 16pt. Is this
> better or worse ?
The problem with both hitting plays is that they risk more than they
stand to gain. If hit, you lose 23+ pips in the race, while only
setting O back 1 or 2 pips. If you get lucky and are missed, then cover
the point, you still haven't accomplished that much, since O has an
anchor on your 5 point. Since you are ahead in the race, it's better to
play safe if possible.
>
> I guess the safe 16-19 17-21 is too cramped ?
Actually, it's fairly flexible. Combinations of odd numbers make your 3
point next time; combinations of even numbers make your 2 point. Worst
case, you might have to hit loose next time.
>
> And I didn't dare to think about breaking my midpoint here, I hope I was right
> there.
I would definitely think about breaking the midpoint, since you can make
your 9 point by doing so. The 9 point is very valuable as a blocking
point in this position. And the midpoint isn't so crucial, since O no
longer has his midpoint. However, if you play 12/16, then you're forced
to leave a shot with the 3.
If you are interested in what Jellyfish says (Level 7):
17-21 16-19 15.4
12-16 12-15 14.1
12-16 19-22 14.0
17-24* 10.9
16-23* 10.1
The plays are fairly close, which is usually the case in this type of
position. However, it appears that the safe play is best, followed by
the plays that make your 9 point.
Ron
Chuck Bower
In article <1548.691...@xs4all.nl>,
Robert-Jan Veldhuizen <veld...@xs4all.nl> wrote:
(snip)
>
> +-1--2--3--4--5--6--------7--8--9-10-11-12-+ O: MMagnani - score: 1
> | O X O | | O O X |
> | O X O | | O O X |
> | O | | |
> | O | | |
> | O | | |
> | |BAR| |v 3-point match
> | | | |
> | X | | X |
> | X | | X |
> | X O X | | X |
> | O O X O X | | X X |
> +24-23-22-21-20-19-------18-17-16-15-14-13-+ X: Zorba - score: 0
>
> BAR: O-0 X-0 OFF: O-0 X-0 Cube: 1 turn: Zorba
>It's your turn. Please roll or double
>Pipcounts: You 143 MMagnani 151
>You roll 4 and 3.
I apologize if I'm taking this position too much out of context.
Zorba presented this as a "play" problem, but I have a question:
"what is the cube doing on 1??"
At this score, the trailer has a "loose cube" oportunity,
particularly if gammons are possible (for BOTH sides!). Match LEADER
on the other hand has to be VERY DELICATE with the cube, both taking
and (especially) offering.
I asked Jellyfish ("look, Ma,...") to roll the position out
(BEFORE X rolled the 43) at level-6 (cubeless). The results:
total gammon
X wins 60.6 18.7
O wins 39.4 10.4
(X's cubeless money equity = 0.302; s.d. = 0.040)
These results, if valid, indicate that O BARELY HAS A TAKE! I show
the match equity calculations below:
O passes double: -2,-2 0.50
compare to: ---------------------------------------
O takes and wins W 1.00
O takes and loses simple -2,-1 0.30
O takes and loses gammon L 0
First, prorate the losses by the gammon fraction (18.7 / 60.6 = 0.31)
giving a 0.21 chance if the double is taken. So O RISKS 0.50 - 0.21 = 0.29
to GAIN 1.00 - 0.50 = 0.50. Thus O needs 0.29 / (0.50 + 0.29) = 37%
game winning chances to take. From X's point of view, this drop point is
at 63%. JF says that X will win almost 61% of the time--only 2% away
from O's drop point! Without doing the "window opening" calculation
in the posting (which I get 44% for, BTW), it seems pretty clear that
O should have doubled before rolling the 43.
There are a couple arguments which could come to Zorba's defense.
1) Zorba didn't have Jellyfish available to ask! (At least I hope not.)
2) Maybe Zorba knew his opponent well enough to conclude that either of
the following potential happenings were likely enough that more equity
could be gained by waiting:
a) his opponent would take next time (when s/he should drop) if the
sequence were to go Zorba's way,
b) his opponent would erroneously double and then Zorba could use
his "free redouble" to put the match on the line.
Suffice it to say that if Jellyfish-X were playing Jellyfish-O (both
at level-6) then X should double (and O should take at the 90% confidence
level).
Chuck
bo...@bigbang.astro.indiana.edu
c_ray on FIBS
Robert-Jan Veldhuizen
On 14-dec-96 21:24:25, Chuck Bower wrote:
OK, I'm slow... ;) But, better late than never, I guess, so here's my
"defense":
CB> In article <1548.691...@xs4all.nl>,
CB> Robert-Jan Veldhuizen <veld...@xs4all.nl> wrote:
CB> (snip)
>>
>> +-1--2--3--4--5--6--------7--8--9-10-11-12-+ O: MMagnani - score: 1
>> | O X O | | O O X |
>> | O X O | | O O X |
>> | O | | |
>> | O | | |
>> | O | | |
>> | |BAR| |v 3-point match
>> | | | |
>> | X | | X |
>> | X | | X |
>> | X O X | | X |
>> | O O X O X | | X X |
>> +24-23-22-21-20-19-------18-17-16-15-14-13-+ X: Zorba - score: 0
>>
>> BAR: O-0 X-0 OFF: O-0 X-0 Cube: 1 turn: Zorba
>>It's your turn. Please roll or double
>>Pipcounts: You 143 MMagnani 151
>>You roll 4 and 3.
CB> (snip)
CB> I apologize if I'm taking this position too much out of context.
CB> Zorba presented this as a "play" problem, but I have a question:
CB> "what is the cube doing on 1??"
[snip, JF says:]
CB> total gammon
CB> X wins 60.6 18.7
CB> O wins 39.4 10.4
CB> (X's cubeless money equity = 0.302; s.d. = 0.040)
CB> These results, if valid, indicate that O BARELY HAS A TAKE!
[calculations snipped]
CB> There are a couple arguments which could come to Zorba's defense.
CB> 1) Zorba didn't have Jellyfish available to ask! (At least I hope not.)
I didn't and don't :) But I hope to get it some time (only using it offline of
course) !
CB> 2) Maybe Zorba knew his opponent well enough [reasons to wait snipped]
Nope ;(
Well I'll admit I'm surprised that I have about 60 percent winning chances here
plus pretty high gammon chances. I thought O's 5pt anchor was quite strong, and
my 3pt anchor pretty weak with O's bar point made. Actually, I thought I was
behind in this game...(and in real life, I probably am)
But I would never double in this situation, simply because I don't know how to
play this kind of a holding game. I just don't know what to do with my 3pt
anchor. If I am to win this game, I should leave it some time. But when ? Only
6'es escape one backman and then the other is likely to get hit again and again
with O building his board and perhaps picking up other blots too...
So, to get at this 60 percent winning chances, what is the best time to leave
the 3pt anchor (generally speaking) ?
--
<tsb>Zorba
Brian Sheppard
Robert-Jan Veldhuizen <veld...@xs4all.nl> wrote in article
<2713.6953...@xs4all.nl>...
> On 14-dec-96 21:24:25, Chuck Bower wrote:
> >> +-1--2--3--4--5--6--------7--8--9-10-11-12-+ O: MMagnani - score: 1
> >> | O X O | | O O X |
> >> | O X O | | O O X |
> >> | O | | |
> >> | O | | |
> >> | O | | |
> >> | |BAR| |v 3-point match
> >> | | | |
> >> | X | | X |
> >> | X | | X |
> >> | X O X | | X |
> >> | O O X O X | | X X |
> >> +24-23-22-21-20-19-------18-17-16-15-14-13-+ X: Zorba - score: 0
> >>
> 6'es escape one backman and then the other is likely to get hit again and
again
> with O building his board and perhaps picking up other blots too...
>
> So, to get at this 60 percent winning chances, what is the best time to
leave
> the 3pt anchor (generally speaking) ?
The decision depends on several factors. The first three factors deal
with safety. The others concern opportunity
First, there is the issue of running to *safety*. If the runner is
exposed to shots, then you might want to stay. In this situation a man
that runs from 3 to 9 is exposed only to 6-5, so safety is not an issue.
Second, there is the issue of the safety of the man left behind. In this
case the man is exposed to 9 pointing rolls, plus 3-1, 4-1, and 5-1 for
pick-and-pass shots. This is rather high, but not disastrously high given
O's weak board. O would have a lot of work to do to close out a blot
on the 3-point. (In fact, O would have to bring another man around to
finish the job.)
Third, there is the safety of the rest of your forces. Computer programs
have shown how dangerous it is to be involved in a scramble when you have
even a single extra blot exposed. It is easy for that blot to get
picked up, and then you are in big trouble. So you need to safety the rest
of your men before trying to escape.
On the opportunity side, you have to assess what you gain by moving a man
around. Usually, an assessment of timing, racing, or outfield control
is involved. In this situation, I would say that X is in danger of losing
his outfield control because the midpoint is stripped and the builders
on the 6, 8, and 9 points do not cover O's likely escape squares.
Therefore,
bringing a man around is very useful.
So in this situation I recommend running provided that X does not
leave a third blot. For instance, if X rolls 6-1, then a very reasonable
play is 3-9 16-17. But if X rolls 4-2, then to play 3-9 is very risky; I
would have to convince myself that there were no reasonable alternatives.
Brian
Ron Karr
Robert-Jan Veldhuizen wrote:
> > >>
> >> +-1--2--3--4--5--6--------7--8--9-10-11-12-+ O: MMagnani - score: 1
> >> | O X O | | O O X |
> >> | O X O | | O O X |
> >> | O | | |
> >> | O | | |
> >> | O | | |
> >> | |BAR| |v 3-point match
> >> | | | |
> >> | X | | X |
> >> | X | | X |
> >> | X O X | | X |
> >> | O O X O X | | X X |
> >> +24-23-22-21-20-19-------18-17-16-15-14-13-+ X: Zorba - score: 0
> >>
>
> CB> total gammon
> CB> X wins 60.6 18.7
> CB> O wins 39.4 10.4
>
> CB> (X's cubeless money equity = 0.302; s.d. = 0.040)
>
> CB> These results, if valid, indicate that O BARELY HAS A TAKE!
> plus pretty high gammon chances. I thought O's 5pt anchor was quite strong, and
> my 3pt anchor pretty weak with O's bar point made. Actually, I thought I was
> behind in this game...(and in real life, I probably am)
>
> But I would never double in this situation, simply because I don't know how to
> play this kind of a holding game. I just don't know what to do with my 3pt
> 6'es escape one backman and then the other is likely to get hit again and again
> with O building his board and perhaps picking up other blots too...
>
> So, to get at this 60 percent winning chances, what is the best time to leave
> the 3pt anchor (generally speaking) ?
A good question. When to leave an anchor can be tricky.
The best way to leave an anchor is by rolling doubles. Then you can
move both pieces without leaving shots. In this position, only 66
works. But 11 and 22 allow X to advance the anchor, thus increasing his
escaping chances. This may not seem like a lot of numbers, but over the
course of several rolls it starts to add up. So one question affecting
the decision is: how much time do you have before you have to give up
important assets? If you have a lot of time, you're more likely to sit
back & wait for doubles. Note that in the position shown, O doesn't
have that option. Since he has 2 blots in X's board, in addition to the
anchor, there are no rolls that allow him to escape everyone. He will
be trying to run the loose checkers, then if they get free, he'll worry
about clearing the anchor. That's one of the reasons X has an edge in
this position.
Another good way to leave is to when your opponent is on the bar. This
reduces his attacking chances. In this position X can hit O when he
runs into the outfield. X can also point on O in X's board. If one of
those things happens, particularly if O dances, X will consider breaking
anchor.
Until one of those things happens, X should normally sit tight, as long
as he has other reasonable plays. The cost of leaving an anchor can be
high. In this position, while the checkers on the 3 point anchor appear
trapped, O's front position is inflexible. It's difficult for him to
make new points. If X runs, O suddenly has lots of new options, i.e.
pointing on X.
If it gets to a point where X's timing starts to become clearly worse
than O's, he may have to break the anchor in a less desirable
situation. Such is not the case yet in this position. X has 2 men on
the midpoint to play with, as well as spares on the 6 and 8 points to
try to attack O or make points with.
Ron