Coming under the gun
tc...@lsa.umich.edu
my 24-point. I desperately need to get a back checker moving, but what if
doing so comes under the gun? Do I step up anyway, or wait and hope for
another opportunity? This is a decision I have difficulty with. Here is
a position where I made the wrong choice. It is a money game.
GNU Backgammon Position ID: cG/CAQaw58EBMA
+24-23-22-21-20-19------18-17-16-15-14-13-+ O: gnubg
| X O O | | O O X | 0 points
| X O O | | O X |
| O O | | X |
| O | | |
| | | |
| |BAR| |v (Cube: 1)
| | | |
| X | | X |
| X | | X O |
| O X X | | X O | Rolled 24
| O X X | | X O | 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
Walt
> Suppose my opponent has made an advanced anchor while I am still stuck on
> my 24-point. I desperately need to get a back checker moving, but what if
> doing so comes under the gun? Do I step up anyway, or wait and hope for
> another opportunity? This is a decision I have difficulty with. Here is
> a position where I made the wrong choice. It is a money game.
>
> GNU Backgammon Position ID: cG/CAQaw58EBMA
> +24-23-22-21-20-19------18-17-16-15-14-13-+ O: gnubg
> | X O O | | O O X | 0 points
> | X O O | | O X |
> | O O | | X |
> | O | | |
> | | | |
> | |BAR| |v (Cube: 1)
> | | | |
> | X | | X |
> | X | | X O |
> | O X X | | X O | Rolled 24
> | O X X | | X O | 0 points
> +-1--2--3--4--5--6-------7--8--9-10-11-12-+ X: tchow
I usually think of utg as having three builders poised to hit the blot.
In this case there are four builders, so it's utg on steroids. With O
holding the anchor, he'll hit loose if he can, and 24/22 can be hit by
any 2,3,4, or 6. Only 5-5 avoids a hit, and that's a joker anyway.
(some other doubles may have better moves than a hit.)
I'd hang back. Something like 6/2 13/11, not that I'd be at all
thrilled to make that play.
//Walt
Piranha
I play 13/7
If he hits, no big deal, the 3rd man gives help to build an anchor
myself, while he has to break his anchor.
Not sure if Snowie would agree, I just prefer open battle over holding
position until you run out of safe moves.
tc...@lsa.umich.edu
> GNU Backgammon Position ID: cG/CAQaw58EBMA
> +24-23-22-21-20-19------18-17-16-15-14-13-+ O: gnubg
> | X O O | | O O X | 0 points
> | X O O | | O X |
> | O O | | X |
> | O | | |
> | | | |
> | |BAR| |v (Cube: 1)
> | | | |
> | X | | X |
> | X | | X O |
> | O X X | | X O | Rolled 24
> | O X X | | X O | 0 points
> +-1--2--3--4--5--6-------7--8--9-10-11-12-+ X: tchow
The rollout results strongly favor splitting 24/22. I'm going to do a
couple of related rollouts to confirm this result. In particular, after
X plays 24/22 6/2, does O have a cube? GNU 2-ply thinks not, but the
decision is close, and if it's making a mistake here then it might be
skewing the rollout results. However, since the equity differences are
coming out so large, I doubt that the final answer will be any different.
By experimenting with a couple of different variations, I think I have
some idea why coming under the gun is correct. Part of it is that X's
position is rather far advanced. Move a checker from the 6-point back
to the midpoint, and 13/7 becomes a much more attractive alternative.
In the given position, 13/7 strips the midpoint and so it is much harder
to cover the blot. Similarly if instead of the 5-point, X held the
bar-point or the 10-point, then 13/7 would be a contender.
Another factor is that hanging back on the 24-point indefinitely is
just not a viable gameplan, and splitting next roll isn't likely to
be any more attractive. In the actual game, I played 8/2, and then
O rolled 4-1, played 13/9 6/5. Then I rolled 21. I compounded my
previous error by playing 8/5; in fact, splitting 24/22 6/5 is mandatory
even though it looks even more unpleasant than on the previous turn.
Given that splitting will be necessary, X is better off splitting now
before the situation deteriorates further (which it will, because timing
considerations mean that X's front position will get worse while O's
will get better).
1. Rollout 24/22 6/2 Eq.: -0.610
0.366 0.085 0.004 - 0.634 0.238 0.008 CL -0.425 CF -0.610
[0.001 0.001 0.000 - 0.001 0.002 0.001 CL 0.003 CF 0.009]
2. Rollout 24/22 13/9 Eq.: -0.706 ( -0.096)
0.343 0.080 0.003 - 0.657 0.248 0.010 CL -0.490 CF -0.706
[0.002 0.001 0.000 - 0.002 0.002 0.001 CL 0.004 CF 0.013]
3. Rollout 8/2 Eq.: -0.755 ( -0.144)
0.329 0.076 0.002 - 0.671 0.220 0.008 CL -0.492 CF -0.755
[0.001 0.001 0.000 - 0.001 0.002 0.001 CL 0.003 CF 0.009]
4. Rollout 13/7 Eq.: -0.763 ( -0.153)
0.325 0.076 0.003 - 0.675 0.217 0.010 CL -0.498 CF -0.763
[0.002 0.001 0.000 - 0.002 0.002 0.000 CL 0.004 CF 0.012]
Full cubeful rollout with var.redn.
1296 games, Mersenne Twister dice gen. with seed 869399626 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]
Walt
>> I'd hang back. Something like 6/2 13/11, not that I'd be at all
>> thrilled to make that play.
>>
>> //Walt
>
> I play 13/7
> If he hits, no big deal, the 3rd man gives help to build an anchor
> myself, while he has to break his anchor.
> Not sure if Snowie would agree, I just prefer open battle over holding
> position until you run out of safe moves.
I didn't even consider 13/7. Now that I'm looking at it, it's a better
move than 6/2 13/11. We don't want to make the 11 or the 2 at this
point so why slot them? Slotting the bar point looks good.
//Walt
tc...@lsa.umich.edu
>> GNU Backgammon Position ID: cG/CAQaw58EBMA
>> +24-23-22-21-20-19------18-17-16-15-14-13-+ O: gnubg
>> | X O O | | O O X | 0 points
>> | X O O | | O X |
>> | O O | | X |
>> | O | | |
>> | | | |
>> | |BAR| |v (Cube: 1)
>> | | | |
>> | X | | X |
>> | X | | X O |
>> | O X X | | X O | Rolled 24
>> | O X X | | X O | 0 points
>> +-1--2--3--4--5--6-------7--8--9-10-11-12-+ X: tchow
>
>The rollout results strongly favor splitting 24/22. I'm going to do a
>couple of related rollouts to confirm this result. In particular, after
>X plays 24/22 6/2, does O have a cube? GNU 2-ply thinks not, but the
>decision is close, and if it's making a mistake here then it might be
>skewing the rollout results. However, since the equity differences are
>coming out so large, I doubt that the final answer will be any different.
Rollouts say that after 24/22 6/2, O doesn't quite have a double, whereas
after any of the alternatives (24/22 13/9 or 8/2 or 13/7), O does have a
marginal double. This is a little different from what GNU 2-ply says, so
there is indeed a minor issue with the rollout here, but it doesn't look
like the overall verdict will be changed.
tc...@lsa.umich.edu
muratk <mu...@compuplus.net> wrote:
>I am very surprised that a cocksucking mathshitter with
>gambling millionaire friends like you can stoop down to
>sweat such stupid little details...?? :(
I don't play backgammon for the money, only for the intellectual challenge.
Looks like you have a lot to learn about the psychology of mathshitters!
N Merrigan
What interests me more about this position is your previous roll! Could you
have split?
Anyway, over the board, I would play 24/22, 13/9. Clearly your oppon is
calling all the shots here, and you need a foothold in the game. Splitting
now before your oppon's position gets even stronger is a must. From my
experience I would have split a roll before if I was able to. Anticipating
the flow of any game allows you to envisage what could happen to you if you
fail to take action. If you see the threat in real time, then, it's too
late, you've missed the bus.
Having missed your bus, you have to bite the bullet and hope it doesn't go
off in your mouth.
When and when not to come under the gun is not as tricky as some might lead
you to believe. If your on the backfoot like in this position, then you are
defending. Your oppon wants to keep it that way. The more you are
defending, the more of an offencive position your oppon is building against
you.
If your oppon's position is inflexible, then s/he will want to hit you at
the earliest opportunity. Hitting with slotting and splitting unstacks an
inflexible position and returns it to a more fluid state. In your position
however, you have to strip down to your birthday suit and say to the world
"here I am". Bare yourself with 24/22, 13/9 offers the quickest route to
disaster or glory.
Seriously, you need to unravel the potential variations of doing XYZ and who
will have the upperhand when the dust settles. Here, your simply after a
midway anchor which will give you a platform of offence.
Questions you need to be asking are: how dangerous is a split here? Can I
recover if hit? Do I have some game if doubled. Do the pros outweigh the
cons?
Consider this opening and reply. From the starting position your oppon
rolls 62 and plays 24/17, 13/11. You roll 54 and play 6/1*, 24/20. Coming
under the gun with tempo removes your oppon's fingers from the trigger.
Your aim is to prevent your oppon from getting an advanced anchor before you
do. Here, you are able to achieve both objectives. It's a matter of seeing
what the picture looks like after XYZ.
N.Merrigan
<tc...@lsa.umich.edu> wrote in message
news:4af22700$0$512$b45e...@senator-bedfellow.mit.edu...
tc...@lsa.umich.edu
N Merrigan <acu...@blueyonder.co.uk> wrote:
>What interests me more about this position is your previous roll!
>Could you have split?
No. The situation I was faced with on my previous roll was this:
GNU Backgammon Position ID: sG/gAQbQc+QBMA
Match ID : cAkXAAAAAAAA
+24-23-22-21-20-19------18-17-16-15-14-13-+ O: gnubg
| X O O | | O X | 0 points
| X O O | | O X |
| O | | X |
| O | | X |
| O | | |
| |BAR| |v (Cube: 1)
| | | |
| X | | O |
| X | | X O |
| O X | | X O | Rolled 65
| O X X | | X X O | 0 points
+-1--2--3--4--5--6-------7--8--9-10-11-12-+ X: tchow
I knew that with my opponent's advanced anchor, I really wanted to get my
back checkers moving. I certainly would have split here if my roll had
allowed it. I did consider the splitting play in the position I posted
originally, but thought that I had missed my chance. I can see now that
I should have split anyway.
muratk
> In article <8dc95297-38c2-4bce-9a88-8b86aa5a4...@y28g2000prd.googlegroups.com>,
> muratk <mu...@compuplus.net> wrote:
>> I am very surprised that a cocksucking mathshitter with
>> gambling millionaire friends like you can stoop down to
>> sweat such stupid little details...?? :(
> I don't play backgammon for the money, only for the intellectual challenge.
Ah, so, is that why you and your millionaire friends jack off with a
cheating piece of shit backgammon bot...??
> Looks like you have a lot to learn about the psychology of mathshitters!
The truth is self evident...! ;)
MK