Run or attack?
tc...@lsa.umich.edu
GNU Backgammon Position ID: zK7BASHg8+ABBg
+24-23-22-21-20-19------18-17-16-15-14-13-+ O: gnubg
| O X O O | | O O X | 0 points
| O X O O | | O X |
| O | | X |
| | | X |
| | | |
| |BAR| |v (Cube: 1)
| X | | |
| X | | X |
| X | | X O |
| X | | X O | Rolled 55
| O O 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
lmfback
tc...@lsa.umich.edu says...
> I offer my latest whopper for your dining enjoyment.
>
> GNU Backgammon Position ID: zK7BASHg8+ABBg
> +24-23-22-21-20-19------18-17-16-15-14-13-+ O: gnubg
> | O X O O | | O O X | 0 points
> | O X O O | | O X |
> | O | | X |
> | | | X |
> | | | |
> | |BAR| |v (Cube: 1)
> | X | | |
> | X | | X |
> | X | | X O |
> | X | | X O | Rolled 55
> | O O X | | X O | 0 points
> +-1--2--3--4--5--6-------7--8--9-10-11-12-+ X: tchow
Or both... anyway, I wont bother with GNU positions anymore as they are
always lacking (possibly) relevant information (like a pip count, money
or match scores etc).
Eskimo
--
//------------------------------
//Remove tämä all the way to and including soomee to mail directly.
//Ascended:W,V (genopolywish),P(ill ath), T,K,H,S,B,C,P,W
(naked),Ro,Ra,A,W,almost pacifist A
//In progress:PAIN
Peter Schneider
"lmfback" <peter.b...@absolutelynos-p-a-m.ericsson.com> wrote
> anyway, I wont bother with GNU positions anymore as they are
> always lacking (possibly) relevant information
> (like a pip count, money or match scores etc).
The score is printed; the length is not, and cannot be inferred from the
position id alone (other than the pip count).
That annoyed me as well. But the match length seems to be irrelevant!
Money, 2-pointer, whatever: my mistake (the same as Tim's?) stays a mistake
(I had stayed behind because I was afraid of gnubg's better board).
Best,
Peter aka the juggler
Walt
> I offer my latest whopper for your dining enjoyment.
>
> GNU Backgammon Position ID: zK7BASHg8+ABBg
> +24-23-22-21-20-19------18-17-16-15-14-13-+ O: gnubg
> | O X O O | | O O X | 0 points
> | O X O O | | O X |
> | O | | X |
> | | | X |
> | | | |
> | |BAR| |v (Cube: 1)
> | X | | |
> | X | | X |
> | X | | X O |
> | X | | X O | Rolled 55
> | O O X | | X O | 0 points
> +-1--2--3--4--5--6-------7--8--9-10-11-12-+ X: tchow
>
I'd blitz. 6/1(2)* 8/3(2)*
If he dances, double. Probably double if he fails to anchor.
X has a good anchor and there are three blots to be had. Bold play
would seem to be called for. The passive 21/11 is probably sound and
may win more games, but I'd go for the attack.
//Walt
Walt
>> GNU Backgammon Position ID: zK7BASHg8+ABBg
>> +24-23-22-21-20-19------18-17-16-15-14-13-+ O: gnubg
>> | O X O O | | O O X | 0 points
>> | O X O O | | O X |
>> | O | | X |
>> | | | X |
>> | | | |
>> | |BAR| |v (Cube: 1)
>> | X | | |
>> | X | | X |
>> | X | | X O |
>> | X | | X O | Rolled 55
>> | O O X | | X O | 0 points
>> +-1--2--3--4--5--6-------7--8--9-10-11-12-+ X: tchow
>
> Or both... anyway, I wont bother with GNU positions anymore as they are
> always lacking (possibly) relevant information (like a pip count, money
> or match scores etc).
My understanding is that if no match length is specified it's assumed to
be money play w/ Jacoby rule. If it's match play, gnubg will output the
match score and length when you generate the diagram.
Pip count is a nice piece of info to have, but you can figure it out
from the diagram. Anyway, it's never provided for you in meatspace play.
//Walt
tc...@lsa.umich.edu
Walt <walt_...@SHOESyahoo.com> wrote:
>My understanding is that if no match length is specified it's assumed to
>be money play w/ Jacoby rule. If it's match play, gnubg will output the
>match score and length when you generate the diagram.
Yes. When there's a match score I always give it, of course. When it's
money (with Jacoby), I sometimes forget to state that. A bad habit, I guess,
but most books I've read use the same convention. There's no other default
that makes sense---I mean, I couldn't exactly expect you to assume that the
match score is 2-away/5-away or something random like that, could I?
Paul
>
> Yes. When there's a match score I always give it, of course. When it's
> money (with Jacoby), I sometimes forget to state that. A bad habit, I guess,
> but most books I've read use the same convention. There's no other default
> that makes sense---I mean, I couldn't exactly expect you to assume that the
> match score is 2-away/5-away or something random like that, could I?
> --
Yes, most books use your convention. But it seems that there's at
least one other convention makes sense:
Match to N where N is sufficiently large (to use the maths lingo.)
For all practical purposes, this is the same as 10000 away, 10000
away.
I think it's the case that large N away, large N away, plays the same
way as money non-Jacoby but I'm not sure.
Another default that would make some sense would be money, non-Jacoby.
Paul Epstein
Walt
> On Nov 12, 3:52 pm, tc...@lsa.umich.edu wrote:
>> Yes. When there's a match score I always give it, of course. When it's
>> money (with Jacoby), I sometimes forget to state that. A bad habit, I guess,
>> but most books I've read use the same convention. There's no other default
>> that makes sense---I mean, I couldn't exactly expect you to assume that the
>> match score is 2-away/5-away or something random like that, could I?
>
> Yes, most books use your convention.
When I first started participating in this group I would often ask about
match score and match/money and Jacoby. I was told (sometimes politely,
sometimes not so much) that this is the convention.
//Walt
N Merrigan
threats. After 55, O will have a 4 pip Ad with X on roll. Pretty even.
With a level playing field after 21/11, X can negotiate a position of Prime,
Attack and Race.
I'm outta here.
Nigel
<tc...@lsa.umich.edu> wrote in message
news:4afb879d$0$510$b45e...@senator-bedfellow.mit.edu...
tc...@lsa.umich.edu
This is roughly what I was thinking. Unfortunately, this is dead wrong.
Let's look at the position after 6/1*(2) 8/3(2), dance. X doesn't have a
double, despite having 9 checkers in the zone and despite O's blots. X's
hitting threes are duplicated. More importantly, O's front position is
very strong, with the 5-point and 3-point made. The importance of the
opponent's home-board strength in a blitz is easy to undervalue, but it
is extremely important. In a more standard 55 opening blitz, even if X
rolls something mediocre, O has trouble catching his breath and building
some structure to turn the game around, so the gammon threat may still be
strong. But here, O has already built excellent structure. Unless
everything goes perfectly for X, the turnaround will be swift and painful.
X will be overextended and stuck with an inferior holding game. The gammon
threat just isn't that strong.
Given that X isn't even good enough to double after 6/1*(2) 8/3(2), dance,
we should pause before launching the attack in the first place. Is there
anything better? Yes, of course: 21/11(2). This won't produce an
overwhelming advantage any time soon, but it is safe and solid. It is
the right play by far for money.
Some side notes:
(1) At gammon go (2-away/1-away Crawford), attacking is correct.
(2) The duplication of threes that I mentioned above is important to note.
Suppose we move one of O's checkers from O's 8-point to O's bar-point.
Then X has a double after 6/1*(2) 8/3(2), dance. However, 6/1*(2)
8/3(2) is still clearly worse than 21/11(2), though not by as much
as in the original position.
tc...@lsa.umich.edu
>would seem to be called for. The passive 21/11 is probably sound and
>may win more games, but I'd go for the attack.
Oh, there's one other thing I wanted to mention here. There's a subtle
distinction between "bold" play and "aggressive" play. "Bold" means
being unafraid of leaving blots, while "aggressive" means playing with
the aim of putting your opponent on the bar and keeping him there. X's
anchor means that he can afford to play somewhat boldly, but it doesn't
necessarily indicate that he ought to play aggressively.
tc...@lsa.umich.edu
<tc...@lsa.umich.edu> wrote:
>I offer my latest whopper for your dining enjoyment.
>
> GNU Backgammon Position ID: zK7BASHg8+ABBg
> +24-23-22-21-20-19------18-17-16-15-14-13-+ O: gnubg
> | O X O O | | O O X | 0 points
> | O X O O | | O X |
> | O | | X |
> | | | X |
> | | | |
> | |BAR| |v (Cube: 1)
> | X | | |
> | X | | X |
> | X | | X O |
> | X | | X O | Rolled 55
> | O O X | | X O | 0 points
> +-1--2--3--4--5--6-------7--8--9-10-11-12-+ X: tchow
Just to show you how big an error it is to attack:
1. Rollout 21/11(2) Eq.: +0.194
0.554 0.040 0.001 - 0.446 0.058 0.001 CL +0.090 CF +0.194
[0.001 0.001 0.000 - 0.001 0.001 0.000 CL 0.003 CF 0.005]
2. Rollout 8/3(2) 6/1*(2) Eq.: +0.040 ( -0.153)
0.497 0.147 0.004 - 0.503 0.103 0.005 CL +0.038 CF +0.040
[0.001 0.001 0.000 - 0.001 0.002 0.000 CL 0.003 CF 0.008]
Full cubeful rollout with var.redn.
1296 games, Mersenne Twister dice gen. with seed 869749722 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]
Peter Schneider
> 1. Rollout 21/11(2) Eq.: +0.194
> 0.554 0.040 0.001 - 0.446 0.058 0.001 CL +0.090 CF +0.194
> [0.001 0.001 0.000 - 0.001 0.001 0.000 CL 0.003 CF 0.005]
> 2. Rollout 8/3(2) 6/1*(2) Eq.: +0.040 ( -0.153)
> 0.497 0.147 0.004 - 0.503 0.103 0.005 CL +0.038 CF +0.040
> [0.001 0.001 0.000 - 0.001 0.002 0.000 CL 0.003 CF 0.008]
>
What I find interesting is that the gain in (back)gammons would warrant the
risk of losing more games overall. (Often that's not the case when
considering riskier games.)
What amazes me is that your *opponent's* gammon chances also rise, which
makes it a bad move. That is contrary to my intuition -- keeping that
anchor should reduce your opp's gammon chances! Probably the bold strategy,
once chosen, demands more risky moves subsequently, including hits from the
anchor; or the timing becomes such that we need to leave that anchor even
without hitting.
Quite interesting.
tc...@lsa.umich.edu
Peter Schneider <schneiderp...@gmx.net> wrote:
>What amazes me is that your *opponent's* gammon chances also rise, which
>makes it a bad move. That is contrary to my intuition -- keeping that
>anchor should reduce your opp's gammon chances! Probably the bold strategy,
>once chosen, demands more risky moves subsequently, including hits from the
>anchor; or the timing becomes such that we need to leave that anchor even
>without hitting.
The former is certainly true. Making the 1-point and the 3-point is very
committal. The best way to win is to try to carry out the blitz. That
includes breaking the anchor to hit the opponent's third blot, and hitting
loose when the opponent enters from the bar. This entails a lot of risk
of getting hit back when the opponent already has three good home-board
points made, and several other checkers ready to pounce. (You could, of
course, try to avoid this risk by not hitting loose, but this will come
at a severe cost in winning chances and gammon chances.)
In Kit Woolsey's article about the five-point match, he uses the term
"gammonish position." The first time I read it, I remember asking myself,
"gammonish for whom? White or Black?" I still think that this is a good
question, and that Woolsey sometimes blurs the distinction incorrectly.
However, later I began to realize that in many cases, increasing one's
own gammon chances also increases one's *opponent's* gammon chances.
To increase my gammon chances I will often need to hit loose, and if I
hit loose then I will often get hit back and gammoned myself.
Regarding the anchor, it's true that making and holding an advanced anchor
is an excellent first step to avoiding getting gammoned. However, this is
only the first step. After all, one cannot hold on to an anchor forever;
at some point it has to be broken. Holding on to the anchor longer than
you should means that you'll likely be breaking it at an inopportune time.
So it doesn't follow that holding the anchor longer is better for avoiding
the gammon. In particular, large doubles are often the best possible time
to swing an anchor around. Having all your checkers escaped is usually
even stronger than having an advanced anchor. You're then likely to
transition into a holding game, a priming game, or a race, and these are
not likely to lead to gammon wins or losses.