Backgame checker-play problem
tc...@lsa.umich.edu
perhaps someone here can shed light on it. This is a money game.
GNU Backgammon Position ID: ZzNbACCwBQIeHw
+24-23-22-21-20-19------18-17-16-15-14-13-+ O: gnubg
| O X O O X O | | O O O X | 0 points
| O X O O X O | | O O |
| O X X | | |
| X X | | |
| X | | |
| |BAR| |v
| | | |
| | | |
| | | |
| X X | | | Rolled 52
| O X X | | X | 0 points
+-1--2--3--4--5--6-------7--8--9-10-11-12-+ X: tchow (Cube: 2)
--
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
Peter Schneider
<tc...@lsa.umich.edu> wrote
> Here's a problem I faced recently. I still don't understand this
> position;
> perhaps someone here can shed light on it.
I can't shed light but I can grope in the dark, just for promoting the
discussion.
This is a money game.
>
> GNU Backgammon Position ID: ZzNbACCwBQIeHw
> +24-23-22-21-20-19------18-17-16-15-14-13-+ O: gnubg
> | O X O O X O | | O O O X | 0 points
> | O X O O X O | | O O |
> | O X X | | |
> | X X | | |
> | X | | |
> | |BAR| |v
> | | | |
> | | | |
> | | | |
> | X X | | | Rolled 52
> | O X X | | X | 0 points
> +-1--2--3--4--5--6-------7--8--9-10-11-12-+ X: tchow (Cube: 2)
Interesting position. Advanced anchors in a backgame, poor block but no
wasted checkers, opp has 4-pt board. But also opp's position for bearing in
is poor (bar pt gap).
1. Should we hit now or wait until our block is improved? Making the opp's
bar pt also looks nice.
2. If we decide to hit, how do we want to continue, and how do we play the
2 so that it serves our winning plan? (Or perhaps we are content with a
not-losing-gammon plan?)
Ad 1: I once didn't hit in a backgame because I wanted to improve my
position first, and a better player remarked correctly "you may not get a
second shot". (Actually, we are likely to get one here, but we may not
hit!) Thus my inclination is to hit under all but the most suicidal
circumstances. (And coincidentally, that's my inclination anyway, in
backgammon.) Of course it's also much much harder four O to bring two
checkers around than one. And if we contain them, O's winning chances are
much smaller than if we only contain one. I may be wrong but my guess is
that we need to hit.
Considering the non-hitting alternatives:
-- 23/18 20/18: Making the opp's bar pt does not improve our block, the
problem keeping us from hitting. It leaves the shot on our bar pt plus one
additional blot on the 13. I think we want to avoid dancing over an
extended period of time which would give O a welcome opportunity to clear
his problematic 17 and 16 pts while leaving only indirect shots, if any.
-- 20/18/13: More solid, but still leaves the direct shot on our bar.
Doesn't do enough to contain O's back checker.
-- Other options seem inferior. 13/8 7/5, 13/6: No, we want to keep
midfield control. 23/18 + x: Slotting the bar seems too risky with too
little gain.
Ad 2: If we need to avoid dancing, it seems wrong to leave 3 blots (20/15
13/11), a move that would give us the best chances to cover our bar and
would enable us to move subsequent 4s nicely. O may well bear in completely
before we have entered all 3 checkers if s/he manages to hit them all.
Then I see 20/15/13 and 20/15 7/5.
20/15/13 is solid and leaves only the single blot on our bar. Now we have
to decide whether that's an asset or a liability.
We'd loooove to leave it there. It's the point we want to make next. If it
weren't slotted already we'd slot it now, right ;-)? It seems wrong wrong
wrong to move it away. It seems so wrong that it actually may be wrong. But
still.
O hits roughly 1/3 of the times. We cover nearly 1/2 of the times. In other
words only about 1/3 of the times it'll actually serve its purpose. If O
enters but doesn't hit and we don't cover, we'll move it onto our 5pt or
6pt if possible, avoiding the double shot (hitting loosely seems too risky,
giving the lack of builders and poor block). If we can't, we are in
trouble. So we must at least to consider whether we do that right now.
O will escape easily, but it will certainly be hard for O to move two
checkers through a midfield controlled by us. Our advanced backgame has the
upside of letting us escape easily as well so that we can build obstacles
in the midfield. Yes, I have seen people make a point with their two lone
back checkers and rolling them home on big doubles. After all, we are not
favourites ;-). But I'd be afraid to play O's side.
And even if O gets the back checkers home, not all is lost.
I'd opt for 20/15 7/5, with the benefit of a quiz move ;-).
Ah well, gnubg seems to disagree :-).
Best,
Peter aka the juggler
Peter Schneider
a few hours ago I wrote
> I'd opt for 20/15 7/5, with the benefit of a quiz move ;-).
>
> Ah well, gnubg seems to disagree :-).
After running a rollout for a few hours on my poor old single core CPU it
appears that gnubg has a hard time to come to a conclusion at all :-)...
I'm curious.
Best,
Peter
tc...@lsa.umich.edu
Peter Schneider <schneiderp...@gmx.net> wrote:
>I'd opt for 20/15 7/5, with the benefit of a quiz move ;-).
This happens to be the move I chose over the board.
>Ah well, gnubg seems to disagree :-).
For once, I believe a gnubg rollout does not end all argument. I am not
yet convinced that gnubg is playing this position correctly. Here, for
example, is the position on my previous roll.
GNU Backgammon Position ID: szlbgACwAQOOTw
+24-23-22-21-20-19------18-17-16-15-14-13-+ O: gnubg
| O X O O X O | | O O O X | 0 points
| O X O O X O | | O O X |
| X X O | | |
| X | | |
| X | | |
| |BAR| |v
| | | |
| | | |
| | | |
| X X | | | Rolled 56
| X X | X | O | 0 points
+-1--2--3--4--5--6-------7--8--9-10-11-12-+ X: tchow (Cube: 2)
I'm on the bar and have only two legal options here. OTB I chose to hit,
but a gnubg rollout prefers bar/14 by over 0.05. Of course gnubg could
be correct, but I'm not yet convinced.
Walt
I'll bite. I usually try to solve these positions without reading what
others have written, but in this case I didn't know where to start so I
read Peter's analysis first. I'm not sure I understand it (need to read
it again) but here's my take on the position:
X has a lot of checkers back, but he's occupying the wrong points for a
backgame. In particular, the 5 point is too far forward to produce the
well-timed shot that you need to turn the game around. However, the 5
point is well positioned to escape and recirculate checkers and this is
what we want to do here - try to transition into something other than a
backgame, rather than hanging back and waiting for the right moment to
hit in the classic backgame style.
The right moment to hit is now. X has the beginnings of a blockade; by
sending a second checker back and moving some men into the outfield X
may be able to contain them. Once contained, O's outfield will crunch.
Granted, containing the runners is not in the bag yet, but I think it's
the best chance.
So, hit with 20/15* . Where to play the 2? I prefer to spread the
builders around with 13/11 instead of the "obvious" 15/13, but I'm not
sure about that. We want to hit him if he tries to escape, and we
shouldn't be afraid of being hit ourself. 7/5 is too passive, 20/18 is
needlessly reckless.
So, that's my play 20/15* 13/11.
//Walt
Peter Schneider
"Walt" wrote
> I read Peter's analysis first. I'm not sure I understand it (need to
> read it again)
Oh pity ... my mother tongue is German but I try to make myself as clear as
possible. Not sure whether what I wrote is worth a second reading.
I agree with most of what you say except
> we shouldn't be afraid of being hit ourself. 7/5 is too passive
Being so advanced also means that O will quickly pass us if we dance. A 4pt
board poses enough of a threat that we should avoid being hit, let alone
multiple times. 7/5 is ro be considered seriously, as much as I hate to
leave the bar pt.
Best,
Peter
Peter Schneider
<tc...@lsa.umich.edu> wrote
> For once, I believe a gnubg rollout does not end all argument. I am not
> yet convinced that gnubg is playing this position correctly. Here, for
> example, is the position on my previous roll.
>
> GNU Backgammon Position ID: szlbgACwAQOOTw
> +24-23-22-21-20-19------18-17-16-15-14-13-+ O: gnubg
> | O X O O X O | | O O O X | 0 points
> | O X O O X O | | O O X |
> | X X O | | |
> | X | | |
> | X | | |
> | |BAR| |v
> | | | |
> | | | |
> | | | |
> | X X | | | Rolled 56
> | X X | X | O | 0 points
> +-1--2--3--4--5--6-------7--8--9-10-11-12-+ X: tchow (Cube: 2)
>
> I'm on the bar and have only two legal options here. OTB I chose to hit,
> but a gnubg rollout prefers bar/14 by over 0.05. Of course gnubg could
> be correct, but I'm not yet convinced.
Ah, gnubg would not have got us into the mess of your original position!
;-) My suspicion is that it's bad to get hit, especially with more than 1
blot around. Partly because of the pip loss (more bgs) and because O passes
us (partly) while we are on the bar.
gnubg's move makes it less likely to get hit while maintaining good chances
to have another shot a move later.
But I'm not sure either.
Best,
Peter
tc...@lsa.umich.edu
Peter Schneider <schneiderp...@gmx.net> wrote:
>gnubg's move makes it less likely to get hit while maintaining good chances
>to have another shot a move later.
How does it make it less likely to get hit? Either way, O hits if and only
if O rolls a 7.
tc...@lsa.umich.edu
Walt <walt_...@SHOESyahoo.com> wrote:
>X has a lot of checkers back, but he's occupying the wrong points for a
>backgame. In particular, the 5 point is too far forward to produce the
>well-timed shot that you need to turn the game around.
I'm not sure that this is correct. As Walter Trice explains in
"Backgammon Boot Camp," a 2-5 backgame is actually pretty decent for
getting shots. It's quite different from a 1-5 backgame. The 2-point
and the 5-point are close enough to each other that they work together,
with the 2-point blocking some moves and the 5-point well-positioned to
hit if a shot opens up.
Walt
> In article <WKeMm.231120$Jp1....@en-nntp-02.dc1.easynews.com>,
> Walt <walt_...@SHOESyahoo.com> wrote:
>> X has a lot of checkers back, but he's occupying the wrong points for a
>> backgame. In particular, the 5 point is too far forward to produce the
>> well-timed shot that you need to turn the game around.
>
> I'm not sure that this is correct. As Walter Trice explains in
> "Backgammon Boot Camp," a 2-5 backgame is actually pretty decent for
> getting shots. It's quite different from a 1-5 backgame. The 2-point
> and the 5-point are close enough to each other that they work together,
> with the 2-point blocking some moves and the 5-point well-positioned to
> hit if a shot opens up.
I'm not sure either. I remember the part about a 1-5 backgame being the
poorest configuration. I figured that 2-5 was similarly bad, but that
may not follow. Anyway, I suck at backgames. I'm trying to get better,
but I know I suck at them.
But, since I know I suck at them I take ever opportunity to transition
into another theme when I can.
//Walt
//Walt
Walt
> Hi Walt,
>
> "Walt" wrote
>
>> I read Peter's analysis first. I'm not sure I understand it (need to
>> read it again)
>
> Oh pity ... my mother tongue is German but I try to make myself as clear as
> possible. Not sure whether what I wrote is worth a second reading.
It's not your command of English (which is better than most native
speakers) but rather that the subject matter is complex and I only
scanned it.
> I agree with most of what you say except
>
>> we shouldn't be afraid of being hit ourself. 7/5 is too passive
>
> Being so advanced also means that O will quickly pass us if we dance. A 4pt
> board poses enough of a threat that we should avoid being hit, let alone
> multiple times. 7/5 is ro be considered seriously, as much as I hate to
> leave the bar pt.
Well, maybe we should be afraid of being hit. Our timing is sufficient,
so we don't *want* to be hit. But we're far enough behind that passive
play will most likely result in a gammon loss. I'd rather leave the bar
point slotted and take our chances than safety the blot.
But see my previous response detailing my backgame acumen. (c:
//Walt
Philippe Michel
> Here's a problem I faced recently. I still don't understand this position;
> perhaps someone here can shed light on it. This is a money game.
>
> GNU Backgammon Position ID: ZzNbACCwBQIeHw
> +24-23-22-21-20-19------18-17-16-15-14-13-+ O: gnubg
> | O X O O X O | | O O O X | 0 points
> | O X O O X O | | O O |
> | O X X | | |
> | X X | | |
> | X | | |
> | |BAR| |v
> | | | |
> | | | |
> | | | |
> | X X | | | Rolled 52
> | O X X | | X | 0 points
> +-1--2--3--4--5--6-------7--8--9-10-11-12-+ X: tchow (Cube: 2)
I would try to go forward with 20/15* and probably 15/13 rather than
13/11 (not 7/5 for sure).
The 2-5 backgame is not great (but not terrible either) but what stands
out is how ugly O's position is. 6 men still in the outfield and already 3
on the ace point. Totally stripped : if he can't play with his blots, he
could sink very fast.
Right now only 1-5 from the bar hurts O but if X makes his bar point
(hence no 7/5), these root numbers will multiply and X should do well if
he is able to keep his back men moving.
N Merrigan
I chose 20/15*, 7/5. Rollouts suggest 20/15* - 13. I understand the pos,
but the complete destination is slightly unclear.
A 5,2 backgame is considered awkward (timing,skill luck etc). O has almost
relinquished all outfield posesssion. X can still fight from the front -
and should do so with 20/15*. O is stripped. X should create blocking pts
whenever he s/he can.
O's position can only get better if she were able to destroy X's timing.
Thats why I opted for 7/5. While on the bar a tempo is established. So I
agree with 20/15 - 13*. Its a juggling act between stronger boards, timing
and blocking.
At all costs, X must not get hit!! A 5,2 backgammon with sad timing is a
"given".
Good luck
N.Merrigan
<tc...@lsa.umich.edu> wrote in message
news:4b008d01$0$514$b45e...@senator-bedfellow.mit.edu...
Peter Schneider
<tc...@lsa.umich.edu> schrieb im Newsbeitrag
news:4b019f43$0$507$b45e...@senator-bedfellow.mit.edu...
> In article <7mdh91F...@mid.individual.net>,
> Peter Schneider <schneiderp...@gmx.net> wrote:
>>gnubg's move makes it less likely to get hit while maintaining good
>>chances
>>to have another shot a move later.
>
> How does it make it less likely to get hit? Either way, O hits if and
> only
> if O rolls a 7.
Ooops, true.
Peter