My question is: in actual play on the Internet (1000 exp. points at
GamesGrid, 800 at VOG), I've almost never seen my opponent do this. Is
this just because no one I've played knows this 'secret', or are there
other good reasons to hold the cube for a while until gaining an advantage?
Any input on this is appreciated...
--
Gregg Cattanach
gcattanach...@prodigy.net
Zox at GamesGrid, VOG, FIBS
>I have read in several articles and books that at the match score of
>2-away, 2-away, it is rational for either player to double on their first
>roll, converting the game to double-match-point. I understand some of the
>basic rationale behind this, and I see that the JellyFish always does this,
>too.
>
>My question is: in actual play on the Internet (1000 exp. points at
>GamesGrid, 800 at VOG), I've almost never seen my opponent do this. Is
>this just because no one I've played knows this 'secret', or are there
>other good reasons to hold the cube for a while until gaining an advantage?
>
Usually a player will delay doubling at -2/-2 because they want to
give their opponent an opportunity to lose their market by waiting too
long to double. Of course this strategy cuts both ways.
Dropping at -2/-2 gives the trailer 30% match winning chance using the
Woolsey-Heinrich table. After the cube is turned gammons have no added
value. So a player should correctly take if s/he has at least 30%
chance to win this game independent of gammons. In other words the
window is 30-70%.
Early in the game at this score I always ask myself a modification of
Woolsey's rule: if I roll my BEST number and my opponent rolls their
WORST, am I >70%? If I see even one such sequence I double, otherwise
I shake. Moreover, I assume somone will always double early. I make my
plays as if it were double match point from the beginning, emphasizing
maximum winning potential and not maximal gammon-adjusted equity [with
the cube at 1].
JF is programmed to double at the first legal opportunity at -2/-2,
irrespective of whether it is a favorite or an underdog. There is
nothing wrong with this approach and it corrects the gammon price
right away, leading to a dmp play strategy after the first sequence.
SW is more clever, but ends up misplaying the early game at -2/-2. SW
also looks for a market losing sequence prior to doubling, but as I
have previously written, SW gives EXTRA value to gammons, assuming a
gammon price of 0.75, when the effective gammon price [Dave
Montgomery's concept] is 0.00. This is a strategic error.
Stuart
SKM> Dropping at -2/-2 gives the trailer 30% match winning chance using the
SKM> Woolsey-Heinrich table. After the cube is turned gammons have no added
SKM> value. So a player should correctly take if s/he has at least 30%
SKM> chance to win this game independent of gammons. In other words the
SKM> window is 30-70%.
Isn't the calculated doubling window just 50-70%?
I'm not sure if doubling is a good idea if you're an underdog. Even if
there are market losers, there are probably also market losers for your
opponent, let's say you're 40% now, roll badly and get at 25%. If you
had doubled, you would've had to play on for the match with 25% MWC,
otherwise, you would just drop the coming double and still be at 30%
MWC.
And then there's also the issue of your opponent making an error. All in
all, I'd say you'd better not double when behind.
--
Zorba/Robert-Jan
>Isn't the calculated doubling window just 50-70%?
>
>I'm not sure if doubling is a good idea if you're an underdog. Even if
>there are market losers, there are probably also market losers for your
>opponent, let's say you're 40% now, roll badly and get at 25%. If you
>had doubled, you would've had to play on for the match with 25% MWC,
>otherwise, you would just drop the coming double and still be at 30%
>MWC.
>
>And then there's also the issue of your opponent making an error. All in
>all, I'd say you'd better not double when behind.
I think these are all valid points; I was in error when I said the
window opened at 30%. The bottom of the window is the point where the
doubler gets marginal positive value from the double and this is
indeed 50% [ignoring second order factors like the free drop].
JF will double immediately at -2/-2 even after an unfavorable
sequence. It may be a small technical error, but the trailer picks up
some equity at least in theory by eliminating the effect of gammons.
I did not mean to imply that one should double when far behind. Thanks
for pointing out my error mapping the market window.
Stuart
If you assume that your opponent is the proverbial perfect player, then
delaying only works if they wait too long and `accidentally` overshoot their
market window. As Stuart says, that cuts both ways. But isn't there also
value in waiting to see if your opponent knows what they're doing at this
match score? If your opponent doesn't know what to do, then the chance of
them losing their market is much higher.
In that case by delaying you are risking the chance of you `accidentally`
overshooting, versus the reward of your opponent `accidentally` overshooting
PLUS overshooting because they didn't know what they were doing.
So doesn't that make it worth waiting? I always like to wait a few rolls and
take the risk of losing my market, just to see how the game's going. If my
opponent looks to be getting the better start, then I hold off cubing and
wait for them to make a cube mistake.
Gavin Anderson
When playing a better player, I double right away. First, I want to
give them just one game, not two, to have their better checker play
grind down my hoped-for better-than-average luck. Second, I consider
that I am more likely to make a cube error than they are. For
precisely the opposite reasons, I hold the cube when playing a weaker
player.
Since you know that it's theoretically right to turn the cube, one way
to look at it is this: "If I were my opponent and the cube came over,
would I be happy?" Surely, if I were playing a world-class player and
I got the opening roll, even a poor one, and he doubled, I'd be happy.
On the other hand if he were a poor player, I'd say "Hmmm, I guess he
knows more theory than I thought, oh, well, of course I have to take."
So that ought to govern how I would play.
>On 16 Oct 1998 22:27:00 GMT, "Gregg Cattanach"
><gcattanach...@prodigy.net> wrote:
>
>>I have read in several articles and books that at the match score of
>>2-away, 2-away, it is rational for either player to double on their first
>>roll, converting the game to double-match-point. I understand some of the
>>basic rationale behind this, and I see that the JellyFish always does this,
>>too.
>>
>>My question is: in actual play on the Internet (1000 exp. points at
>>GamesGrid, 800 at VOG), I've almost never seen my opponent do this. Is
>>this just because no one I've played knows this 'secret', or are there
>>other good reasons to hold the cube for a while until gaining an advantage?
>>
(Stu responded:)
>Usually a player will delay doubling at -2/-2 because they want to
>give their opponent an opportunity to lose their market by waiting too
>long to double. Of course this strategy cuts both ways.
(snip)
BTW, this is a key part of the "proofs" that it is correct to double
at the first sight of a market loser, because "the opponent will never
make the mistake of losing his/her market". If you throw out this
assumption, then, as Stu points out, you can gain by waiting to double.
There is ANOTHER way to gain by not doubling at first sight of a
market loser, and this second opportunity seems to manifest itself more
often than many would guess. There is an example in the Ortega/Kleinman
book ("Cubes and Gammons at the End of a Match") involving even top level
players. I saw one at the Indiana Open (where Doug Roberts was the 'lucky'
recipient of his opp's error). That other path is:
Your opponent may drop a double that is actually a take.
In simple terms, if you offer an opponent a cube at the -2,-2 score when
the position is close to equal, very few will drop. If you wait until
you are a significant but not overwhelming favorite (maybe between 60 and
65% game winning chances) you will sometimes find that your opp will drop!
Of course, if you wait too long (until you are more than a 70% favorite)
and your opp correctly drops, then you have lost your market, which is
why Stu said waiting "cuts both ways".
On the flip side of the coin, an argument for doubling immediately
(even without sight of market losers) is that you then simplify your thinking
to just trying to win the game (under DMP conditions) with checker play.
If you are constantly trying to decide "should I double yet", it could hurt
your concentration. And there is also the extreme case where you actually
FORGET that the cube is alive. This tactic also enters in post-Crawford
cubes. Doubling immediately may be a SMALL error, but forgetting to double
may be a LARGE one. Making a small error consistently might leave you better
off in the long run than taking the chance that you will make a large error
by forgetting to double.
The bottom line on the -2, -2 doubling debate is that it doesn't make
much difference unless someone REALLY screws up. Some (like Rob Maier) are
of the opinion that it receives WAY too much discussion here, and wastes
time which could be better spent attacking REAL backgammon problems. (But,
should we really be listenting to Rob??)
Chuck
bo...@bigbang.astro.indiana.edu
c_ray on FIBS
Claes Thornberg
--
______________________________________________________________________
Claes Thornberg Internet: cla...@it.kth.se
Dept. of Teleinformatics URL: NO WAY!
KTH/Electrum 204 Voice: +46 8 752 1377
164 40 Kista Fax: +46 8 751 1793
Sweden