FIBS Etiquette and doubling from behind
Aaron Casser
Being sort of new to the FIBS enviroment, and to rec.games.backgammon,
I apologize in advance if my question has been posed in the recent past.
However, since I have not seen anything on this, I will pose it.
What is the etiquette concerning doubling from behind during a lengthy
match, e.g. trailing 2-1 in a 3-pointer and doubling early since there
is no cost to you if you lose (since you were losing anyway)? Is it
considered to be good etiquette to seize upon it as an advantage, or
is this seen as manipulating your opponant into doing what you want
since he has nothing to gain and a lot to lose? Any opinions?
-----------------------------------------------------------------
Aaron Casser, Sophomore/Junior at the American University
ac8...@auvm.american.edu
Opinions expressed are solely those of the author and are
not those of anyone else, especially those of the American
University.
Opinions within may not be replayed, rebroadcast, or
transmitted in any form without the express written consent
of Aaron Casser (ha ha just kidding).
Paul Ferguson
AC8...@american.edu writes:
>
> What is the etiquette concerning doubling from behind during a lengthy
> match, e.g. trailing 2-1 in a 3-pointer and doubling early since there
> is no cost to you if you lose (since you were losing anyway)?
It's not a question of etiquette, it's a question of winning the match.
If you're trailing 2-1 in a 3 point match (and assuming that the
Crawford rule isn't in effect) then on your first turn you should
double your opponent. By doubling immediately, if you win the game,
you win the match. It turns that game into a winner-take-all game
even though you are behind. Otherwise, you have to win TWO games
(or win a gammon or backgammon) to win the match.
Note that this advice only applies to match games like FIBS where
the final score doesn't matter (i.e. losing 4-1 is no worse than
losing 3-1; the ratings change is the same).
--fergy
--------------------------------------------------------------------
Paul Ferguson | "It's a sick world, I'm a happy guy..."
pfer...@kaleida.com |
--------------------------------------------------------------------
Matthew N. Kleiman
:
: What is the etiquette concerning doubling from behind during a lengthy
: match, e.g. trailing 2-1 in a 3-pointer and doubling early since there
: is no cost to you if you lose (since you were losing anyway)? Is it
: considered to be good etiquette to seize upon it as an advantage, or
: is this seen as manipulating your opponant into doing what you want
: since he has nothing to gain and a lot to lose? Any opinions?
Etiquette...hell, who ever gave a damn about etiquette on fibs. With
all the shouting and network errors and kibbitzers, its like playing
backgammon in a friggin' subway station. If you've got time for etiquette,
then you're a better player than me.
Well, seriously (and I think others will back me up on this), if you can
double--go for it. If you're down 5-1 in a seven point match, double
away. And if you opponent is foolish enough to redouble, then tip it
again!
As for doubling when down 2-1 in a 3 point match, you better try 'help
crawford' next time you're on fibs...
(Ah, Crawford....many is the time I have cursed and blessed this mad
genius who invited the rule. Don't you just love it when the newbies
*shout* "fibs is broken, I can't double.")
Best of luck.
- Matt (berner)
Kit Woolsey
:
:
: What is the etiquette concerning doubling from behind during a lengthy
: match, e.g. trailing 2-1 in a 3-pointer and doubling early since there
: is no cost to you if you lose (since you were losing anyway)? Is it
: considered to be good etiquette to seize upon it as an advantage, or
: is this seen as manipulating your opponant into doing what you want
: since he has nothing to gain and a lot to lose? Any opinions?
It is entirely ethical and expected. In fact if your opponent reaches
match point and you win the Crawford game (as you have undoubtedly found
out, you have to wait one game to double when he reaches match point --
that is the Crawford rule), it is definitely your best strategy to double
as soon as legally possible. Similarly, suppose your opponent is two
points from winning the match and doubles you. If you choose to accept
the double, once again your best strategy is to redouble immediately
since you have everything to gain and nothing to lose. There is nothing
at all unethical about this -- you are simply taking the rules of the
game and using them to maximize your winning chances, which is the way
any competitive game should be played.
Kit
Karl Mosgofian
>Crawford rule isn't in effect) then on your first turn you should
>double your opponent.
Is this true? I have not run into this topic yet in books or articles.
What I wonder is, does it make sense to wait a little while to see what
your winning chances are? It occurs to me that if you were to double
at a point where you were an underdog, but not a huge underdog, you
might get a drop which would return you to 50/50 in the next game.
Of course, it after a few rolls you are ahead and get a drop, you lost
a chance to win the match. Maybe it balances out in the end, so the
best thing is just to double right away... anyway, it seems like
something that some of the experts might have some insight into.
Experts?
-karl
--
Karl Mosgofian ka...@csi.com
"You got any Stan Getz?" "No, I got some Smothers Brothers..." Tom Waites
Albert Steg (Winsor)
In a previous article, ka...@fcca.fcca.csi.com (Karl Mosgofian) says:
>>If you're trailing 2-1 in a 3 point match (and assuming that the
>>Crawford rule isn't in effect) then on your first turn you should
>>double your opponent.
>
>Is this true? I have not run into this topic yet in books or articles.
>What I wonder is, does it make sense to wait a little while to see what
>your winning chances are? It occurs to me that if you were to double
>at a point where you were an underdog, but not a huge underdog, you
>might get a drop which would return you to 50/50 in the next game.
>Of course, it after a few rolls you are ahead and get a drop, you lost
>a chance to win the match. Maybe it balances out in the end, so the
>best thing is just to double right away... anyway, it seems like
>something that some of the experts might have some insight into.
>Experts?
>
>-karl
Hmmm, one doesn't have to be an expert to respond to this one, but the
challenge is to explain it in the simplest terms possible... I'll try to
address the confusions I see in your thinking.
1) Does it make sense to wait a little to see what your winning chances
are? No. What may be the result? If you find yourself losing, retaining
the cube doesnt help you: if you lose the game, you lose the match anyway.
It might as well be on two.
But what if your game has improved and you are a favorite, even a
narrow one? Emboldened buy your position, you double. . .and your opponent
drops. Now, instead of being a favorite to win the match, you are back at
%50. You *wish* you had doubled earlier.
2) Might your opponent drop your cube, even if you are a slight underdog?
No. Why should he? You imply that he is sitting on, say, %54 winning
chances. Why should he "drop" that position to get back down to a %50
scenario?
3) Remember, it always hurts to get to a position where you cube your
opponent OUT at this score. For you, this game is "do or die." If you lose,
the match is over, whatever the cube is at. ...But if you win, don't you
want the cube to be at 2, so that you, too, have the match as your reward
for winning this game?
Hope that helps.
Albert
--
"When it was proclaimed that the Library contained all books,the
first impression was one of extravagant happiness. All men felt
themselves to be the masters of an intact and secret treasure.
-Jorge Luis Borges, "The Library of Babel"
Willis Elias
was saying:
*snip*
|>It is entirely ethical and expected. In fact if your opponent reaches
|>match point and you win the Crawford game (as you have undoubtedly found
|>out, you have to wait one game to double when he reaches match point --
|>that is the Crawford rule), it is definitely your best strategy to double
|>as soon as legally possible. Similarly, suppose your opponent is two
*snip*
I have a question about the strategy Kit discussed..
Supposing I am two away after winning Crawford's. Say my opponent gets the
first roll and it happens to be an excellent opener: say 3-1, 4-2 or 6-1.
Is it still in my best interest to double right before my first roll of
the match? Is there any possible benefit to me holding off on the double
and waiting until I gain an advantageous position, _then_ double with the
hopes that my opponent may drop?
I know what Kit is saying is probably the best strategy and I am probably
"over-thinking" the issue (or just being stupid at this point..). But, I
have played a few games after winning Crawford's. In these games I have
either doubled immediately or, alternatively, I have held off on the double
(because of my opponent's great opener) until I could gain a clear
advantage. Doubling after this point _has_ led to some drops and some
takes (to my joy). My sample size is not adequate enough to draw any sound
conclusions.
Any advice/analysis about Kit's recommended strategy and the reasons
against using any other strategy, will be greatly appreciated. Also, what
about on the other side of the cube? If the cube is turned after the first
or second roll, what kind of information does the person 1-away use to make
a sound decision to take or drop..?
Thanks <in advance>.
avernesse on FIBS
=======================================================================
Fortune:
--------
America was discovered by Amerigo Vespucci and was named after him,
until people got tired of living in a place called "Vespuccia" and
changed its name to "America".
-- Mike Harding, "The Armchair Anarchist's Almanac"
Willis Elias
was saying:
*snip*
<karl's question about wether to automatically double deleted>
It seems I had the same "confusions in my thinking" as karl.. Thanks for
putting things into better perspective for me Albert.
avernesse on FIBS
=======================================================================
Fortune:
--------
For every credibility gap, there is a gullibility fill.
-- R. Clopton
Gerald E Mortensen
: Willis Elias (wi...@BIGSKY.AERO.ORG) wrote:
: : I have a question about the strategy Kit discussed..
: : Supposing I am two away after winning Crawford's. Say my opponent gets the
: : first roll and it happens to be an excellent opener: say 3-1, 4-2 or 6-1.
: : Is it still in my best interest to double right before my first roll of
: : the match? Is there any possible benefit to me holding off on the double
: : and waiting until I gain an advantageous position, _then_ double with the
: : hopes that my opponent may drop?
[this is my second attemp at answering this; hopefully the first one
got deleted...]
your opponent will [should] drop any time you're 50% to win the game
since by dropping he gets one last winner-take-all chance. but
if you're really over 50% for the game you don't want him to drop,
since that drops you back to 50% to win the match. so basically you
never want him to drop (assuming perfect play. if you think your game
equity judgement is a whole lot better than his that might be a
different story... )
split all the post-crawford games -2:-1 games into 2 groups:
1) games where at some point in the proceedings you're over 50% to
win the game
2) the other kind
now for type 1), when you get to the point where you are the favorite
you will wish the cube was already at 2 without your opponent being
able to cut his losses
as for type 2), you lose either way. better luck next time :)
wilfo
--
** Gerald E. Mortensen (Jay) Syracuse Research Corp. ***
** Research Engineer Merrill Lane ***
** (315)426-3269 -- j...@syrres.com Syracuse, NY 13210 ***
Dick King
|> Gerald E Mortensen (gmor...@newstand.syr.edu) wrote:
|> : Willis Elias (wi...@BIGSKY.AERO.ORG) wrote:
|>
|> : : I have a question about the strategy Kit discussed..
|>
|> : : Supposing I am two away after winning Crawford's.
|>
|>
|> since by dropping he gets one last winner-take-all chance.
Indeed when you are -1 and your opponent is -2 you drop any disadvantage. This
is called the "free drop". Your opponent needs to win _that_ game to win if
you take, but he* needs to win the _next_ game if you drop, and the question is
which game you would rather play.
When you are -1 and your opponent needs an odd number of points you do not have
a free drop because your opponent will need one fewer game to win the match
[assuming taken doubles** from then on] if you do drop. For example, if he is
at -3 he needs this game and another if you take the double, but he only needs
that other game, or perhaps yet a third game if you take a free drop then, if
you drop.
Life might get interesting if your opponent is at -4.
If your opponent rolls a 31 and you roll a 52 you will drop when he doubles,
because you would rather play the next two games than this one and another.
This is of course a free drop.
HOWEVER
suppose your opponent gets his first chance to double after you are only
bruised, not battered. It's a legitimate question whether you should play on,
therefore forcing your opponent to win this game plus your choice of the next
two games, or take your almost-free drop at which point he needs to win the
next two games and you don't get to duck out on your obligation to finish what
you have started.
Kit, you deal with this sort of stuff. Do you ever think about whether you are
willing to play a slightly unfavorable game in order to preserve a free drop
for later in the match when you might be faced with a real stinker, or do you
drop even slightly unfavorable games when your opponent is some even number >2
of points away from snatching victory out of the jaws of defeat?
** I promised to come back to odd numbers.
Like many things in games distinctions blur when numbers get large.
Is it wrong to drop a truly miserable start if your opponent is at -437 and
you are at -1?
-dk
* please don't ding me for sexist language. Thank you.
** i'll come back to this
Christopher Yep
1. If you want to consider the limit case, then make a few reasonable
assumptions and see where it takes you. Here are some reasonable
assumptions:
For large k, with match score -2k:-1 or -2k+1:-1 [both post-crawford]
the leader will win 50-52%
trailer will win 48-50% (since he will take a few extra risks to try
to win a gammon or bg)
of the trailers wins,
70-80% single wins
20-30% gammons
< 2% bgs [I'm being very generous with the bounds]
2. From this, you may solve the limit equation. I actually did it, this
last summer out of curiosity, using the brief data (for my assumptions) that
Kit Woolsey provided in his "How to Play [or maybe "Win"] Tournament
Backgammon" book. My conclusion was that for large k,
lim [-2k+3:-1] lim [-2k+2:-1]
(k --> infinity) ---------- = (k --> infinity) ---------- ~= 0.57
[-2k+1:-1] [-2k:-1]
Here, [-m:-n] represents the match equity of the m-away player
[post-crawford].
I believe the values that I used to obtain the .57 ratio were 49% wins for
trailer, and 21 (or 22%) gammons out of these trailer wins (no bg's). I'll
try to find my work and perhaps post some of it after I edit it a bit.
Btw, the ratio converges very quickly. (Note that even [-3:-1]/[-1:-1] =
.3/.5 = .6, initially very close to the final value of .57)
3. Some more:
Consider the trailer's 1st opportunity to double. It will either
occur after 1-ply (ie leader made 1 move), or after 2-ply (both players
moved once).
Suppose k is large
Case A: Leader is -1:-2k+1
Claim: if trailer doubles, leader should not drop.
A corollary of the .57 limiting ratio (above) is that the leader
should not drop at -1:-2k+1 [post-crawford], unless his chance of being
gammoned is .57/.43 ~= 1.33 times his chance of losing [work out the
numbers]. (This assumes no bg's; if you prefer, you can also take into
account bg's, but it will hardly change the results.) From this, if we
make the very reasonable assumption that after the 1st roll, the leader
will always have at least 40% chance to win the game, and losing gammon no
more than 20% of the time (i.e. the "worst" case being trailer wins 40%
single wins, 20% gammon wins), then he still must take a double. The
numbers are actually not so "bad." The relevant equities are:
leader wins --> trailer's eq. = 0.
leader drops --> trailer's eq. = only very slightly less than .57x
leader takes and loses single --> trailer's eq. = .57x
leader takes/loses gammon --> trailer's eq. = x
By playing on, the leader risks ~ .43 (if he's gammoned) for the chance
to gain .57 (if he wins).
Btw, the leader's strategy at -1:-5 (or even -1:-3) is almost identical to
his strategy at -1:-437 or -1:2k+1 (large k). Nothing "blurred."
Case B: leader is -1:-2k
Now the leader has a free drop. Since k is very large, we can assume
that if the trailer is to win this match, that every possible starting
2-ply dice sequence will occur during his "comeback run" of games. In
particular, choose the worst possible sequence leading up the trailer's
initial double. By "worst," consider match equity. Using the relevant
equities (0, .57x, x) and an estimate of the different game likelihoods
(leader win, trailer single win, trailer gammon win), figure out which
sequence would be the worst for the leader to take. This particular
sequence, then, is the one that he will want to drop. He should accept
all other sequences, even those which put him at a disadvantage, and then
drop the 1st time that the "worst" sequence occurs. Actually, the
trailer may choose to play on for gammon on a roll-by-roll basis after
rolling his best possible 2-ply sequence. This would complicate things a
bit. However, the original question was whether or not the leader should
always drop (if he has a free drop available) if he is at a
disadvantage. The answer to this is "no."
For k not so large, he can make some adjustments. For k=1, he should
drop whenever he wins < 50% of the time. For k=2,3,4,5,... he can keep
lists of which sequences he wants to drop. The size of these lists will
converge to 1, and for some M, if k > M, then this k-list will only include
one particular sequence.
I hope this article is somewhat understandable.
Chris
Kit Woolsey
: HOWEVER
: suppose your opponent gets his first chance to double after you are only
: bruised, not battered. It's a legitimate question whether you should play on,
: therefore forcing your opponent to win this game plus your choice of the next
: two games, or take your almost-free drop at which point he needs to win the
: next two games and you don't get to duck out on your obligation to finish what
: you have started.
: Kit, you deal with this sort of stuff. Do you ever think about whether you are
: willing to play a slightly unfavorable game in order to preserve a free drop
: for later in the match when you might be faced with a real stinker, or do you
: drop even slightly unfavorable games when your opponent is some even number >2
: of points away from snatching victory out of the jaws of defeat?
This is really getting into some technical stuff, yet it is a very valid
point. Certainly when you have the potential for a future free drop you
will be more inclined to save it if your are at a *slight* disadvantage.
However, usually the disadvantage isn't as slight as one might imagine.
Pretty much any time the opponent wins the opening roll and you don't
fire back a clearly good return you are at a fairly substantial
disadvantage, which means the free drop should be exercised immediately.
I would recommend exercising it any time it is clear to you that you are
at a disadvantage, even if you think the disadvantage may be slight.
Kit