opening roll 6-4
Unknown
I did some rollouts of opening moves a while back and top play for
this this roll was to make the 2 point.
Chuck Bower did a more comprehensive rollout in the latest
Flint Area Backgammon News and came to the same conclusion.
So..... why is it whenever I look at matches from some of the best
players they invariably play 24-18/13-9 or 24-14 ?
Chuck Bower
Well, as long as you're going to bring my name into this, I
guess I have to comment (and thanks for the plug)!
If you read my article CAREFULLY (and I recommend this to ANY
interested person) then you will see that the three most common
plays for the opening 64 are QUITE CLOSE:
Play Opener's Statistical
cubless Significance
$ equity (in std.dev.)
8/2,6/2 +0.010 ---
24/14 +0.007 1
24/18,13/9 +0.005 1.7
Without repeating the ENTIRE article (and for $1.66 per issue,
why aren't you a subscriber already!), I summarize that the
third column asks "What are the chances that this play is
better than the 'best' play?" in standard deviation units.
In addition, JF does not play PERFECT backgammon (yet) so
small differences could be overshadowed by better play.
Bottom line is: JF rollouts say that these three plays
are so close that other factors (like "how does your opp
play and how do YOU play") override the small differences
shown here.
Now that I've painted a clearer picture of what I
presented, I'll let the REAL EXPERTS argue why their favorite
moves for the 64 opening are best...
Chuck
bo...@bigbang.astro.indiana.edu
c_ray on FIBS
P.S. "you" above is meant to be plural and general, and not
directed at midas. Now, if this newsgroup were in just about
ANY OTHER LANGUAGE I wouldn't need to make this disclaimer!
illium
In article <34b8ea2f...@news.demon.co.uk>,
(midas) wrote:
>Subject: opening roll 6-4
>From: midas
>Date: Sun, 11 Jan 1998 15:53:37 GMT
>Newsgroups: rec.games.backgammon
>
>I did some rollouts of opening moves a while back and top play for
>this this roll was to make the 2 point.
>Chuck Bower did a more comprehensive rollout in the latest
>Flint Area Backgammon News and came to the same conclusion.
>So..... why is it whenever I look at matches from some of the best
>players they invariably play 24-18/13-9 or 24-14 ?
Hi midas
Not being an expert I hesitate to comment, but since you asked...
The advantage of course in making the 2-point is that it makes re-entry of
men from the bar more difficult, however it does tie up two men on a point
that you really don't need to make this early in the game. These men are
probably better disposed of in making the 5,4,bar and 3-points. Also, the
2-point is not part of a possible prime spanning the 8 to 3-points.
I think the thing with rollouts is that the further you are from the end of
a game the less accurate they are, there is a greater margin of error at
least. This means that rolls that are very close together can effectively
be equivalent. I usually consider rolls that are within about 3% of each
other roughly equivalent but this of course depends on the circumstances.
I would imagine that experts still play the two moves you suggested because
they offer greater flexibility with little, if any, sacrifice of equity.
Personally I would probably play the 24-18/13-9 move if I was behind and
wanted to attack the game, and I'd play 24-18-14 if I was well in front and
wanted to consolidate. In a single point game making the 2-point could well
be the thing to do, but with matches of 5 points and more the other moves
are certainly worthy alternatives.
Regards
William Hill, illium on FIBS.
Stuart Katz, MD
On Sun, 11 Jan 1998 15:53:37 GMT, (midas) wrote:
>I did some rollouts of opening moves a while back and top play for
>this this roll was to make the 2 point.
>Chuck Bower did a more comprehensive rollout in the latest
>Flint Area Backgammon News and came to the same conclusion.
>So..... why is it whenever I look at matches from some of the best
>players they invariably play 24-18/13-9 or 24-14 ?
My choice is tactical. Usually I will run. If I think I'm the
significantly stronger player I will split unless I need only a single
game to close a match. If I need gammons as the trailer at
Crawford/even or post-Crawford/>2 situations I will anticipate a
potential blitz by making the 2-point.
Stuart
Donald Kahn
I could be all wet about this, but I like making the deuce point
when I am ahead in the match. If I get a third point, I will be a lot
more comfortable taking doubles which would be too dangerous
gammon-wise with no points made. (A 3-point board has a discouraging
effect on loose hits inside.) And even a 2nd point (at least
according to JF) is a big equity improver.
I also like the deuce point when opponent starts with a pointmaking
roll.
deekay
Don Woods
ill...@whills.demon.co.uk (illium) writes:
> I think the thing with rollouts is that the further you are from the end of
> a game the less accurate they are, there is a greater margin of error at
> least.
Also, you have to remember that the rollouts are being done by a program
(probably Jellyfish), so the result really means that if your playing
style matches that of Jellyfish, and you're playing against an opponent
whose style is that of Jellyfish, then making the 2-pt on an opening 6-4
might be the best move.
Of course, since JF plays pretty darn well, maybe your playing style
SHOULD be that of Jellyfish, but there's still that pesky opponent to
consider. I.e., against human opponents it may be that other opening
rolls are more effective. Or it may not. The point is that rollouts
with the computer (perforce) playing both sides won't take into account
how best to take advantage of weaknesses in your opponent's play.
On the topic of how to play 6-4, I've found there's one other positive
effect of making the 2-pt (besides making it harder for your opponent
to enter if you later hit). It means he can't split his runners on a 1.
Thus he might have a harder time playing rolls like 2-1, 4-1, 5-1.
Certainly there are satisfactory plays for those rolls that do not
involve splitting the runners, but restricting your opponent's options
is often good.
-- Don.
-------------------------------------------------------------------------------
--
-- Don Woods (d...@clari.net) ClariNet provides on-line news.
-- http://www.clari.net/~don I provide personal opinions.
--
Gary Wong
(midas) writes:
> I did some rollouts of opening moves a while back and top play for
> this this roll was to make the 2 point.
That seems to be the normal conclusion. However, you should take this result
with a grain of salt, for the following reasons:
1) Every rollout I've seen has ranked the three plays 24/14, 24/18 13/9, and
8/2 6/2 very close -- within a couple of standard deviations, anyway. The
evidence seems to indicate that 8/2 6/2 is probably the strongest play
(for Jellyfish), but it is not conclusive. The difference is so small that
it would take a huge number of samples to say with any certainty. Besides,
"knowing" something is the best move isn't as good as knowing _why_ --
otherwise, how can you make use of the knowledge in other situations?
Suppose an extensive rollout concluded 8/2 6/2 is the best move for money --
would you play it at double match point? Leading or trailing post-Crawford?
For money against an opponent that you know slots and plays purely, and you
know from experience that games against her tend to turn to protracted
positional battles where strategic weaknesses will come back to haunt you?
For money against an opponent who is scared of blitzes and tends to drop
cubes he ought to take when in danger of being blitzed? Knowing all the
alternatives and their relative strengths and weaknesses is more valuable,
in my opinion.
2) Jellyfish is a very good player in most positions, but it's not perfect.
It is known to be somewhat better at blitzes than back games, for instance.
8/2 6/2 strengthens your inner board for a blitz; 24/18 13/9 is somewhat
more likely to lead to a back game if your blots are hit. Who's to say
that the reason 8/2 6/2 produces more wins for Jellyfish isn't simply
because it leads to positions it is good at, rather than being an
intrinsically `better move'?
> So..... why is it whenever I look at matches from some of the best
> players they invariably play 24-18/13-9 or 24-14 ?
I'm no "best player" so to get an answer from an expert you'd have to ask one;
however I expect the answer you'd get is that most good humans feel that making
the 2 point as early as the opening move is too impure -- it's just begging
for the opponent to anchor in front of it and leave you with two chequers out
of play. It hurts your future blockading strength, because you already start
the game with the 8 point and two points 6 pips apart do not go well together
(they can never be part of the same prime). It also hurts your timing somewhat
which is an important factor in back games and prime-vs-primes -- the problem
is that the men you bury on your 2 point have nowhere useful left to go and
force you to play other moves elsewhere on the board which may hurt your
position.
Having bashed the 8/2 6/2 move so much it's worth mentioning its good points
too. It does of course secure another inner board point which is not to be
sneezed at -- a stronger board than your opponent is always an advantage and
any time you put him on the bar, the more points you have, the less chance he
has of entering and the less choice of moves he has when he does enter. This
is true whether it's your 5 point or 1 point that's made. If the next few
rolls start to suggest a blitz is a good approach for you, you'll be glad you
made your 2 point.
Overall I prefer 24/18 13/9 on an opening 64 (though I'd play 24/14 at
double match point). I wouldn't argue against anybody (or any computer)
who played 8/2 6/2, though -- in my opinion the plays are close enough that
personal preference can and should decide between them. Making a play that
leads to a position you're familiar and comfortable with is more of an
advantage to you than selecting a move that Jellyfish rollouts say is 0.002
better.
Cheers,
Gary (GaryW on FIBS).
--
Gary Wong, Computer Science Department, University of Auckland, New Zealand
ga...@cs.auckland.ac.nz http://www.cs.auckland.ac.nz/~gary/
Anthony Patz
There is one more argument to be made for the 2 point. For years this
move has been conventionally been such a no-no that many opponents
will write you off as a novice.
Two wrongs dont make a right, but two Wrights do make an aeroplane.
(and three rights make a left)
Patti Beadles
One of the things that I find intriguing about 8/2 6/2 is that it
cramps some of your opponent's aces. What does he do with 21, 41,
51 now?
Splitting is no longer an option, so the only reasonable play is
slotting. But now your board is stronger, which means more good
things can happen if you hit him. If his style is to always split
with aces, then he's probably going to be on unfamiliar ground after
these sequences.
Just some random 2 a.m. ramblings,
-Patti
--
Patti Beadles |
pat...@netcom.com/pat...@gammon.com |
http://www.gammon.com/ | "I trust you. It's just
or just yell, "Hey, Patti!" | that I'm scared of you."
ches...@feist.com
In article <pattibEn...@netcom.com>,
Good point about taking away your opponent's ones. It follows, then,
that if your 6-4 is in response to an opening 3-1, you would not make
your 2-point, since he now has a good one (man from his 6-point to his
5-point).
Jerry Weaver
chessart on FIBS
-------------------==== Posted via Deja News ====-----------------------
http://www.dejanews.com/ Search, Read, Post to Usenet