5-3 opening roll.
William C. Bitting
: Simply put why is making the 3 point with a 5-3 stronger
: than bringing up builders?
[cut]
: Am I missing something basic, started reading
: Magriel recently and the players I usually play with
: prefer the 5-3 build. Whereas I've been making the
: 3 point on faith.
I'm not much on an authority on this, but the way backgammon is played at
the intermediated and expert level has changed dramatically since the mid
1970's ..perhaps in 2 or 3 steps. Beginning with TD-Gammon, the neural net
programs have quanitified the 'equity' differences between plays. I'm not
sure I understand it all at more than a general level, but the end result
is that many of the standard moves of the 1970's are both very dated (and
recognizable) and out-dated.
In December, mloner and jellyfish completed their 2nd 100 match series, 5
points each, and in the 792 games, there were 46 opening rolls of 5-3, 31
by jellyfish, 15 by mloner. (These are both neural net programs, btw.) All
46 were played 8-3 6-3. Whether some experts still use a different play, I
don't know. ..wcb on FIBS
PS. Keep reading Magriel! While I've not yet been able to find a copy, it
is still regarded as perhaps the best backgammon book ever, and I'm sure
much of the strategy and tactical elements it developes are basic to
learning backgammon. You'll find Jeremy Bagai, in his notations of his
match with Kit Woolsey, refers to this book numerous times. Virtually all
the moves are discussed not only by Kit and Jeremy, but also TD-Gammon
(Gerry Tesauro). You will find the match available on Stephen Turner's
web site ..well worth studying.
Roger Hall
Simply put why is making the 3 point with a 5-3 stronger
than bringing up builders?
It essentially kills your chances of building a solid
prime against your opponent. Admittedly it gives you
a second point in your home board but....
Am I missing something basic, started reading
Magriel recently and the players I usually play with
prefer the 5-3 build. Whereas I've been making the
3 point on faith.
Anyone?
Kit Woolsey
: Simply put why is making the 3 point with a 5-3 stronger
: than bringing up builders?
: Anyone?
In the late 1970's, a majority of the experts were playing 13/10, 13/8
with an opening 5-3. The emphasis was on playing purely, and going to
the three point seemed too deep.
Since then the tide has shifted. Now almost all experts make the three
point with an opening 5-3. There are several reasons.
1) A point is a point. Just making an second inner board point increases
an opponent's chances of flunking when hit from 1/36 to 1/9.
2) The three point is part of a potential full prime from the three point
to the eight point.
3) Playing 13/10, 13/8 starts to strip the midpoint very early, which can
lead to awkwardness later on.
4) The blot on the ten point gets hit with 5-4 and 6-3. Since these are
otherwise very unproductive responses, the cost when the opponent rolls
one of them is quite severe.
5) The builder on the ten point is not functioning fully due to
duplication -- it is two away from the men on the eight point, while the
men on the eight point are two away from the men on the six point. This
means that followup rolls of 3-1 and 5-3 both make two decent points.
If you stylistically still like bringing the builders down, go for it.
However computer rollouts have consistently indicated that the play is
considerably inferior to making the three point, which is the same
conclusion experts came to over the past decade without the computers.
Kit
J S Pluim
>
> Simply put why is making the 3 point with a 5-3 stronger
> than bringing up builders?
[snip]
> Am I missing something basic, started reading
> Magriel recently and the players I usually play with
> prefer the 5-3 build. Whereas I've been making the
> 3 point on faith.
Making the 3 point does leave an ugly gap, but the 3 point still
forms part of the natural 6-prime, from the 8 to the 3 point.
It doesn't leave any blots, and it wins more gammons should
you be in a position to carry out a blitz.
Magriel himself changed from bringing down builders to making
the 3 point.
pluim.
Daniel Murphy
>Simply put why is making the 3 point with a 5-3 stronger
>than bringing up builders?
>It essentially kills your chances of building a solid
>prime against your opponent. Admittedly it gives you
>a second point in your home board but....
First, the difference in value between the two plays (making the 3-point,
or playing two checkers from the midpoint) is not large, so it's not
surprising that over the years expert opinion has swung back and forth
between the two plays. The current consensus is
that making the 3-point is best. Computer programs such as Jellyfish,
TD-Gammon and Expert BG agree (which should be viewed not as directing
expert opinion but as confirming it; note that experts have *not* rushed
to adopt all the favorite openings of the programs, such as the 43 and
32 rolls, which the programs tend to play as building moves and experts
tend to play as splitting moves).
Second, making the 3-point does *not* kill your chances of priming
your opponent; it *does* tend to direct your efforts to filling in the
5-point and 4-point -- points you usually want to make anyway -- in
order to make your 3-point even stronger.
You start the game with two points of a potential prime already
made -- your 8- and 6-points. Your 3-point works well with these points,
since all three points can become members of the same prime.
Theoretically, you could build a prime starting all the way out at the
11-point, but the lower primes (8-4, 8-3, 7-3, 7-2 are usually better.
They are easier to make (because you have more builders to make them with)
and tend to be more effective, because they force your opponent to stay
anchor on a low point. The endgame result is that *when* the back checkers
finally escape, they'll escape later in the game and have farther to go to
get home. Outside primes (extending from the 11- to 6-points, or 10- to
5-points) are often ineffective, because the endgame result is likely to
find you well behind in the race, struggling to break up your prime, bring
10 checkers into your home board, avoid leaving shots, and still manage to
keep your opponent's last two checkers from galloping on home.
Third, that second point you've made in your home board can be very
useful if your game turns into a blot hitting war. In such games, every
additional homeboard point you make greatly increases your opponent's
chance of fanning or having to play his rolls awkwardly.
Fourth, making the 3-point allows you to immediately put one of the extra
checkers on your 6-point to use.
Fifth, bringing two checkers from the midpoint leaves only one spare
checker there. Your midpoint is often a valuable point to keep until
quite late in the game; you want to be a little reluctant, at the game's
beginning, to lose the flexibility that midpoint spares give you.
(Example: Play 53 off the midpoint. Your next roll is 61, practically
forcing you to make the bar point. You now have no spares left on the
midpoint; you might very well wish you still had one spare left to
hit or point with, without having to give up the midpoint entirely.)
Sixth, although the 5-point is certainly more valuable than the 3-point,
a *made* 3-point is more valuable than a *potential* 5-point.
Finally, if memory serves, Paul Magriel didn't actually
recommend *either* play in his book. I would venture that how your
friends play the move depends very much on which move was most popular
when they first learned to play.
/Raccoon
Stephen Turner
>
> Simply put why is making the 3 point with a 5-3 stronger
> than bringing up builders?
>
> It essentially kills your chances of building a solid
> prime against your opponent. Admittedly it gives you
> a second point in your home board but....
>
You say "admittedly", but that's the overwhelming factor. That gives you
an advantage every time you hit from now on. Home table points are
Good Things(TM). Also it starts to unstack your 6 point.
It doesn't kill your chances of making a prime either. You can still bring
builders down from the midpoint to build a prime from 3 through to 8.
Yes, I know Magriel (as other writers of the time) prefers 13/10 13/8, but I
don't think many people now would agree. (The computers now seem to think that
even 64:8/2 6/2 isn't as bad as we thought).
Bonus question: Are there any match scores at which 13/10 13/8 is better?
--
Stephen R. E. Turner
Stochastic Networks Group, Statistical Laboratory, University of Cambridge
e-mail: sr...@cam.ac.uk WWW: http://www.statslab.cam.ac.uk/~sret1/home.html
Michael J Zehr
>Bonus question: Are there any match scores at which 13/10 13/8 is better?
Interesting question. Let me show everyone how much of this game I've
forgotten by responding. :)
Making the 3pt is better whenever you expect to be hitting your opponent
and might end up blitzing. It seems like it ought to lead to more
gammons than 13/10 13/8. On the other hand, 13/10 13/8 is a bit purer
in that you're going for the 5pt first, then the 4pt, etc.
At double match point, purity is more important than at other scores
because the game has to run to completion. In addition, gammons don't
help, and your opponent is more likely to stay back on the ace point
anyway (thus taking away your opportunity to hit and get the use of your
3pt) at double match point because ace point games are more effective at
double match point than at other scores because they trade losing more
gammons for winning more games. (From my own limited knowledge of double
match play, anyway.)
Are these factors enough to favor 13/10 13/8 at DMP? I'm not sure. If
so, you could make the same argument for -1:-3 Crawford (gammons don't
help either side).
I don't think you would want this play at post crawford (and
non-DMP) because either one side or the other is going to be trying to
win a gammon. If you're trying to win the G you make the 3pt, if you're
trying to avoid it you make the 3pt because you know your opponent will
be attacking you hard when you split your anchor and your 2nd home board
point makes youre return shots a lot more deadly.
Any other opinions?
(Good question, btw!)
-michael j zehr
David Montgomery
>>Bonus question: Are there any match scores at which 13/10 13/8 is better?
>> [than making the 3pt]
I don't think so. The second best play with 5-3 is 24/21 13/8,
which probably gains relative to 8/3 6/3 when the opener either
wants to avoid gammons (as at -1:-2 Crawford), or maximize winning
chances (as at DMP). I don't know if it gains enough in either
of these situations -- I would still play 8/3 6/3.
David Montgomery
monty on FIBS