BG by Magriel
raf...@aol.com
Can someone give me an idea of the street price of Backgammon by Paul
Magriel. I have an oportunity to buy one, and would like to know what
a reasonable price is. Thanks!
Raffi
raffis on FIBS
M Louis
$50 at Fred Wilson Books in NY 80 East 11th St. 212-533-6381
I just got mine there, excellent condition inside, book jacket good
Robertie and others over $100 are greedy$$$!
Happy BG!
Raccoon
raf...@aol.com writes:
>Can someone give me an idea of the street price of Backgammon by Paul
>Magriel. I have an oportunity to buy one, and would like to know what
>a reasonable price is. Thanks!
A reasonable price for the book depends on where you are and how badly
you want to buy it.
Copies can be found occasionally in used bookstores in the USA for $5 to
$20 dollars. Some used bookstores, and some individuals advertising on
rec.games.backgammon, have asked for and received $100 or $120 per copy.
Magriel is a fine book, a great book, really just a wonderful book on
backgammon, truly it is, but I would only pay more than $20 for it
knowing I could it resell it to you for $100. :-)
Raccoon
Albert Steg
> raf...@aol.com writes:
>
> >Can someone give me an idea of the street price of Backgammon by Paul
> >Magriel. I have an oportunity to buy one, and would like to know what
> >a reasonable price is. Thanks!
>
> A reasonable price for the book depends on where you are and how badly
> you want to buy it.
Certainly true. If you wait long enough and can travel to some obscure
bookstores, your chances are good of finding a cheap copy. Meanwhile,
though, if you're playing for $5 a point in a field of players who are
already "past" Magriel strength, you may wind up "paying" more for the
book by waiting.
I recall playing for about a year against someone who *needed* to read
that book. It became a running catch-phrase with us. When he made a big
& basic equity blunder I would shake my head sagely and murmer "Magriel."
The book does pay for itself, in the long run, even at $100. That's why
people have been willing to pay that price.
Albert
Albert Steg
> raf...@aol.com writes:
>
> >Can someone give me an idea of the street price of Backgammon by Paul
> >Magriel. I have an oportunity to buy one, and would like to know what
> >a reasonable price is. Thanks!
>
> A reasonable price for the book depends on where you are and how badly
> you want to buy it.
Certainly true. If you wait long enough and can travel to some obscure
bookstores, your chances are good of finding a cheap copy. Meanwhile,
though, if you're playing for $5 a point in a field of players who are
already "past" Magriel strength, you may wind up "paying" more for the
book by waiting.
I recall playing for about a year against someone who *needed* to read
that book. It became a running catch-phrase with us. When he made a big
& basic equity blunder I would shake my head sagely and murmur "Magriel."
James Eibisch
On Sun, 09 Jun 1996 21:43:40 -0700, raf...@aol.com wrote:
>Can someone give me an idea of the street price of Backgammon by Paul
>Magriel. I have an oportunity to buy one, and would like to know what
>a reasonable price is. Thanks!
$80 seems to be the current going price, although $50 seems a common
next step down. I got my copy through a quick shout on FIBS. Someone
replied, offering theirs for $50 which I took up. I was rather
disappointed when I read through it for the first time, but from the
second reading I could see its value - it is undoubtedly a required
purchase IMO.
Joseph A. Wetherell
In <31c04bc8.1761829@news> jeib...@revolver.demon.co.uk (James
Hi All,
Is Magriels book necessary if one has studied THE BACKGAMMON HANDBOOK?
I also studied Bill Robertie's BACKGAMMON FOR WINNERS.
Regards,
Joe joeweth on FIBS
Raccoon
jeib...@revolver.demon.co.uk (James Eibisch) writes:
>I was rather
>disappointed when I read through it for the first time, but from the
>second reading I could see its value - it is undoubtedly a required
>purchase IMO.
It gets even better the third time through.
Raccoon
tex...@ix.netcom.com(Joseph A. Wetherell) writes:
>Is Magriels book necessary if one has studied THE BACKGAMMON HANDBOOK?
Magriel's is the rare (the only) introductory book that is not just for
beginners; reading (and rereading) it can be very helpful to more
experienced players trying to improve their game.
I was not enthusiastic about Heyken and Fischer's _The Backgammon
Handbook_ (and if it's worth the US$31.45 I paid for it, then Magriel is
easily worth $100).
I found much of H & F's advice good but insufficiently explained. For
example, they say "Play safe. Never give your opponent the chance to hit
your pieces without good reason" -- not very useful if you don't
understand what those good reasons are. Elsewhere their advice is wrong:
"In a long-running game, double if you lead by over 10% of your pipcount.
If this lead is over 15%, your opponent should concede."
>I also studied Bill Robertie's BACKGAMMON FOR WINNERS.
This book is an adequate beginner's introduction to backgammon. I don't
think it aspires to be more. Of other beginner books in print, the only
one I would recommend is Jacoby's _The Backgammon Book_. It's outdated but
fairly complete, well written, cheap, and there's not too much it gets
wrong.
Of beginner books which are out of print but often found in used book
stores, Tim Holland's is ok, John Longacre's is fair, Prince Obolensky's
is poor, and Walter Gibson's is, I think, the worst book ever written on
backgammon (I loved the part where he encourages you to double when
losing, so that your opponent will immediately double you back, after
which you might get lucky and double to 8. Whoa!).
Daniel Murphy
San Francisco CA
rac...@netcom.com
SEM
> example, they say "Play safe. Never give your opponent the chance to hit
> your pieces without good reason" -- not very useful if you don't
> understand what those good reasons are. Elsewhere their advice is wrong:
> "In a long-running game, double if you lead by over 10% of your pipcount.
> If this lead is over 15%, your opponent should concede."
>
Claes Thornberg
As far as using percentage to determine doubling in race i use the
following
>= 8% lead (Your opponent has 8% more pips than you) ==> Double
>= 9% lead ==> Redouble
If you are 12% behind your opponent (you have 12%, or more, pips more
than your opponent) ==> drop
Regards,
Claes (claest @ FIBS)
--
______________________________________________________________________
Claes Thornberg Internet: cla...@it.kth.se
Dept. of Teleinformatics URL: http://www.it.kth.se/~claest
KTH/Electrum 204 Voice: +46 8 752 1377
164 40 Kista Fax: +46 8 751 1793
Sweden
Stephen Turner
SEM wrote:
>
> > Elsewhere their advice is wrong:
> > "In a long-running game, double if you lead by over 10% of your pipcount.
> > If this lead is over 15%, your opponent should concede."
> >
> What is the correct version of this advice?
Roughly, 8% lead to double from centre, 9% to redouble, at most 12% behind to
take. (Incidentally, I'm always amazed that there is a working formula which
is a constant percentage of the pipcount).Short races and positions with
odd bearoff distributions may need modification, which can be done by the
Thorpe count and other specialised counts.
Before anyone asks, the Thorpe count is pipcount, plus 2 for each chequer
still on the board, minus 1 for each inner board point with any chequers on,
plus 1 for each chequer on the ace point. If the leader's count is over 30,
he must then add 10%. The leader should double if his count is at most 2
greater than the trailer's, redouble if only 1 greater, and the trailer should
drop if the leader is 2 ahead. Got that?
At different match scores, you need to know your percentage winning chance.
Does anyone know a SIMPLE formula for that? (That doesn't rely on me having to
multiply by numbers like 73 over the board!)
--
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
"You may notice that your Customer Reference Number has changed" British Gas
Robert Scibelli
On Jun 21, 1996 10:24:27 in article <Racing doubles (was Re: BG by
Magriel)>, 'Stephen Turner <sr...@statslab.cam.ac.uk>' wrote:
--some deletia--
chance.
>Does anyone know a SIMPLE formula for that? (That doesn't rely on me
having to
>multiply by numbers like 73 over the board!)
>
>--
>Stephen R. E. Turner
formula for estimating the values in his match equity table. And now you
want me to post it, but my copy is at home and I am in the office, and I
haven't memorized it yet, and I am not sure about the copyright thing. So
why am I posting this? Well, now you can ask Kit next time he is in the
neighborhood.
--
Robert Scibelli
xav...@pipeline.com
"If the election were held today, 72% of registered voters would be
completely surprised because they thought the election was in November."
Claes Thornberg
In article <4qeod2$1...@news1.t1.usa.pipeline.com> xav...@nyc.pipeline.com(Robert Scibelli) writes:
> 'Stephen Turner <sr...@statslab.cam.ac.uk>' wrote:
>
> >At different match scores, you need to know your percentage winning
> chance.
> >Does anyone know a SIMPLE formula for that? (That doesn't rely on me
> having to
> >multiply by numbers like 73 over the board!)
>
> Kit Woosley's fine pamphlet "How to play Tournament Backgammon" contains a
> formula for estimating the values in his match equity table.
I don't think Steve wants you to post formulae to approximate Kit's
match equity table. I think he's quite comfortable with his own
formula (right Steve?). What he wants to know, I believe, is if there
is any formula giving an answer to questions like:
"If I'm behind X% (pips) in the race, what are my winning chances in
this game"
Knowing the answer of this, would greatly improve your possibility to
make a correct take/drop decision at different match scores.
So, does anyone know of such a formula?
Claes Thornberg (claest @ FIBS)
"I can't believe there are so many books on backgammon. All you have
to do is roll the dice and move your men according to the pips."
______________________________________________________________________
Claes Thornberg Internet: cla...@it.kth.se
Dept. of Teleinformatics
Stephen Turner
Robert Scibelli wrote:
>
> Kit Woosley's fine pamphlet "How to play Tournament Backgammon" contains a
> want me to post it, but my copy is at home and I am in the office, and I
> haven't memorized it yet, and I am not sure about the copyright thing. So
> why am I posting this? Well, now you can ask Kit next time he is in the
> neighborhood.
Not a formula for the match equity table. I know about that. A formula for
your percentage winning chances in a race.
--
Stephen R. E. Turner
Chuck Bower
In article <31CA6A...@statslab.cam.ac.uk>,
Stephen Turner <sr...@statslab.cam.ac.uk> wrote:
>SEM wrote:
>>
>> > Elsewhere their advice is wrong:
>> > "In a long-running game, double if you lead by over 10% of your pipcount.
>> > If this lead is over 15%, your opponent should concede."
>> >
>> What is the correct version of this advice?
>
>Roughly, 8% lead to double from centre, 9% to redouble, at most 12% behind to
>take. (Incidentally, I'm always amazed that there is a working formula which
>is a constant percentage of the pipcount).Short races and positions with
>odd bearoff distributions may need modification, which can be done by the
>Thorpe count and other specialised counts.
(snip)
>At different match scores, you need to know your percentage winning chance.
>Does anyone know a SIMPLE formula for that? (That doesn't rely on me having to
>multiply by numbers like 73 over the board!)
>
>--
>Stephen R. E. Turner
(snip)
I have "rules" which convert both formulas to winning percentage. It is
based on the following assumption for pure races and money play:
Initial Double (i.e. centered cube) requires 70% cubeless winning chances.
Redouble requires 72% winning cubeless chances.
Drop/take line occurs at 78% winning chances.
(Note: I apologize to the persons who came up with these numbers for not
giving them due credit, but unfortunately I don't remember their
names. I will look up and post in a followup message.)
The "8,9,12" formula is from Robertie's "Advanced Backgammon" and sez:
Determine the roller's pip count (=R).
Determine the non-roller's pip count (=N).
Calculate F = N/R - 1.
Then (for money) roller has an initial double with F .ge. 0.08
a redouble with F .ge. 0.09
non-roller has a take if F .le. 0.12
(Note: .ge. means "greater than or equal to". You figure out what .le. means)
My first "rule" says cubeless winning chances for roller are:
S = 0.54 + 2*F
I leave it to the interested reader to verify the 8,9, and 12 % correspond
to 70, 72, and 78% cubeless winning chances. I CAUTION THAT THESE RULES
APPLY FOR ROLLER'S PIP COUNT IN THE RANGE 60 - 100. With less than 60,
use Thorp Count (see below). With more, see "Advanced Backgammon" for a
further discussion.
Here is the Thorp Count. (This is named after its creator, Edward O. Thorp,
author of "Beat the Dealer". It also is featured in "Advanced Backgammon".)
Define the following quantities:
P = roller's Pip count
C = number of roller's Checkers which remain on the board
W = number of roller's checkers on the Won point (er, I mean "one point")
B = number of home Board points covered by at least one of roller's checkers
Calculate T = P + 2*C + W - B.
If T .ge. 30, then T = 1.1 * T (that is, increase T by 10%).
Now, calculate the same quantities for the non-roller (and denote them
with lowercase letters), EXCEPT, don't increase by 10%, regardless of the
value for t.
Now compute V = T - t.
If V .ge. -2, Roller has an initial double.
If V .ge. -1, Roller has a redouble.
If V .le. +2, non-roller has a take.
Now for the punch line, to convert to cubeless winning chances use the
following:
S = 0.74 + V/50
(As above, the interested reader should verify that (-2,-1,+2) of the Thorp
formula correspond to 70, 72, and 78% cubeless winning chances.
In practice, these two rules work pretty well for cubeless winning chances in
the range 55% to 85%.
Try them out in a match sometime!
Chuck Bower
Bloomington IN
bo...@bigbang.astro.indiana.edu
Chuck Bower
In article <4qmshb$c...@usenet.ucs.indiana.edu>,
Chuck Bower <bo...@bigbang.astro.indiana.edu> wrote:
(snip)
>
>Initial Double (i.e. centered cube) requires 70% cubeless winning chances.
>Redouble requires 72% winning cubeless chances.
>Drop/take line occurs at 78% winning chances.
>
>(Note: I apologize to the persons who came up with these numbers for not
>giving them due credit, but unfortunately I don't remember their
>names. I will look up and post in a followup message.)
>
(snip)
The following table is taken from "Optimal Doubling in Backgammon" by
Emmett B. Keeler and Joel Spencer in OPERATIONS RESEARCH, vol 23 # 6, p. 1063
Nov-Dec 1975:
Pip Count for player on Roll
30 50 70 90 110
First dbl 0.65 0.68 0.70 0.72 0.73
Redouble 0.69 0.71 0.72 0.73 0.74
Fold 0.76 0.77 0.78 0.79 0.79
(Note: The numbers in the body of the table are Roller's winning chances.)
So, my memory was of the column with Roller's pip count = 70.
As an aside, another article of that same time period also discussed
the same topic--"On Optimal Doubling in Backgammon" by Norman Zadeh and
Gary Kobliska in MANAGEMENT SCIENCE, vol 23 #8, p. 853, April 1977. Both
articles (which deal only with NON-GAMMON games) are still valid today,
IMHO. That is more than can be said of much of the literature of that
time period.
Stephen Turner
>
> As an aside, another article of that same time period also discussed
> the same topic--"On Optimal Doubling in Backgammon" by Norman Zadeh and
> Gary Kobliska in MANAGEMENT SCIENCE, vol 23 #8, p. 853, April 1977. Both
> articles (which deal only with NON-GAMMON games) are still valid today,
> IMHO. That is more than can be said of much of the literature of that
> time period.
>
Zadeh & Kobliska's article is still correct today GIVEN THE CONTINUOUS GAME
HYPOTHESIS. In my view, this makes it still incorrect today. For those who
don't know, the continuous game hypothesis says that your equity changes
continuously, not (as it really does) in jumps. It may be a necessary
approximation in order to get any maths done at all, but really it ignores
the complexity of the problem, because it means that there can be no market
losing sequences. In Jellyfish terminology, you have to know your volatility
as well as your equity when deciding whether to double, which is just as hard.
--
Stephen R. E. Turner
Gerry Tesauro
Stephen Turner <sr...@statslab.cam.ac.uk> wrote:
>In Jellyfish terminology, you have to know your volatility
>as well as your equity when deciding whether to double, which is just as hard.
Excuse me, but let's give credit where it's due.
As far as I know, TD-Gammon was the first program
both to define the "volatility" of a position
(the standard deviation in equity averaging over
the upcoming dice rolls), and to use both equity
and volatility to make doubling decisions.
It's been doing this since 1992.
-- Gerry Tesauro (tes...@watson.ibm.com)
Stephen Turner
>
>In Jellyfish terminology, you have to know your volatility
>as well as your equity when deciding whether to double
>
>
> Excuse me, but let's give credit where it's due.
> As far as I know, TD-Gammon was the first program
> both to define the "volatility" of a position
> (the standard deviation in equity averaging over
> the upcoming dice rolls), and to use both equity
> and volatility to make doubling decisions.
> It's been doing this since 1992.
>
I'm sorry, Gerry. I stand corrected.
Tom Keith
>
> Chuck Bower wrote:
> >
> > As an aside, another article of that same time period also discussed
> > the same topic--"On Optimal Doubling in Backgammon" by Norman Zadeh and
> > Gary Kobliska in MANAGEMENT SCIENCE, vol 23 #8, p. 853, April 1977. Both
> > articles (which deal only with NON-GAMMON games) are still valid today,
> > IMHO. That is more than can be said of much of the literature of that
> > time period.
> >
>
> Zadeh & Kobliska's article is still correct today GIVEN THE CONTINUOUS GAME
> HYPOTHESIS. In my view, this makes it still incorrect today. For those who
> don't know, the continuous game hypothesis says that your equity changes
> continuously, not (as it really does) in jumps. It may be a necessary
> approximation in order to get any maths done at all, but really it ignores
> the complexity of the problem, because it means that there can be no market
> as well as your equity when deciding whether to double, which is just as hard.
While Z&K don't use the term "volatility" in their article, they do
account for it in their calculations. They do this by defining
"effective doubling points" which are used to approximate the loss
in value of the cube caused by volatility. This method works
surprisingly well considering how simple it is.
It is interesting to note that any system that proposes different
"double" and "drop" points must be taking into consideration the
possibility of market losing sequences. Why double now if there is
no sequence of rolls which will cause opponent to drop next time?
Tom
Bob Koca
>In article <31D251...@statslab.cam.ac.uk>,
>Stephen Turner <sr...@statslab.cam.ac.uk> wrote:
>>In Jellyfish terminology, you have to know your volatility
>>as well as your equity when deciding whether to double, which is just as hard.
>Excuse me, but let's give credit where it's due.
>As far as I know, TD-Gammon was the first program
>both to define the "volatility" of a position
>(the standard deviation in equity averaging over
>the upcoming dice rolls), and to use both equity
>and volatility to make doubling decisions.
>It's been doing this since 1992.
>-- Gerry Tesauro (tes...@watson.ibm.com)
I agree that volatility is very important to consider but am not sure
if the TD-Gammon definition is the best way to quantify it for
doublling purposes.
I propose a measure which I call market losing factor (mlf).
Average over the 1296 sequences the value of
min{ 0, market loss }.
Let's illustrate with a hypothetical example:
Suppose gammons are no longer possible and that you
currently have 68% cubeless winning chances. (cubeless equity = +.3)
After the next 2 roll sequence you will be at 88% cubeless (cubeless
equity = +.76) with probability 40%, 53% cubeless with probability
40% and 58% with
probability 20%.
Suppose you consider your cashpoint to be 78%.
Then mlf = .40 ( +.76 - +.56 ] / 1296.
I think this is better for two reasons
1) When deciding to double the key volatility considerations are
how often will I lose my market and by how much will I lose the
market. mlf combines these two concepts succinctly (and does so in
the proper way, but that would take a very long post to explain
fully.)
Standard dev. adds to volatility when it shouldn't. Suppose in my
example that the 40% of time get 88% win chance is changed to
20% of time get 83% and 20% of time get 93%. This should not make
doubling more attractive and indeed mlf stays constant. Standard dev
increases however.
2) It doesn't matter to the computers but mlf is reasonable to attempt
to estimate over the board. Std dev is much omre difficult.
,Bob Koca
bobk on FIBS
Chuck Bower
>>
>> As an aside, another article of that same time period also discussed
>> the same topic--"On Optimal Doubling in Backgammon" by Norman Zadeh and
>> Gary Kobliska in MANAGEMENT SCIENCE, vol 23 #8, p. 853, April 1977. Both
>> articles (which deal only with NON-GAMMON games) are still valid today,
>> IMHO. That is more than can be said of much of the literature of that
>> time period.
>>
>
>Zadeh & Kobliska's article is still correct today GIVEN THE CONTINUOUS GAME
>HYPOTHESIS. In my view, this makes it still incorrect today.
(snip)
I agree that the "continuous game hypothesis" is inaccurate for real
backgammon games. BUT, Zadeh and Kobliska also agree! The second part of
their article deals with the non-continuous race, if I read correctly.
Chuck Bower
bo...@bigbang.astro.indiana.edu
Bob Koca
>Chuck Bower wrote:
>>
>> As an aside, another article of that same time period also discussed
>> the same topic--"On Optimal Doubling in Backgammon" by Norman Zadeh and
>> Gary Kobliska in MANAGEMENT SCIENCE, vol 23 #8, p. 853, April 1977. Both
>> articles (which deal only with NON-GAMMON games) are still valid today,
>> IMHO. That is more than can be said of much of the literature of that
>> time period.
Was MANAGEMENT SCIENCE an appropriate journal for that article?
It seems to me that a journal dealing with game theory or a more
theoretically based operations research journal would have been
appropriate. I doubt if it would have been accepted in such a journal
though, the article has several unsound mathematical
statements/practices.
Bill Weinman
I am looking for a copy of Backgammon by Magriel. Does anyone know
where I can find one? Or have one for sale?
--Bill
1911@primenet.com Mr 1911
I have two copies of Backgammon by Paul Magriel...Email me or post if you
are interested in one of them.. thank you..You may find me on fibs under
the name of seinfeld if you are a user..
The Marksman