The aim of reading a good book should not be to learn to play like the author,
but to open your eyes to concepts that they express that you had not
experienced yourself.  For instance, timing is a concept that a beginner
might find extremely difficult to deduce by themselves; through trial and
error they might notice that in some types of games, their own position
tends to deteriorate faster than their opponents.  However it's something that
a book can provide a great deal of insight into, just by giving the reader
another point of view to the same properties they had already noticed
informally in their own games, or perhaps previously been unaware of but
now able to look for.  Even if the conclusion the author makes is blatantly
wrong ("when you find yourself short of timing, dump extra chequers onto your
1 point to avoid leaving blots"), the background is valuable.

I guess my opinion then is to steal as many ideas as you possibly can about
the game from books; read them critically whether they're Magriel or Becker;
form your own opinions, and never stop thinking.  Feel free to disagree with
the author -- whether you just learnt to play this morning, or have beaten
Jellyfish in a 21 point match every day before breakfast for the past year.
Try playing the way the author suggests; try any reasonable alternative you
can think of -- see which you prefer.

Perhaps most importantly, don't be afraid of being "wrong".  There are so
many ideas in backgammon that it's impossible to be "right" about all of them
anyway.  But I guarantee that if you thought about something as a beginner,
make an elementary mistake and then eventually learn you were wrong all along
and then start conforming to generally accepted expert opinion, then you'll
have learnt a lot more than somebody who just accepted the expert opinion in
the first place without challenging it.  Don't ever be embarrassed or afraid
to change your mind.  I'm no expert and have only been posting here for a
few months, but I bet I could already find a hell of a lot to argue with
against what I've written before.  That doesn't worry me one bit (nor will it
make me shut up I'm afraid, you have to killfile me for that :-)  And I'm
not the first to change my mind, even the experts have been changing their
minds about backgammon for thousands of years.  Since the dawn of history,
players have preferred to split with an opening 21 for instance -- until the
modern play of the 1970s started favouring slotting the 5 point for a
stronger offence.  And then after seeing neural nets play differently and
beat them, experts have changed their minds again and decided that gee,
maybe splitting was right after all, these damn computers have made us look
a bit silly.  But what matters isn't what play is currently fashionable
(or even "right", since nobody can say for sure) -- the fact is that we now
know more about the strengths and weaknesses of splitting and slotting than
anybody ever has before.

"Backgammon In A Week" is a reasonable beginners guide that explains the rules
and doesn't pretend to touch advanced concepts -- certainly not something a
serious player would take a huge amount of interest in once they are familiar
with the game, not that there's anything wrong with it.

"Teach Yourself Backgammon" is a somewhat deeper text that so far I as
remember (I don't own a copy myself) is a 70s-style strategy guide -- it
could be quite valuable to a beginner/intermediate but be aware that many
of the ideas have progressed considerably since the discussion.  It has
a reference table of opening moves and replies for instance that is
somewhat different than I expect most top players would make today.

"Backgammon: Winning Strategies" in my opinion is by far the strongest of
the three.  It is pretty much current (frequently referencing Jellyfish 2.1
evaluations and rollouts and Scoyners database equities) and generally
presents reasonable alternatives modern players might consider and gives
a fair comparison of them.  It touches many concepts without going into a
great deal of depth about them which is probably reasonable for a book of
this size (244 paperback pages) -- for instance it presents a match equity
table derived from Janowski's formula and analyses how it should affect
match play, without justifying the background of the table or formula
itself.  It makes a good effort to address the discrepencies between
generally accepted expert and computer preferences ("...Jellyfish 2.1
prefers this move but many players dislike placing a chequer on the 1 point
so early in the game...", "surprisingly Jellyfish thought that hitting was
the best move, but when it rolled out the position 1,296 times it
established that move D was superior.")  It goes into reasonable detail
about many tactical situations (inspecting various possible back game
positions and explaining why the 5 point is weak bearing in against a 1,3
back game; walking a prime backwards to pick up an extra chequer, etc.)  My
main criticism (not that I'm claiming that my opinion is in any way better
than the book -- just as an example of how a text should be critically
considered rather than accepted at face value) is that his treatment of
cube handling is weak.  Clay claims that cube decisions should be made on
three criteria: race, position, and threat -- personally I think that's
horribly overgeneralised: you can't base decisions on just 3 factors, and
even if you could, they would be threat, threat, and threat.  But never
mind.  I don't think he mentions volatility at all, oversimplifying to the
point of being wrong: "the ideal lead at which to double is 70 per cent..."
(admittedly he qualifies the assertion as dependent on the match score and
the claim appears restricted to races, but even so presenting a number like
that without justification ought to raise questions in the mind of the
reader).  The quiz section of the book (100 questions) is well done --
again I think it would be a shame if a reader was overly concerned about
choosing the "right" answer in the quiz, but thinking about a position,
selecting a move, and then reading the comments about the alternatives
could be valuable if Clay mentions factors you hadn't considered, or seen
in the same way.  Overall I think most intermediate players could learn a
lot from it, though advanced readers would be frustrated at the lack of
detail in many areas.