Winning With the Doubling Cube
David Montgomery
Most of the discussion of basic concepts and theory (like pip
counting, take points, doubling points, etc.) is pretty good.
The positions that he uses to illustrate the concepts are very
often misanalyzed. Apparently no one associated with the book
checked the positions with a bot, or even showed the positions
to a strong player. It's really ridiculously bad. There are
positions called too good that are beavers, and vice versa. Or
perhaps just very close to this -- my notes are at home.
The novel contributions, like the blitz count and so forth, are
too simplistic to be of use. I am all in favor of developing
models, but these models are probably worse than useless. A
strong player can do much better than the models; a weak player
might place too much faith in the models and fail to realize
the limits on their applicability. A player who believes Bell's
analyses and models and who plays props would be a gold mine.
I recommend this book as an excellent gift to give to weaker
players that you play for money.
David Montgomery
mo...@cs.umd.edu
monty on FIBS and GG
Gavin Anderson
that a lot of what you say is true, but I still think the book is useful.
When I started reading it, my first thoughts of course were 'what do the
bots have to say about this', and sure enough, when I checked many of his
positions through on Jellyfish, it came up with different answers. I was
actually intending to add another paragraph to my original post, querying
what other people thought of the book, and whether the analysis was
reliable, because I'd come up with several positions where JF didn't like
his decisions. (On reflection I really should have included that last
paragraph, but I did all the JF checking after writing the post, and
unfortunately forgot to update the post before I sent it).
So - in big warning letters to potential readers DON'T TAKE HIS ANALYSIS AT
FACE VALUE.
But - having said that, I still think there is a lot of useful information
in the book.
> Most of the discussion of basic concepts and theory (like pip
> counting, take points, doubling points, etc.) is pretty good.
I also think that his discussion of calculating winning chances is very
good - but the %'s he uses are too high. Once I got the hang of his idea,
and found that JF disagreed with his numbers, I simply modified all of the
percentages that I would apply (eg - very simple example - where Bell would
give 10% for having escaped a back man, I would only give 4%, etc). Doing
that I have found that in a large number of mid-game positions I'm managing
to estimate my winning chances to within 4% of what JF says, and gammons to
within 8% or so. That may still be hopeless, but it is way better than I
could do before, so I'm grateful to Bell's book for that at least.
So - his basic idea seems really useful, but his figures are too optimistic.
The alert reader should check them and scale them down.
>
> The positions that he uses to illustrate the concepts are very
> often misanalyzed.
This is true. And I guess it's a very damning indictment. But - taking
Bell's ideas, and modifying them as I mentioned, I'm managing to come up
with the correct cube decisions anyway (ie the one's that JF recommends,
rather than Bell - and I'm managing that BEFORE I look at JF's answers)
> The novel contributions, like the blitz count and so forth, are
> too simplistic to be of use. I am all in favor of developing
> models, but these models are probably worse than useless. A
> strong player can do much better than the models; a weak player
> might place too much faith in the models and fail to realize
> the limits on their applicability. A player who believes Bell's
> analyses and models and who plays props would be a gold mine.
Again I think that David is mostly right - following Bell's analysis blindly
would NOT be a good idea. I rather liked the models as an aid for getting a
feel for particular positions, like primes or blitzes. I would not - and I
hope alert readers wouldn't - take the counts etc as gospel. Again Bell's
numbers seem over-optimistic, and I've already scaled back the numbers he
gives as indications of when to double etc.
As David says - to strong players the models are no longer of much use, and
weak players might try to apply them blindly - but I think there is a
category of alert 'intermediates and improving' (where I put myself) who
could pick up a lot from Bell's book - and then know what to use and what to
alter. You'd need JF or Snowie to help you with that process though.
David - thank you for posting your thoughts. I'm really glad to hear what
you think, and I would say I'm mostly in agreement with you, but that I
still think there's value to be found in the book.
> I recommend this book as an excellent gift to give to weaker
> players that you play for money.
:)
Cheers,
Gavin Anderson
Sapporo, Japan
brit...@mbf.sphere.ne.jp