[PREV - OPTIMAL_MIST]    [TOP]

BERNOULLI_DEFECT

                                             January 28, 2014

Dan Gilbert, Ted talk, "Why we make bad decisions".

Dan Gilbert is a true believer in                        SPINOZA_CYCLE
risk-neutral decision-making:
multiply your best estimates of the         This idea is
stakes times the odds, and use that         attributed to
to compare your alternatives.               17th Century
                                            work by
                                            Bernoulli.

                                                I'm more or less a believer
                                                myself, but Daniel Kahneman's
                                                work makes me wonder if this
                                                is an over-simplification...

Gilbert goes off into a silly tangent
(it's silly because it's unnecessary
to get his point across): do you see
more dogs or pigs on leash at Oxford?

    He then purports to explain why we said
    dogs, but his explanation is wrong in
    various details in my case (though perhaps
    not his immediate audiences' cases): he
    says you quickly review in your mind cases        This is a *very* quick
    where you've seen dogs on leashes and pigs        "review process" I think,
    on leashes, it's easy to remember dogs,           so quick that it's
    not so easy to remember pigs, and so you          debateable whether it
    went with dogs.                                   really happens, or if
                                                      there's something else
    I myself guessed dogs, but I've never in fact     that goes on that
    been to Oxford, and I could imagine various       simulates it.
    tricks he might've been about to play on us
    with this question: could there be exceptional
    circumstances, like an annual pig convention
    held at Oxford where there are thousands of
    pigs on leashes, maybe enough to throw off the
    average value for an entire year?

    So myself, I felt like the reason I tentatively
    went with dogs has more to do with stereotypes
    of life in english-speaking urban communites
    (and I'm not conscious of doing any review
    process where I counted up instances of seeing
    dogs on leashes).

    In any case, that's supposed to be an example
    of this review process working.


Gilbert then goes on to a word game where I presume he
has actual data-- so that when he tells all of "you"
how you think, he's at least referring to a statistical
average.  He once again, does not actually get "me"
when he says "you":

  The question was are there more four letter words
  that fit one pattern or the other:

        "_ _ R _"   or   "R _ _ _"

  since it's easier to come up with examples of the
  second, apparently most people answer the second,
  though the first is correct ("bare", and so on...
  though the first example I got was "mere").


In my case, I could see the difficulty he was
getting at, and essentially I was unwilling to
commit, it wasn't clear to me how to really       When dealing with psych
resolve the issue quickly, it could be one or     boys, you've got to constantly
the other.                                        expect them to be playing
                                                  little manipulative tricks.



It generally isn't considered polite to quibble about
little details like this during a presentation. After
all, you can see what he's talking about: people have
trouble estimating probabilities, the things that stand
out in your mind aren't necessarily more probable, and
so on.

The reason I've inclined to quibble (outside of 
my knee-jerk contrarianism at being confronted
with that "you"): guys like this spend their lives
digging into quibbles like this, they're supposed
to be experts in how we think, and the fact that
they keep telling me things that don't actually
match what I'm thinking does not inspire
confidence.
                                                         THE_GREAT_KAHN_KNOWS
I think they've convinced themselves that
overstatement is the only way to make a point
because we suck at dealing with qualified
statements, but overstatement does not actually
impress me that much.





--------
[NEXT - THE_GREAT_KAHN_KNOWS]