[PREV - SAN_FRANCISCO_DREAMING]    [TOP]

WHICH_WAY

                                                 April 6, 2008
How can we continually make predictions
as though "Black Swans" don't happen?            BLACK_SWAN

That wild changes will occur is indeed           (As I was just beginning
likely, but knowing which will happen is         to read the book)
next to impossible: hence any prediction
that is to have any hope of being correct
at least in the short term has to ignore
*all* of these unlikely occurrences,
however unlikely it is that none of them
will happen.

    You won't get it right
    all the time (ever?)          This is similar to (though
    but at least you won't        not quite an application of)
    be quite so grossly           Occam's Razor: yes, the world
    wrong.                        is a complex place, but our
                                  explanations of it should be
       E.g. should                no more complex than need be to
       we believe                 cover the evidence.
       judgement
       day approaches                Similarly: yes
       because we                    the future will be
       expect the                    weird, but our
       unexpected?                   predictions of it
                                     should be no weirder
                                     than required by
                                     extrapolation of
                                     present trends.



              Betting on a miracle--
              any particular miracle--
              remains insane.

                The venture capital
                strategy, though,
                would be to bet                 Taking Taleb
                that *some* miracle             literally, his
                is going to come through.       claim would be
                                                that you should
                                                not bet that
                                                no miracle is
                                                ever going to
                                                happen.

                                                  Does that sound
                                                  useful to you?

                                                  (It must be, Taleb
                                                  *insists* that he's
                                                  a practical, pragmatic
                                                  fellow.)




    "In the end this is a trivial decision
    making rule: I am very aggressive when
    I can gain exposure to positive Black           You can't estimate
    Swans -- when a failure would be of             your exposure to a
    small moment -- and very conservative           Black Swan without
    when I am under threat from a negative          first knowing it's
    Black Swan.  I am very aggressive when          there, and *by
    an error in a model can benefit me, and         definition* these
    paranoid when the error can hurt.  This         dark swans are hard
    may not be too interesting except that          to see.
    it is exactly what other people do not
    do."       -- p. 296





--------
[NEXT - TOY_WORLDS]