[PREV - SUPERFORECASTING]    [TOP]

SUPERFORECASTING_ISSUES

                                                     November 30, 2015

"Superforecasting" (2014) by Tetlock and Gardner

This is an extremely important and
profoundly annoying book.

As I mentioned: the primary thing that bugs
me about it is that the debiasing training
it mentions is pushed off to an Appendix           SUPERFORECASTING
without much discussion.

The second thing that's
tremendously annoying about this
book is that it's written like a      It's tremendously condescending
trashy business-advice book, full     in places (don't worry your
of folksy anecdotes, and personal     pretty little head about *hard*
descriptions of some of Tetlock's     stuff, like *equations*, it turns
"superforecasters".                   out you don't need 'em!  Yay!).

                                      Even the phrase "superforecaster" is
                                      more than a little annoying, what
                                      Tetlock's actually talking about is
                                      something more modest like
                                      "not-quite-as-stupid forecasters".


The third annoying thing, and perhaps one
of the biggest in terms of how grating it          A possibly related
is, is that throughout Tetlock keeps invoking      annoyance has to with
the names of pundits like Thomas Friedman          Iraq war/Osama bin Laden
and Niall Fergusson as though anyone with          intelligence, where
a brain would take them seriously.  Tetlock's      again Tetlock has a
dream is apparently to challenge the big           little too much resepct
name pundits to commit to making precise,          for the Very Serious.
testable predictions without any weasel-room
in them, so we can at last see for sure who           Tetlock accepts the
knows what they're talking about.                     testimony of a single
                                                      insider that the Iraq
But it's *not* as difficult as Tetlock                War intelligence was a
makes out to examine their track records,             sincere mistake.
it's been done, e.g FAIR's analysis of
Friedman's tendency to predict things                 Tetlock goes with
"within six months", which is now known               establishment sources
as a Friedman Unit in some circles.                   showing Obama agnoizing
                                                      over the location of
Notably *it doesn't do any good*.                     Osama bin Laden
No one cares that Tommy is demonstrably               (Seymour Hersch paints
an idiot, he remains Tommy.                           a different picture).
                                                                          
Here's my prediction: if you got             Concerning macroecon, Tetlock
Niall Fergusson to commit to a               pilously asserts that he is not
definite prediction, being proved            competent to express an opinion.
definitely wrong would not stop              In other fields, he is not so
him from walking it back with a              cautious.                    
very impressive looking article
in the Financial Times full of
quasi-bogus stats and figures
that wouldn't hold up given a
moment's thought.

Being wrong just gives these guys
something else to write about.            But then, we do seem to have a deep
It doesn't undermine their status.        abiding fascination with anything
                                          that looks like Sports: getting
                                          pundits to duel each other might
                                          be a successful way of getting
                                          people to focus on Who Won.




--------
[NEXT - QUESTIONING_DEBIASING]