INTUITION VERSUS ALGORITHM 553
His book The Signal and the Noise is a tribute to the triumph of
algorithm over intuition.^8
It would be a mistake, however, to dismiss intuitive expert
judgment generally, just because it fails at predicting the results
of presidential elections. Not all expert opinion based upon
experience can be reduced to “divination.” Through repeated
experience, people develop expertise of all kinds, ranging from
chess playing^9 to medical diagnosis.^10 No one accuses the best of
such people of using a divining rod simply because they do not
rely on computer algorithms. Moreover, we make judgments all
the time about what is likely to happen next, including, for
example, the judgment that it is safe to cross the street when the
traffic signal favors us and the cars are all stopped. Most of the
time, there is no computer algorithm with which we can
compare our rate of success, but we have a good sense—
confirmed by repeated experience—that we are making the right
decision.
On the other hand, the use of algorithms seems to neutralize
some obvious biasing factors that plague the pundits routinely.
Why is it that experts paid by Fox News (a Republican-oriented
network) predicted a Romney victory, whereas those paid by
MSNBC (a Democrat-oriented network) predicted that Obama
would win? One possibility is that most of the pundits are
sufficiently corrupt to misstate their actual views if they are paid
enough to do so. More likely, though, their prior commitments
contribute to what information they regard as significant and
color their analyses, which are sincere. This phenomenon, called
confirmation bias, is well studied by psychologists. It is an
november-7-2012/nate-silver.
(^8) See generally NATE SILVER, THE SIGNAL AND THE NOISE (2012)
(investigating how statisticians distinguish meaningful indicators in ever-
increasing amounts of data in order to make accurate predictions).
(^9) Fernand Gobet & Neil Charness, Expertise in Chess, in THE
CAMBRIDGE HANDBOOK OF EXPERTISE AND EXPERT PERFORMANCE 523, 532–
34 (K. Anders Ericsson et al. eds., 2006).
(^10) Geoff Norman et al., Expertise in Medicine and Surgery, in THE
CAMBRIDGE HANDBOOK OF EXPERTISE AND EXPERT PERFORMANCE, supra
note 9, at 339, 350.