"Whether you can observe a thing or not depends on the theory which you
use. It is the theory which decides what can be observed."
http://en.wikiquote.org/wiki/Albert_Einstein
It is incredibly naive to say that
physical scientists stole all of the easy problems. It is probably more
meaningful to say that weaker physical scientists stole all the low
hanging fruit. But then there is a clearly defined hierarchy of problems
in the physical sciences when it comes to what is important, and that
is why Einstein is revered but many others are not.
We should not be surprised at all if Nate Silver misses the
results at times. After all, the domains he deals with are probably not
governed by laws even in a statistical sense (that does not mean
statistical analysis has no role to play).
In the old days when I took my econometrics course, OLS and
all that stuff was disposed off rather quickly and we were told to
concentrate on building a structural model of the problem domain, study
it, derive from it the reduced form model and so on. I think those days
are gone, and these days most (not all) feel quite satisfied with a
single equation model. The concept of a rich model to represent
mathematically some aspect of nature, the hallmark of physical sciences
borrowed in to the social sciences, has just abut disappeared. In the
physical sciences, usually you start with a system of equations
(differential/difference) and ultimately come up with a solution that is
very easy to comprehend. In accounting we start with an equation and
end up with a cartload of statistical mumbo jumbo.
In the physical sciences there is a clear distinction made
between those who model and those who test them. Theoretical physicists
work with mathematical models, but experimental physicists just verify
them in nature. Nevertheless, the two collaborate in rather peculiar
ways so that the science can advance (I have given instances of this in
earlier postings). This enables theoreticians to hypothesize things that
have not even been shown to exist. Bosons were conjectured by Satyen
Bose with Einstein in the 1920s, but experiments to confirm their
existence is now becoming possible 40 years after the death of Bose.
In accounting, on the other hand, we all pretend to do both
with the result that we have perverse incentives to claim advances in
both. We do not recognize that the skills needed for theoretical and
empirical/experimental research are very different and rarely does one
come across a person who can do justice to both.
Most sciences see this dichotomy between theoretical and
experimental/empirical aspects of research but collaborate in ways that
help the domain. When I switched to Computing towards the tail end of my
career I found the way these two collaborate absolutely fascinating. I
wish there was more of that in accounting.
Incidentally Keynes was a mathematical tripos student, and
graduated twelfth wrangler at Cambridge. His thesis was in mathematics
(Treatise on Probability, a classic). )Unfortunately, he lacked the
genius in mathematics and so chose Economics as told by the economist
Harrod in his biography of Keynes, "
... logical faculty, his accuracy and his lightning speed of thought
made him a thoroughly competent mathematician. He had no specific genius
for mathematics; he had to take pains with his work; ... he did not
seek out those abstruse regions which are a joy to the heart of the
professional mathematician.". (See http://www-groups.dcs.st-and.ac.uk/history/Biographies/Keynes.html).It
is likely that, Keynes said those words. On the other
hand, I think it was Einstein who said something to the effect that he
opted out of economics because it was too difficult. It is possible that
Einstein found the verbosity of the then German economists tiring.