Markets are efficient if and only if P = NP

[coderman]

On Tue, Feb 23, 2010 at 5:27 AM, Joseph Ashwood <ashwood at msn.com> wrote:
> ... the inability of a massive multiprocessing
> system to build proper market efficiency indicates that no one currently has
> a solution to prove P=NP.

correlation != causation

also, the strong-form efficiency present in such an analysis fails in
the face of biased and flawed real-world market actors. cognitive and
informational biases render such approaches masturbatory at best.


> There is also far more money in economic research
> than in algorithmic research...

leveraging economic research interest to fund algorithmic research is
a clever approach to funding perhaps.. :)


> ... Maybe Maymin should apply the same processes and become a hedge fund
> manager, they make a lot more money than mere engineers.

until they blow up*. the quant meltdown brought such computational
hubris back to earth...

(Renaissance Institutional Equities Fund was down 6% at the start of
this year. and 2009 should have been a good year for a hedge!)

* nassim talib == the man
he, by the way, recognized inherent market instability as an
opportunity. the Universa fund he advises made returns of 65% to 115%
in oct 2008 via his "black swan protection protocol.