"History is a wonderful thing, if only it was true"
-Tolstoy

Wednesday, February 15, 2006

Bada Bing - I agree on the concept

Interesting piece
Challanges conventional market wisdom.

"efficient market hypothesis (EMH), which holds that markets aggregate information efficiently and investors form their financial expectations rationally. The reality may be much messier."


Dr. Andrew Lo: Darwinian Investing:
The MIT finance professor's market theory borrows from neuroscience, evolution, and econometrics.

...

"Sci-fi was an important influence on Lo, whose family moved from Taiwan to Queens, N.Y., when he was 5. Raised by his mother, he became an academic star. He skipped eighth grade, sped through Bronx High School of Science and Yale University, and nabbed a PhD in economics from Harvard University at age 24. But it was Isaac Asimov's Foundation trilogy that steered him toward finance economics. Asimov sketched out a branch of mathematics called psychohistory, whose practitioners sample the proclivities of large numbers of people, then accurately predict the future based on what they learn. Sound familiar?"

...

My comments submitted to BusinessWeek:

"All very very interesting

Patterns to my thinking very well.
Markets are NOT rational, but emotional. Patterns change, players change (both the players themselves and what they pay attention to). Anybody remember M1? M2?
This is why trading was fun, as well as, mostly profitable while I was doing it. Now I’m just a long term investor.

I always viewed the markets as a large experiment in mass psychology. Trade on the basis of how you expected the market to react to news.

I have argued that “Rocket Science” does not work on the street.
Engineering models are “brittle” and prone to failure when exogenous variables are introduced. They work for a while, but eventually fail.

Asimov tie in is very interesting. It influenced my work in college (PoliSci, Econ), but very weak carryover. Flaw in the analogy is that psychohistory demanded subtle measurement of millions if not billions of individuals. That said, there are interesting areas of study in Data Mining."

No comments: