free4now
Thinks s/he gets paid by the post
- Joined
- Dec 28, 2005
- Messages
- 1,228
Interesting idea about risk, that feels very apropos today:
http://www.nytimes.com/2009/03/10/science/10quant.html?pagewanted=1&hp
I wonder if incorporating this idea into monte carlo simulations would improve their accuracy.
Original article on "quants":Dr. Taleb has waged war against one element of modern economics in particular: the assumption that price fluctuations follow the familiar bell curve that describes, say, IQ scores or heights in a population, with a mean change and increasingly rare chances of larger or smaller ones, according to so-called Gaussian statistics named for the German mathematician Friedrich Gauss.
But many systems in nature, and finance, appear to be better described by the fractal statistics popularized by Benoit Mandelbrot of IBM, which look the same at every scale. An example is the 80-20 rule that 20 percent of the people do 80 percent of the work, or have 80 percent of the money. Within the blessed 20 percent the same rule applies, and so on. As a result the odds of game-changing outliers like Bill Gates’s fortune or a Black Monday are actually much greater than the quant models predict, rendering quants useless or even dangerous, Dr. Taleb said.
http://www.nytimes.com/2009/03/10/science/10quant.html?pagewanted=1&hp
I wonder if incorporating this idea into monte carlo simulations would improve their accuracy.