11

I saw an article in which the writer had mentioned that he used chaos theory to predict stock prices and ended up with a profit over 30%. Chaos theory is basically about finding patterns called fractals in the data. It is used in various fields such as weather prediction and stock price determination.

How we can forecast stock prices using chaos theory? It would be great if someone can provide me with some examples.

Tal Fishman
  • 13,449
  • 7
  • 63
  • 110
Egalitarian
  • 219
  • 2
  • 3
  • +1 for the question. I would be interested to know also how exactly chaos theory can be used, or was used by that user in particular case. I suspect that if there is something it might be just academic speculation. – mpiktas Jul 20 '11 at 10:05
  • 3
    It should be noted that chaos theory is a catchy name, which covers applications of usual mathematical methods, so it might be that chaos theory is already applied, but under different name. Furthermore since the claim of 30% profits should be taken with the grain of salt, the same should be applied for the claim that chaos theory was used. People like catchy names for naming their work. – mpiktas Jul 20 '11 at 10:10
  • 4
    This is not a good question at all. Someone who asks about using chaos theory in finance clearly doesn't understand chaos theory or finance. This reminds me a lot of this question on machine learning. I'm going to echo Shane's advice: Learn statistics first. Then learn backtesting. Any concepts stated here are going to be useless without a solid foundation in quant trading first. – chrisaycock Jul 20 '11 at 14:49

2 Answers2

7

Benoit Mandelbrot, pioneer of fractals theory, has a set of papers on the Multifractal Model of Asset Returns. I believe when most people in finance are saying they use "chaos theory", they are really referring to fractals, and particularly multifractals. I am skeptical as to whether it can be useful to predict stock prices or even to predict volatility, but searching down this line of research (see, e.g. here and here) could lead you to your answer.

Tal Fishman
  • 13,449
  • 7
  • 63
  • 110
5

I think you got something wrong here: Chaos theory states that even when the system is deterministic but chaotic it is unpredictable!

This is the very definition of a chaotic system. So it is rather paradoxical to ask how to use the statement that something is unpredictable to predict it...

See here for more details: http://en.wikipedia.org/wiki/Chaos_theory

vonjd
  • 27,437
  • 11
  • 102
  • 165
  • 1
    Chaos theory imposes some structure on the processes. If there is structure you can exploit it. Unit-root processes are also can be considered unpredictable, but all stochastic calculus is based on such processes. So I would not be so fast in dismissing the notion that chaos theory cannot be used for forecasting of stock prices. However the claim in the question can be dismissed :) Real life example or some papers would be interesting to read. – mpiktas Jul 20 '11 at 10:03
  • 2
    @mpiktas Just because there's structure doesn't mean the structure is exploitable. – chrisaycock Jul 20 '11 at 14:50
  • 1
    Isn't wheather chaotic and its forecast better than random? – DaveBall Jan 01 '12 at 21:24