5
$\begingroup$

Yesterday, I was going through an article in which the user had mentioned that he has used chaos theory to predict stock prices and ended up with 30% + profit.(I am not intersted in the profits :P) After that I read a bit about chaos theory and found out that its basically finding patterns called fractals in the data available. After that I came to know it is used in various fields as weather prediction and stock prices determination.

I just wanted to know how if anyone has an idea on how we can do stock forecasting using chaos theory.

Like, it will be great if someone can provide me with some example w.r.t chaos theory.

  • 3
    I traded stocks professionally for many years. We used a lot of math (PDEs, statistics, linear optimization), but we never used anything like chaos theory. It's an interesting topic, but I don't think you find a lot of money in it. There are books out there on chaos theory as applied to the financial markets, but they don't actually explain how to make money. Books never do! If you had a secret formula, would you write a book about it?2011-07-20
  • 0
    You can't use chaos theory to predict stock prices. The fundamental tenet of chaos theory is 'prediction is hard, if not impossible'. To say that a system is chaotic means that there is a *deterministic* rule for its future evolution, but that small differences in initial conditions grow exponentially quickly over time - meaning that a tiny error in estimating the initial conditions renders your predictions useless sufficiently far in the future. So there are two questions: (i) do stock prices follow a deterministic rule (they don't) and only then (ii) if so, do they display chaos?2011-07-20
  • 2
    P.S. If you *really* want to test a series of stock market returns to see if they're chaotic, then the modified Kolmogorov 0-1 test may be useful, see [this paper](http://arxiv.org/abs/0906.1418). P.P.S. I think your keyboard may be broken, as it is automatically appending a string of periods (..........) to all of your posts.2011-07-20
  • 0
    @Chris: wouldn't you clarify why prices do not follow the deterministic rule?2011-07-20
  • 2
    If prices followed a deterministic rule then whenever you observe the system state to be $x_i$ at time $i$, you would expect to see a very similar state $x_{i+1}$ at the next time. Accounting for measurement error, the difference $||x_{i+1}-x_i||$ will either be small and roughly constant (if the system is deterministic and regular), exponentially increasing with time (if it's chaotic) or randomly distributed (if it's stochastic), see [this paper](http://www.santafe.edu/media/workingpapers/91-07-029.pdf). I admit that it's possible that the 'state' of the financial system ...2011-07-20
  • 3
    ... follows some incredibly convoluted deterministic rule dependent on many factors other than prices, but I don't think that's a particularly helpful model as you'd never be able to extract such a rule from the data. Instead, a stochastic modelling approach is the way forward.2011-07-20

3 Answers 3