Fractals, Chaos and Non-linear Dynamics

[aricho]

in the movie "the bank" a mathematical genius predicts the exact movements of the sharemarket after years of research and attempts. he uses Fractal geometry, chaos theory, non-linear dynamics and of special interest to him was the work of mandelbrot and his work regarding fractals.

He says, "if we can predict $$\infty$$, then predicting the stock market is easy."

Firstly, is this possible?

Secondly, To be able to predict the exact movements of the market in a defined 30 minute space for example, before it happens, what inputs would be needed? (like previous price data?, economic probability?) how would it be applied?

Thirldy, where is the best place to learn about this mathematics?

I am immensley interested in this and would appreciate any help,
Thanks

 

[Zurtex]

Well, correct me if I am wrong, but stock markets depend on what a lot of experts think. But due to the Heisenberg's Uncertainty Principle we would not be able to fully measure such a thing.

We could probably get a lot of approximations of the stock market with varying degrees of accuracy my simplifying the whole thing, but beyond that I don’t see what can be done.

 

[aricho]

hmm

well excuse my language but this actually is a correct term for traders. a "big swinging dick" is a person who is famous for their trading, and are so closley watched, every thing they do is copied and therefore has an effect on the market. the main forces on the market are supply and demand.

Keeping this in mind, and also "if we can predict $$\infty$$" do you think it's possible?

 

[neurocomp2003]

any school with classes in Neural nets(optional)/ stochastics / markov models and operations reserach shouhld give you the materials you need. Be sure to learn programming as well since this will be your application

...as for inputs like weather and disease you input the prior few years depending on the range of research your doing. mmm i think it was the measles was done at a period of 60 years input

 

[saltydog]

Thirldy, where is the best place to learn about this mathematics?

I am immensley interested in this and would appreciate any help,
Thanks

Start simple: the quintessential reference, "Chaos and Fractals" by Peitgen.

 

[aricho]

...as for inputs like weather and disease you input the prior few years depending on the range of research your doing. mmm i think it was the measles was done at a period of 60 years input

ok, so i imput the past into the magical black box formulae to predict the future?
would anything else be needed? like probability ect...?

 

[Zurtex]

well excuse my language but this actually is a correct term for traders. a "big swinging dick" is a person who is famous for their trading, and are so closley watched, every thing they do is copied and therefore has an effect on the market. the main forces on the market are supply and demand.

Keeping this in mind, and also "if we can predict $$\infty$$" do you think it's possible?

I don't even know what you mean by the phrase "if we can predict infinity", it seems nonsensical to me, I deal with infinity day in and day out in maths classes, I don't quite get what that means.

But yes, these people are very good at "approximating the market", if they knew things exactly then they'd probably be a lot richer...

 

[matt grime]

it sounds lke nonsense (predicitng infinity) and is just film making dressing thigns up. dynamical systems and stochastic processes are used to study the markets (what do you think quant analysts do?). there is a limit to any model's accuracy since the mathematics is only an approximation (you realize that the boom of the 80s when people got disgustingly rich was a mathematical one - a bunch of mathematically trianed people figured that they could do a better job of predicting the markets than traders operating on hunches). Heck, you can even do degrees with financial mathematics in them. but to absolutely and completely predict anything with 100% accuracy that is that complicated we know to be impossible, and that isn't even allowing for "human factors" such as a terrorist attack.

 

[saltydog]

You know what, how fitting for such a thread: We got, I don't know, 5 or 6 different models to predict the path of hurricane Dennis and they're all different! Way to go guys . . . I'm watching two butterflies outside of my window right now fluttering aroung the azelaes . . . a storm is brewing . .

 

[Curious3141]

Well, correct me if I am wrong, but stock markets depend on what a lot of experts think. But due to the Heisenberg's Uncertainty Principle we would not be able to fully measure such a thing.

We could probably get a lot of approximations of the stock market with varying degrees of accuracy my simplifying the whole thing, but beyond that I don’t see what can be done.

Heisenberg's Uncertainty has little to do with stock market fluctuations, or any other deterministic classical chaotic system. It pertains only to quantum indeterminacy.

 

[Zurtex]

Heisenberg's Uncertainty has little to do with stock market fluctuations, or any other deterministic classical chaotic system. It pertains only to quantum indeterminacy.

What I meant is that you could not measure what somebody was actually thinking because that would alter it. Surely stock market predictions of 100% accuracy would require to actually know what everyone is thinking who would affect the market?

Or am I in fact over thinking the situation?

 

[aricho]

he says stuff like, "with this mathematics, things that we never though possible to predict now can be. e.g. cloud formations" also, he is trying to predict "sentiment" in a way.

"all the current methods of prediction do the same thing, try to predict what one person is going to do, but trying to predict what 100 people will do is much much eaiser"

He uses fuzzy logic, Mandelbrot fractals, chaos theory, fuzzy logic and nonlinear dynamics.

Is it possible to use these things to predict the markets?

 

[Zurtex]

he says stuff like, "with this mathematics, things that we never though possible to predict now can be. e.g. cloud formations" also, he is trying to predict "sentiment" in a way.

"all the current methods of prediction do the same thing, try to predict what one person is going to do, but trying to predict what 100 people will do is much much eaiser"

He uses fuzzy logic, Mandelbrot fractals, chaos theory, fuzzy logic and nonlinear dynamics.

Is it possible to use these things to predict the markets?

He uses fuzzy logic twice then?

That is nonsense, what you are able to do is 'approximate' what things are going to do. When you have more data (e.g more people) it means that you take an average that will generally follow a trend, however it means you have to take out the possibility of extreme circumstances.

 

[Curious3141]

What I meant is that you could not measure what somebody was actually thinking because that would alter it.

But what does even a single person's thought have to do with Heisenberg's UP ? Quantum effects are important in the state of individual electrons, less important in the behaviour of a single neurone, and surely completely annulled by averaging out in the behaviour of an entire forebrain. At least that's what I would assume, but it doesn't seem unreasonable.

Surely stock market predictions of 100% accuracy would require to actually know what everyone is thinking who would affect the market?

The individual may be a difficult quantity to predict, but the herd is predictable, for the most part.

Or am I in fact over thinking the situation?

Maybe just a little.

 

[matt grime]

can we please just dismiss thi sfilm as being a complete load of &$^£*&$&&£ (use your preferred expletive).