Page 1 of 4
Multi-fractal stuff, Mandelbrot is mostly right
Posted: Sat Mar 18, 2006 12:23 am
by crowlogic
I know for sure that the anti-Mandelbrot crowd is full of shit now. Markets are fractal at some time scales. Here's a 5-minute price change forecast using some methods inspried by fractals. This is out-of-sample data, and the results are robust from any starting point.
[img]/User%20Files/1210/priceForecast.png[/img]
Multi-fractal stuff, Mandelbrot is mostly right
Posted: Sat Mar 18, 2006 12:31 am
by chiral3
I would agree that markets are fractal at most time scales. However, how many tries did it take you to get your line to follow the price process? So what's the method?
Multi-fractal stuff, Mandelbrot is mostly right
Posted: Sat Mar 18, 2006 12:50 am
by crowlogic
good question. I've been working on this for a few years now so...I've tried thousands of things in that time. I don't want to give it away, but you need a creative filter bank.
Multi-fractal stuff, Mandelbrot is mostly right
Posted: Sat Mar 18, 2006 12:59 am
by chiral3
Can you do me a favor? At 1000EST on Monday can you post the next five minutes for QQQQ?
Multi-fractal stuff, Mandelbrot is mostly right
Posted: Sat Mar 18, 2006 2:54 am
by apine
how about high noon? i'm joking, but i wanna see this too.
Multi-fractal stuff, Mandelbrot is mostly right
Posted: Sat Mar 18, 2006 6:10 am
by crowlogic
[i]Can you do me a favor? At 1000EST on Monday can you post the next five minutes for QQQQ?[/i]
I'll try.. I normally am working my 'real job' during those times, and also, I need to port this stuff over to my realtime trading platform from matlab, hopefully I can get it done this weekend.
Multi-fractal stuff, Mandelbrot is mostly right
Posted: Sat Mar 18, 2006 7:08 am
by FDAXHunter
I don't think there are many people that question that markets are fractal. I certainly don't know any. I think most people quite readily admit that there is self-similarity at many levels of resolution. So who is this "anti-Mandelbrot" (I'm anti-Mandelbrot, but not because I don't believe in self-similarity and the emergence of fractal structures from chaotic processes, but because I think he's a dick, behavior-wise).
Multi-fractal stuff, Mandelbrot is mostly right
Posted: Sat Mar 18, 2006 9:22 am
by crowlogic
Well, some people still believe the EMH and all that crap, my old roomate was a finance major and was constantly telling me I'm insane because the CAPM model is true, etc, etc. I gave him one of Taleb's books but he didnt bother reading it.
I've no idea about him personally, I'm sure he's a complete lunatic, most genius-types are, but, I meant anti-mandelbrot as far as people still assuming normal distributions, complete random walk behaviour, etc.
Multi-fractal stuff, Mandelbrot is mostly right
Posted: Sat Mar 18, 2006 9:29 am
by Nonius
ÿÿ
not as much of a lunatic as Rene Thom.
on second thouht, mandelbrot is more of a lunatic.
thom's quote, "If one must choose between rigour and meaning, I shall unhesitatingly choose the latter" pokes a hole in my conjecture that the phrench are just into rigour.
Multi-fractal stuff, Mandelbrot is mostly right
Posted: Sat Mar 18, 2006 11:31 am
by crowlogic
Not familiar with Thom.. I read a bit, his big thing was catasophre theory it seems. Doesn't seem very different from chaos theory. Some writeups in his obituary says that neither theories have given us much useful stuff in the form of predictible models.. are they crazy or am I? There are plenty of methods that can predict chaotic processes pretty well in the short term.