Sure, then go ahead and point to a SINGLE source that proves that NN are successfully employed in financial trading. There are countless strategy ideas and approaches in no-NN public domain that work one way or the other and with specific assumptions, so surely there must be information available in public domain where a researcher, academician, or practitioner published NN approaches to financial trading that work. There are none, not one single paper, to my knowledge that proves that "training" a strategy algorithm produces sustainable risk-adjusted positive returns after application of costs of execution that warrant an adoption of such process. Not a single one. Please disprove me or anyone whom you seem to so much disagree with. I concur with you that NN have their applications and usefulness in many areas...just not in financial trading. Weights? Can you be very specific which weights you are talking about?
"Can you point to a SINGLE source that proves there are jam donuts orbiting Jupiter?" This is the essence of the Ignorance Fallacy. Now you're asking for a free lunch.
just what I thought. You are a troll, at least keep your BS to yourself and stop to confuse and lure unassuming beginners. PI, lol, at least you duped numerous other lemmings on this thread which is in a way hilarious.
Interesting as always. I thought people might find looking at some irrational charts inspiring. My aim is to provide value. In this file we see 1000 of your posts. I'm sure value can be found in there, but i'd bet there's a good deal more loitering under a bridge.
well, as you are well aware, even your weird text file is "somewhere" encoded in one particular PI digit sequence. Lol, chuckle, yes, you do add value my friend, not. Another idiot on ignore. The programming sub-section on this site is getting cleaner by each day.
Looking back on the posting of volpunter and I, we can see how just a small amount of data can be used to create a desired effect. Data and data analysis is the number one utility in the toolbox of todays traders. A hybrid, human/machine approach is like using binoculars to observe a distant object. The markets are chaotic, and yet with the application of a mechanised microscope, they become more orderly, and predictable.
It does not matter. PI is irrational and infinite sequence of digits. Up to which sub-sequence embedded in pi you use, theoretically you can get any chart you like. Like https://en.m.wikipedia.org/wiki/Multiverse, everything exists in pi, big bubble, big crash, every chart of stocks that ever existed, and everything you can think of.
Are you saying that there are zero traders on here making money daytrading the market? You sound like you've got a chip on your shoulder
Did you ever read any of my posts carefully. If not then you can simply ignore my posts. No I don't have any chip on my shoulder. I simply stated that there is no reported case in existence of someone who managed to employ a neural network to developing and running profitable trading strategies in production. Not one. There are uncountable public occurrences and proof of working and profitable implementations of trading strategies perusing other technology. That is not proof enough? It's called empirical evidence. Unless someone is naive he or she should believe in summary statistics and put faith in numbers. The dumb and outright retarded claim that "if something does not work for me it does not mean it works for nobody else" is simply bizarre and usually comes from naive people who pretend to know stuff they don't. I may come across as an assholes and maybe I am. But I posted uncountable content that gave away code and implementation details of technology that assists in creating and running a fully systematized trading architecture. I simply do not respect those who have never added valuable content on the technology side and suddenly come out of the closet with unrelated PI content and claim they have done what nobody else in the professional HFT and quant trading space ever achieved. It is simply ludicrous.