You are halfway correct. I did indeed trade for almost 12 years at various sell side banks and hedge fund prop desks and have applied my skill set to creating my own systematized trading architecture from the ground up. You would have seen that reflected in various posts in technical and programming forums on this site. I have traded external and my own assets now for almost 6 years independently and believe I have plenty enough experience on the quant trading side to be able to judge that neural networks are unable to capture financial asset dynamics. I have previously posted Screenshots of my quant program masters degree and I am happy to debate in depth why neural networks are not able to be successfully applied to financial trading. Let me know if interested. In the meanwhile feel free to rant what some of the brightest in the world at rentech and related quant funds are capable and not capable of doing. You are certainly incorrect in your assumptions so far at least. By the way are the following of your past posts also outcomes of your neural network ? -> "SPX > 2300 between July and August 2015. May 26, 2015" -> "When looking at your chart, it obstructed my view of price action. It's like trying to predict the weather from behind frosted glass." Not sure how you reconcile your neural network addiction with your reading tea leaves via staring at some charts. Thanks for the entertainment though.
Do you feel that the time spent programming, debugging, and forward testing a strategy (who's edge could eventually be depleted) could be better spent developing an edge and then immediately trading it by hand? Some traders would rather be a master of one trait instead of a jack of them all. I like your thoughts though. If and when you do succeed i'm sure you'll be very wealthy (if not already)
Except neural networks is simply not the way to riches. The identical same fallacy some retail dreamers commit deals with the rather rosy term "machine learning". Think for one second about the whole concept: we are talking about the possibly most competitive career and field there is on this planet. A field hundreds of millions embark on. We are talking about a field that employs some of the smartest walking human beings. And we talk about a field where there are individuals and institutions with sheer unlimited amount of resources. Yet a little wannabe self-declared wiz kid comes along , turns on his quad core cpu powered computer , kicks into gear a bit of python code and maybe throws in a dose of matlab and hey, out comes the magic parameterization for the holy grail MA crossover or whatever other ridiculous trading algorithm. Of course none of the other 100 million guys who are smarter than him figured it out before. Nor did any of the high frequency trading houses or Goldman Sachs or other institutions with sheer unlimited financial resources. Are you seriously telling us that this is a viable scenario? I rather believe and am a living example that years of screen time and risk taking/management experience in the market allows one to discover a small edge here and there and if one has a strong enough determination one might be able to exploit such edge for limited amount of time before everyone else discovers it. But certainly do I not believe some arrogant and snotty python prick walks along and beats them all. The odds of such happening alone are infinitismally small. But it makes for a great bed time story to fan the hopes of losers who are too lazy to put in the work required to discover and use the rare edges in this field. Of course I am sure some will claim "what does not work for you does not mean it's impossible " . If someone does not want to use at the very least their common sense that some on this threat put up blatant lies then that is of course their choice.
My system is a hybrid, it works out the probabilities for me, and then I take the trades in a discretionary manner. You can check the SPX thread, for some of my recent activity.
Congratulations for your 3 points you scored in SPX. Are you really not feeling embarrassed to show this as your "recent activity"? It is quite hard to take you seriously that way...
This is what I'm talking about, when I say you attack people. I have no problem with posting my real name, but if I were to release your information, I think you would be very upset. No, these are not the results of NN prediction, the first is a joke, the second is the truth. No problem bro, I hope you went out yesterday (Saturday night), and didn't only look to the internet for your fun. Accumulated so far I'm up around (56 + 16 + 30 =) 94 points SPX.
Congratulations for your virtual gains. Well done on paper. By the way I did not release your real name you did, it's on your GitHub site in several of the documentations.
Intense scrutiny? I always give credit where it is due. I found the article " System Trading? Trading Experts? Let’s Be Honest" at price action lab quite spot on. Pretty much resounds with what I have said.
I consider myself to be a pretty good trader, I've paid my dues, and I happened to learn the correct information from the very beginning. I think it's amazing how Academics, and 'People on the Internet', can say categorically that something doesn't work. There is a logical fallacy named Appeal to Ignorance: Just because you've never seen something in action, does not mean that it does not act. The things we're doing with NNs these days are incredible. From predicting the weather to election results, data manipulation has become king. The markets are a difficult breed, since they actively resist being twisted by mathematics. However, when training our network, weights are everything. Going back to how I learnt to trade, I had good information from the beginning. This weighted my decision making process. If someone cannot make a NN work for the market, then they are using the wrong weights.