Kevin Davey, an (apparantly) successful algorithmic trader/author claims simple strategy works and will always work. Ernest Chan, another guru in this field, however, says simple quant strategies don't work anymore and machine learning is a must if you want to suceed in trading. He also says (along with Marcos de Prado) that it is impossible to do ML-based system trading on your own; you need a team based approach, since it is so labour intensive work. One of the main motivating factor to strive to be a trader must be for a lot of people, the freedom and independence it entails, the idea that you can be succesful by working on your own. What the latter two guys seem to suggest is demorilizing to some degree. What do you guys think? Can a retail trader succeed in algorithmic trading? Even if I have to use ML, is it really impossible to do ML based trading on my own? Thanks in advance for your replies.
They can, but to reach the highest level in this game, you absolutely must have a gambler mindset, who is willing to lose his shirt, pants and possibly wife and kids. If you don't like that idea, don't even initiate the learning process. Just invest long-term and enjoy a normal life.
Do what big business does, divorce your wife and give her any asset that is positive and you get everything that is negative. Only thing you have left is trading account and trade like no tomorrow, run that sucker as high and as risky as you can, if you make it big, close out and live a good life, if you miserable fail, file bankruptcy, hopefully your wife takes you back to live with her. USA citizens bail out big business, and yours be like a trickle. "algorithmic-trading" has a shelf life of many of them lasting months until "AI" figures them out to join and too many to the party means at some point "no beer", well hell, better have kegs going to replace what others seems to have with your Keg system which either hits or misses. You know what always works? Death, so you better learn how to hedge.
I believe Ernie has gone on record saying the strategies in his book are old strategies that don't work for him anymore. I wouldn't be surprise if @kevinkdog did the same in his book. Seems kind of silly to put strategies in your book you are still using. Both of chan's major publications on this, as well as kevin's, are seminal works in retail algo trading. Kevin is one of the few people with a verifiable record in a trading competition (though we sort of take his word he did it algorithmically). I'd listen to what he has to say. I'm not an expert in ML with finance and I have some decent experience with regular garden variety ML. As you pointed out De Prado's opinion is important when weighing Chan's opinion on the subject. The summary is financial machine learning is virtually impossible to do right because of the large amount of data you need, awareness of bar types and binning, and many other seemingly trivial but hugely important things. To me, it's borderline worthless to even try. I have been able to generate some models based on minute data, but the sheer amount of data and the constant retraining was a waste of time. Leave ML to the MMs and big desks. In summary: I do not think it's possible as a retail to succeed with machine learning in algorithmic trading. I will go against Chan here and say If you define algorithmic trading as "coding a system to fire signals for you", sure, if you have the API access and the money. Algorithmic trading systems are not a panacea, and if you're undercapitalized an algorithm can blow through all your cash + more very quickly. I would strongly discourage undercapitalized retails (perhaps, under 50k at a minimum) from trying to do fully autonomous algorithmic trading. Most retails would be better served by discretionary, or systems based trading - perhaps augmented by software to help you detect and capitalize on signals.
INMO, yes, a retail trader can succeed in Algo trading. It'd depend on his strategy. On the other hand, I don't put much trust in M.L because it depends on pattern recognition and markets are always changing them. It's more useful to have a more global abstract idea that could work regardless of chaotic intraday patterns.
Hello, I already asked this question in another thread, but did not get an answer: What other marketplace - than mql5.com - do you know, where private (retail) traders can buy or rent financial algorithm robots?! I searched the net, but couldn't find any. Also asked here, but didn't receive answers. That's somehow speaking for itself... as most don't even know where to get advanced algorithms from... How can you compare something, that you don't even know of?!
You took the words right out of my mouth. Don't believe these authors. My saying is,"if you have to sell me something, you're not a successful trader." Just a professional bull shitter.
I think you need to determine what the word "works" aka edge is. For example, a simple buy-pullback-while-in-trend strategy works as long as you are in a trend. Is that an edge? A simple mean reversion strategy like Bollinger Bands fade works in mean reverting environment. Is that an edge? I don't know ML, but I suspect it's used to find persistent anomalitties in the markets which are independent of current market regime. That's my take on the difference between the two opinions above. As usual with trading, there's no right or wrong. BTW, I am disappointed with @kevinkdog as he doesn't seem to contribute to the forum. I understand he is paying to be a sponsor so his choice, but other sponsors do.