I always hear about how day trading or even professional point and click (HF/prop) trading is going to be a thing of the past in 10 years, but why do people believe this? Why isn't it possibly for humans to exploit algos and AI assuming it's impossible for them to become perfect traders. Algos and AI would have to have stops on too, so it's interesting to listen to traders who believe there is an invincible force coming to destroy all the markets...

Although I do understand why algos and AI can easily dominate spaces like news driven stock trading, options trading, and forex trading, but those advantages aren't as pronounced in futures. The same data that AI would be using would also be available to us. Sure it could sift and organize data faster, but im not entirely convinced that this is going to become a useless space. Is there something im not seeing?


Sent from my iPhone using NexusFi

I've been learning a lot about auction market theory and market profile lately and I think that the people who believe that algos and AI will take over maybe don't have the greatest understanding of how/why markets actually move? I haven't actually heard anyone with a lot of trading experience making those kinds of claims. In my opinion what will happen as they get implemented is markets will thicken with liquidity and they will trade "tighter" as an AI system is more exact with things like profit taking etc, but the market will still seek out the fair price and will have to move to do so, that will not change. A discretionary trader will still be able to see the order flow and balance of market/limit orders and therefore imo there will always be opportunity for humans to profit.

What's especially important is wherever there is access to order flow information there is opportunity for institutional traders to profit off of trapped shorts/longs with their large position size, so I don't see that changing unless the liquidity in the futures markets gets significantly higher to where that's impossible. Idk long story short I agree with you, humans will be able to exploit algos and vice versa as long as the information required to do so (order flow) is there. That's my 2 cents

Thanks for the incite I totally agree. I think it's definitely possible for markets to thicken up, but imo that's not exactly a bad thing, too.


Sent from my iPhone using NexusFi

@Raider I think machine learning/AI can blend well with discretionary trading. However, yes there are valid arguments for why machine learning can so called "take over". There are a range of arguments. But, I will share just one. But, in order to follow the reasoning requires for me to layout some background.

The background supposition is that trading at a high level, i.e profitably, is "difficult". Difficult will apply both to discretionary and system traders. The difficulty for the discretionary trader is a product of a few factors but namely biological and behavioral. We might compare it to becoming a professional basketball player: you need to get the proper training but also physical and mental characteristics are important too. A system trader also needs certain attributes, training, etc to become successful. There is also a certain amount of time/work required to produce a good system.

The argument for ML taking over is that machine learning could allow for anyone with computing resources to basically develop not only similar but better strategies then the ones in existence and without needing selective advantage. The advantage will be used first by big money. In order to understand how this works, big money can deploy hundreds or thousands of strategies into the markets. Of course, system traders do that already. However, what we can imagine is that system traders and discretionary traders maybe don't compete for the exact same trades or that maybe discretionary traders can develop themselves beyond the trading systems. Today, we can imagine these advanced funds are deploying a bunch of "so so strategies". But, with advanced machine learning they could deploy more advanced strategies that could trade at a higher level. So, they won't need to hire traders. They won't need quants, algo traders, etc either. The result will be a stratification and polarization. Perhaps in years past, a discretionary trader could learn and apply their craft by programming systems or algos or becoming a junior level quant but in the scenario where AI/ML take over then only the star quants will be able to keep their jobs. The ML will develop the algos that traders might have developed before. The new trading jobs will be lower skilled and low payed.

The next step is commoditization. The technology eventually flows to the masses. The effect of all of this would be to make the markets very unpredictable and drive down profits for all traders to zero. ML and AI could also do what automation did to the floor traders to system traders and quants: basically a large firm with advanced ML and AI tools doesn't need to purchase strategies or hire developers. The machines will develop the strategies. So, the effect will be similar to other industries with the net effect being job losses. The good news is that brainacs at Ren Tech can go back to contributing the real world but for traders who don't have the aptitude for being either a top 1% PHD level quant nor can no longer compete in the "new markets": the only offer will be the fast food line.

As for "competing", the idea is that the ML and AI will be able to learn both faster and better then a human trader. We can imagine the efficiency of the markets might ebb and flow and eventually they will learn that too. As soon as profits start to go up, they will just trade more aggressively or frequently. We can imagine passive investors will be able to take advantage of these funds in a way that would not have generally been available. The net effect would be, one can imagine, to reduce all above normal profits to random unpredictable factors. We already have seen that sort of behavior for years with well-documented trend following strategies. And of course, I suspect the reality is that far fewer traders are trading today for a living then in years past because these sorts of changes aren't new.