really? that answer should be obvious ir u have never had a consistently prifitable system" As long as it doesn't cannibalize your own trade and eat into your own profits, you can do both. And like I said, you can offer a basic AI deep hedging product to the alpha-hungry public that empirically determines best hedging practices for an individual underlying asset going as far back as a month or 2. While for your own operation, you keep a much more sophisticated proprietary version that can handle all Black Swan tail events, includes a predictive pricing model, and is scalable on a grand scale...able to handle any amalgamation of similarly traded products, baskets, and uncorrelated or inversely rated portfolios.
How does AI relate to the market market narrative per Wyckoff, Livermore, et al?, that: o The overwhelming portion of market-investable funds concentrate in the hands of relatively few "pools" (large investors/funds, today perhaps not so conspiratorial as implied by "pool"), which want/need to BUY LOW SELL HIGH WITHOUT MOVING THE MARKET. o The interplay of the pools and all the others, mostly not so well-informed or capable players, forms the structures of price action - flags, tops, bottoms, impulses, ... So long as capital is concentrated in pools, will the forms of price action stay the same, regardless of AI ? AI players who have huge success -- will their increased size force upon them the preoccupation of not moving the market, for which the solutions are those of old, same as now, and mostly not related to recent AI approaches ? Is the import of AI on price action that the forms of price action get "fuzzier" and require a more acutely sensitive player to be read ? (Years ago Brooks mentioned that 3push structures are rarely perfect, I think he ascribed that to computers). Or will (do?) pools using new forms of AI actually change the form of price action ? Perhaps the history of Wall St. entries into small, developing markets offers some indication of what may be in store.
everyone knows due to assumptions like random the black sholes would never predict complexities known to the elite. you can teach a neural net blacck sholes but you would need a lot of data
black sholes is obviously trivial since it has inadequate high dimensional structure on the randomness unlike basic viscous flows which are just bell curve liked without even correaltion or greeks and are fully understood by investment professionals. the globalism.
Trading is not like the closed-system, fixed-rules games (e.g. chess, poker, Starcraft) which have been "beaten" by AI. The cleverest "deep learning" algo can't change the fact that buying $1 billion of stock will move the price against you. The growth of AI, like the growth of passive funds, will cause some edges to recede or vanish while new ones will open up. 2018-19 has been great for traders with these repeating V-crashes and subsequent rips higher - a pattern likely caused by the bifurcation of the market into passive dumb money and a backs-to-the-wall active segment that's been replacing humans with quant and AI strategies. Lo and behold, these ostensibly sophisticated robots all end up placing the same trades and crowding to the same side of the boat. Personally, I think the growth of AI will actually make trading easier overall for small-scale players. Competition will get tougher for the types of edges that AI can easily recognize and exploit, while falling off quite a bit for those strategies which require even a modest amount of human experience and skill.