I have recently become interested in Reinforcement Learning, mostly as a result of Alpha Zero’s success in the chess (and my own enthusiasm for the game). While I understand the utility of RL in board games, I am struggling to comprehend it‘s use in trading - which is constantly being mentioned as another important application. What I don‘t understand is the following:
In Chess, the next state of the Markow Decsion Process (e.g. the board position) is clearly a consequence of the action taken (e.g. the move played). If we formulate a trading MDP with the values of your current stocks as the states and three actions (Buy, Sell, Hold) for every stock, this is not the case. The future value of your portfolio mostly depends on how the prices develop, and only partly on the specific action you took before.
I am probably missing something here and would be very grateful for some clarification!