Drawdown Reward System

[Claud_Badau_9eb9P (Claud Badau)]

I'm going through the reinforcement learning course and I was wondering if anyone has any experience with a reward system that minimizes drawdown.



Any resources, code or examples would be much appreciated!

[Suleyman_Emre_Yesil_9LQXr (Suleyman Emre Yesil)]

Hi Claud,



Instead of directly creating a model for drawdown minimisation, I suggest either improving the accuracy of a model for price prediction or building a model for portfolio optimisation. Eventually, both will help to minimise maximum drawdown.



For example, one of the main objectives of portfolio optimisation is hedging through diversification which reduces maximum drawdown. In addition to the course you are taking, you may check here for a collection of GitHub pages containing codes for portfolio optimisation with RL. 



Hope this helps.

[Claud_Badau_9eb9P (Claud Badau)]

Thank you, this is helpful!