You most likely could, all of the "backtesting" in terms of how much profit the bot is making is established via the wallet class. You could look through the code and see everywhere that the wallet class is referenced and replace it with a backtesting framework. You will also need to derive the reinforcement learning reward from your backtesting framework (currently this is % profit per season).
You most likely could, all of the "backtesting" in terms of how much profit the bot is making is established via the wallet class. You could look through the code and see everywhere that the wallet class is referenced and replace it with a backtesting framework. You will also need to derive the reinforcement learning reward from your backtesting framework (currently this is % profit per season).