notadamking / RLTrader

A cryptocurrency trading environment using deep reinforcement learning and OpenAI's gym
https://discord.gg/ZZ7BGWh
GNU General Public License v3.0
1.71k stars 537 forks source link

Curve-fitting seems to be result of high profits #64

Closed developeralgo8888 closed 5 years ago

developeralgo8888 commented 5 years ago

The agents have been trained and fitted with same data. Marco de Prado has great book to avoid over-fitting but also use stationary synthetic data with similar distribution to train the agents

TheExGenesis commented 5 years ago

What do you mean? They're tested with unseen data

notadamking commented 5 years ago

This is incorrect. The agents were trained/validated/tested on separate data sets.