I have a machine learning algorithm that trains on some technical indicators
input: RSI, BB bands, stochastic, etc..
output: 1 (buy)/ 0 (sell) based on whether price closed higher than MA(5)
It has generated good results in the backtest but i believe it is fundamentally flawed. Can anyone explain to me why such a method is flawed?
Submitted October 27, 2020 at 12:20AM by geriophile