Take historical data, and extract on average two time points that had the greatest degree of variance. Since you don't know the direction, pivot trade the time point and close out at end of each session.
This is so so true. Just like you don't need 6 monitors with all those flashy charts to trade manually, you definitely do not need to overload your algo to the max to get the full benefit.
Hi, guys I'm a software developer with algorithmic background, done some hobby tensorflow/keras coding for intraday scalping strategies. Let me know if you are interested in me getting involved, I'm looking for interesting projects to join.
you'll know when you've hit a winner - time of day fluctuation to a fixed profit should be an interesting place to start - record expected draw down based on profit target over time - weigh profit to loss percentage