The idea here is to run multiple different algorithms "in-parallel". Each day, each algorithm indicates long or short, and a relative strength of long or short. Those outputs are normalized on a -1 to 1 scale, with -1 meaning really short, and +1 meaning really long.
Average all the indicators to come up with a composite score for the day (-1 to 1), and order accordingly.
Need to:
- add more strategies
- potentially weight strategies heavier than others (or based on market conditions? i.e. mean reversion does better during a volatile market)