As a result, the strategies I obtained through my work on this project leave a bit to be desired since they have achieved a Sharpe ratio of at best about 1.2 on

As a result, the strategies I obtained through my work on this project leave a bit to be desired since they have achieved a Sharpe ratio of at best about 1.2 on the backtesting data Keep in mind that market heavyweights have been using machine learning for some time to turn market patterns into money for themselves. They've spent many years more than you have and poured enormous financial and computing resources into doing so. My assumption is that juicy patterns that would be easily identifiable and exploitable by machine learning have largely been squeezed out of the system as low hanging fruit. It would be interesting to use data from many decades past in order to compare the effectiveness of several basic machine learning techniques with what those same techniques are able to achieve today.