Quick Code: Predictive Power of Risk-Adjusted Returns
Building variables and running regressions
We can measure reward/risk as return from X days ago to today, divided by the max drawdown seen in that time. Max drawdown equals the minimum price in X days, divided by the price at the beginning of the time period.
Now we can divide previous return by previous max drawdown. Before using it, let's check the statistical properties of this input variable…