Insights into the alpha pipeline and effective machine learning
Often you will see researchers working with X period trade bars and developing strategies that rebalance their portfolio at a given interval (commonly that of their data interval).
This is great for keeping your logic simple and reducing the need for large data processing/storage on the research side, but it leaves money on the table.
It also makes us easier to front-run and leads to us missing out on potential short-term seasonality gains in execution. Excessive turnover also poses a problem when we keep the logic this simple.
A Simple & Dirty Example
When Should We Exit?
Portfolio Optimization [Code Here]