Decoding Quantile Back test report

While alpha analysis is a very good starting point for discovering alpha, it should not be the end of your research. Before taking your strategy live with your newly discovered alpha, you should also conduct a rigorous quantile back test.

At times, even though you may have good alpha, the strategy could still not make money after factoring in transaction costs and slippages. So you need to back-test to check if you have a “tradable” alpha.

You can conduct 2 types of back tests. The first one is a quantile back test, which essentially creates a separate portfolio for each bucket that you have decided and conducts a historical back test based on the methodology mentioned here.

You can play around with position sizing, rebalancing frequency and other parameters as explained in these articles – Strategy Parameters.

The other is to develop a full-fledged strategy using your alpha and backtest the performance of your strategy against a benchmark of your choice. The workflow is similar to the strategy creation workflow described here. Once you create an alpha, you can directly use this alpha to create, back test and paper trade a strategy.

Also, when building a strategy, you can load any of your saved alphas directly into the workflow.

Ideally, your alpha research flow should start with 1. Alpha Analysis, and then move on to 2. Quantile Back Test and finally end with 3. Strategy creation and back test.