I have tested the 107 stock factors and now it is time to share what is perhaps the most important lesson I have to offer. The reason I used the screener backtest for factor testing as opposed to ranking system performance is important to understand what is going on "under the hood".
Below is a gallery of screenshots I took when I was testing the stock factors. As you can see there are many anomalies. Some of these factor anomalies are already known, such as institutional factors. But not all.
Most of these anomalies occur at the start of the database, and for this reason I only develop portfolios based on the most recent 10 years, not the entire database.
The next important lesson for stock market students is to never let anything go until it is resolved. Now I am going to study each anomaly, determine if it is a true problem and then raise the issues with Portfolio123. It isn't my job and I don't really have time for this but I want to work with quality tools and I don't believe these issues will be addressed unless the issues are pushed.
So now you know why I have been dragging my feet in writing the next installment of the Ranking System tutorial.