Factor Ranking and Long–Short Portfolio Returns in Grouped Tests
Summary
The document discusses how to evaluate a stock factor using portfolios formed by sorting stocks on factor values. In a long–short comparison, the return spread is calculated between the group ranked highest by the factor and the group ranked lowest by the factor. This tests whether the factor ranking separates stocks with different subsequent returns, rather than simply selecting stocks that happened to have the highest or lowest returns.
A reply explains that grouped backtests are intended to assess this return discrimination and that the group numbered zero is conventionally assigned the largest factor values in the platform’s setup. The discussion also raises a question about whether a template should show groups zero through nine and place the return curve at the top. It offers no backtest results, detailed portfolio construction rules, or evidence that a particular factor works; group ordering is a convention that should be checked when interpreting charts.
Key ideas
- A factor-sorted long–short portfolio compares returns of the highest-ranked and lowest-ranked factor groups.
- Grouped backtests test whether factor values distinguish stocks by subsequent returns.
- The platform convention described assigns the largest factor values to group zero.
- Group labels and chart ordering should be verified before interpreting a factor analysis.
Tags
This summary was written by Stratmill's research agent from the original; it is not a copy of the source.