Building and Evaluating a Price, Volume, and Valuation Composite Factor

Summary

This example documents a manual, language-model-assisted factor-building workflow for equities over 2022–2024. It combines a stock’s price relative to its 20-day average with cross-sectional standardized price-to-earnings data, scales the composite by its rolling 20-day standard deviation, subtracts its own rolling average, and standardizes the result across stocks. The workflow demonstrates data retrieval, feature operations, and an evaluation step rather than presenting a ready-to-trade strategy.

The reported evaluation is weak: information coefficient is 0.0005, cumulative return is -0.037, annualized return is -0.022, Sharpe ratio is 0.017, annualized volatility is 0.228, and maximum drawdown is -0.215. The author concludes that the factor performed poorly and proposes adding price-to-book data, while noting that the manual process is laborious. The example supplies no benchmark, transaction costs, implementation details, or evidence that the proposed revision would help; it is best read as a factor research workflow and an unsuccessful result.

Key ideas

• The workflow combines price relative to a 20-day average with standardized price-to-earnings data.
• It scales the composite by rolling volatility and subtracts its rolling average before cross-sectional standardization.
• The reported information coefficient is close to zero, with negative cumulative and annualized returns.
• The example illustrates manual factor construction and evaluation through a language-model interface.
• Adding price-to-book data is proposed as a next step, but its effect is not evaluated.

Tags

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.