Iterative Factor Evaluation and Context Management in Quant Research

Summary

This project note describes iterative construction and evaluation of equity factors using amount, turnover, price, market capitalization, and price-to-book data. Candidate expressions combine standardized ratios and rolling statistics, including volatility, percentile rank, and correlations. The note records historical information-coefficient and portfolio metrics for several candidates; most listed tests show negative cumulative returns and Sharpe ratios, while one price-rank combination has positive reported cumulative and annualized returns and a positive Sharpe ratio. These are the author’s reported results, with no details on sample, universe, costs, or validation method.

The note also focuses on the workflow problem: carrying every intermediate expression into each model prompt causes the working table and token use to grow, increasing redundancy and the risk of model errors. It proposes keeping factor results in a primary table and removing temporary data tables between rounds. It warns that increasingly complex expressions also make analysis and backtesting slower. The evidence is exploratory and insufficient to establish a robust factor.

Key ideas

• The factor experiments combine standardized trading activity measures with rolling price and turnover statistics.
• The note reports mixed historical results, with most evaluated candidates showing negative returns or Sharpe ratios.
• A candidate adding a rolling price percentile rank has positive reported cumulative and annualized returns.
• The author recommends retaining factor outputs in a primary table and clearing temporary data between evaluation rounds.
• Growing factor complexity can increase model confusion and lengthen analysis and backtesting.
• The document does not specify enough methodology to establish that any reported factor is robust.

Tags

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