factor-research

Quantile backtests generate group equity curves across sorted factor quantiles to assess monotonicity, long-short spread, and nonlinearity, allowing you to visually confirm whether higher quantiles consistently outperform lower ones over time.

The best way to combine multiple effective factors is using equal-weighted, IC-weighted, or orthogonalized aggregation techniques, which merges screened effective factors into a multi-factor portfolio to streamline quantitative stock selection.

You run a quantile backtest by inputting aligned factor CSV and return CSV files, which generates group equity curves to assess monotonicity, long-short spread, and nonlinearity across sorted factor quantiles for cross-sectional research.

Factor backtesting requires aligned factor and forward return matrices formatted as CSV files, along with a designated output folder, to properly compute IC/IR statistics and quantile backtest curves without look-ahead bias.

Unstable IC performance indicates suspicious factor validity, flagged by computed summary statistics and stability clues that compare your daily IC series against established thresholds for basic, strong, or suspicious performance.