Reproducing and Analyzing WorldQuant’s Alpha101 Factors
Summary
The article explains how WorldQuant’s 101 formulaic alphas combine short horizon price and volume features, often mixing momentum and mean reversion. It distinguishes signals traded on the same day as their latest input from those traded later, and walks through representative formulas using cross sectional ranks, rolling correlations, moving statistics, decay weights, and industry neutralization. It also describes a platform implementation for Chinese equities and outlines a daily factor analysis workflow.
Reported characteristics from the original study include holding periods of roughly 0.6 to 6.4 days, average pairwise correlation of 15.9%, a strong relationship between factor returns and volatility, and no significant dependence of returns on turnover. These observations are descriptive rather than proof that any factor will remain profitable. The article supplies many implementation examples and formula definitions, but its platform specific SQL and Chinese market classification choices may not transfer directly to other datasets or trading settings. It does not establish that the factors survive realistic costs or out of sample validation.
Key ideas
- Alpha101 expresses short horizon price and volume behavior through formula based signals.
- The factor set includes both mean reversion and momentum elements with differing trading delays.
- Cross sectional ranking and rolling time series operations are common construction tools.
- The article reports short average holding periods and relatively low average pairwise factor correlation.
- The reported return and volatility relationship does not establish future profitability.
Tags
This summary was written by Stratmill's research agent from the original; it is not a copy of the source.