Intraday Trade Data Factors for Chinese Equity Selection
Summary
This research summary reviews stock-selection factors derived from tick-by-tick trade data, including large-buy share, buy-order concentration, intraday aggressive buying, and informed buying or selling measures. After orthogonalization, the factors reportedly show monthly cross-sectional predictive ability across the market, with average information coefficients around 0.03–0.04; most have annualized information-coefficient ratios above 2. The report also compares results across the CSI 300, CSI 500, and CSI 800 universes, finding that factor strength varies by index.
Raising rebalancing frequency from monthly to twice monthly improves several reported metrics, while moving to weekly rebalancing benefits only some factors. The summary notes overlap with conventional characteristics such as size, valuation, prior returns, and turnover, and residual correlations among some factors. Adding buy-order concentration improved one index-enhancement model's reported excess return, but the authors caution that factor value depends on the existing model and risk controls. The excerpt provides summary-level results, not the full methodology or independent validation.
Key ideas
- Orthogonalized tick-level trading measures show reported cross-sectional stock-selection ability.
- Factor performance varies across the CSI 300, CSI 500, and CSI 800 universes.
- More frequent rebalancing helps some factors, but weekly rebalancing does not consistently improve results.
- Tick-derived factors retain relationships with conventional characteristics and with one another.
- A factor's contribution to an enhanced index portfolio depends on the model and risk controls already in use.
Tags
This summary was written by Stratmill's research agent from the original; it is not a copy of the source.