AI Timing for a Manually Screened China A-Share Portfolio

Summary

The author describes a China A-share swing strategy that first narrows the universe to selected high-quality companies and blue chips, then uses a machine-learning model to identify timing patterns. The training labels are based on relative performance against the broader market, aiming to identify stocks that hold up during market declines. The approach is framed as a small index-enhancement strategy, with market risk controls, stop rules, periodic rotation, and a five-stock portfolio mentioned as components.

The post reports a competition Sharpe ratio and favorable backtest and live results, but offers no supporting charts, detailed model specification, or transaction-cost analysis in the supplied text. It acknowledges overfitting risk and argues that a narrow stock universe can reduce noise, while maintaining a long history of observations for training. Those claims are not independently demonstrated here; survivorship and selection effects, validation design, and changing market regimes remain important limitations.

Key ideas

• The strategy manually restricts the stock universe to selected quality companies and blue chips before modeling.
• Market-relative returns are used to identify stocks that show resilience during broad declines.
• Machine learning is assigned the timing task within a swing and index-enhancement framework.
• The post combines timing with market risk controls, stop rules, and periodic portfolio rotation.
• Reported results are author claims and lack enough methodological detail here for independent assessment.

Tags

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