Skip to content
All library documents

Combining Stock Selection Models for Changing Market Regimes

Article BigQuant

Summary

The article describes combining two Chinese equity selection models designed for different market conditions: a rebound after an index decline and a mild index pullback. Each model includes market environment features so that it produces candidates only when its intended regime is present. If multiple models qualify, their historical backtest performance is used to prioritize which model’s picks to buy.

The examples use signals related to limit-up stocks, recent price ranges, prior strong returns, and aggregate limit-up activity. The proposed combination pairs a low-level strength breakout approach with a strategy that follows stocks after a limit-up and a modest pullback. The article reports that combining the models reduced idle periods and increased returns compared with using the models separately, but supplies no numerical results or detailed validation. It also acknowledges that the combined strategy still spends substantial time out of the market and would need additional models. Historical model ranking may not predict future performance.

Key ideas

  • Separate stock selection models can be tailored to distinct market regimes.
  • Regime features can gate each model’s signals to the conditions it was designed for.
  • When several models match, the article prioritizes picks using their historical backtest performance.
  • The examples combine rebound strength with post-limit-up pullback entries.
  • The reported reduction in idle periods lacks numerical evidence, and the combined strategy remains incomplete.

Tags

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