Introduction

When stocks receive positive news (earnings beats, analyst upgrades), sentiment spikes. But how long does the positive sentiment effect on price persist? Plotting sentiment decay curves—tracking how positive sentiment decays over days and weeks—reveals the half-life of sentiment effects and informs holding period decisions. Machine learning enables precise decay curve estimation.

Sentiment Decay Modeling

For each positive news event, track sentiment on event day and subsequent days. Fit exponential decay models: Sentiment(t) = Sentiment(0) × exp(-λ × t), where λ is decay rate. Extract half-life: time for sentiment to fall to 50% of initial level. Estimate half-lives by stock, sector, and news type. Results: analyst upgrades decay with 5-day half-life, earnings beats with 3-day half-life, short-squeeze rumors with 1-day half-life.

Trading Applications

If positive sentiment decays in 3 days, mean-reverting traders sell rallies on day 1-2, expecting regression to mean by day 4-5. Momentum traders exit shortly after the decay half-life. Understanding sentiment decay improves timing of exits.

Conclusion

Quantifying sentiment half-lives reveals the temporal structure of behavioral effects, enabling more precise strategy timing.