Skip to content
All library documents

Ornstein–Uhlenbeck Spread Models Compared with Rolling-Window Pairs Trading

Article arXiv papers · Author: Jirat Suchato et al.
Read the original

Summary

This preliminary analysis applies an Ornstein–Uhlenbeck process to model spreads between paired stocks for a mean-reversion trading strategy. It compares that approach with a simpler method that estimates the spread’s mean and standard deviation over a rolling window. The comparison addresses both signal behavior and risk-adjusted performance.

The authors report that the OU model captures signals and trends effectively, but performs worse than the rolling-window approach on a risk-return basis. They suggest that non-stationary pairs and limitations in parameter tuning may explain the weaker result. The description does not state the sample, asset selection rules, trading costs, or specific performance metrics, so it supports only a preliminary comparison. Pair stability and parameter choices remain important caveats when applying either model.

Key ideas

  • The study uses an Ornstein–Uhlenbeck process to model spreads between paired stocks.
  • It compares the model with a rolling-window method for estimating spread statistics.
  • The OU approach captures signals and trends but underperforms on a risk-return basis.
  • The authors point to non-stationary pairs and parameter tuning as possible limitations.
  • The brief report does not specify data, transaction costs, or detailed performance measures.

Tags

EquitiesPairs tradingMean reversionStatisticsBacktesting
Full text
# An Application of the Ornstein-Uhlenbeck Process to Pairs Trading


# An Application of the Ornstein-Uhlenbeck Process to Pairs Trading









We conduct a preliminary analysis of a pairs trading strategy using the Ornstein-Uhlenbeck (OU) process to model stock price spreads. We compare this approach to a naive pairs trading strategy that uses a rolling window to calculate mean and standard deviation parameters. Our findings suggest that the OU model captures signals and trends effectively but underperforms the naive model on a risk-return basis, likely due to non-stationary pairs and parameter tuning limitations.

Shown in full with attribution under the source's licence. Licence: abstract CC0

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