Back
Autocorrelation Signal Detector
Compute ACF and PACF plots for a return series, highlight significant lags, and identify momentum vs mean-reversion signals.
Feb 23, 2026, Eric
The Core Idea
Autocorrelation measures whether past returns predict future returns. Positive autocorrelation at lag 1 means yesterday's up predicts today's up — a momentum signal. Negative means yesterday's up predicts today's down — a mean-reversion signal. This is the empirical foundation of trend-following and mean-reversion strategies. The ACF shows the total correlation at each lag; the PACF isolates the direct effect, removing indirect influence through intermediate lags.
Data Selection
500 observations | 95% confidence band: ±0.0877
Autocorrelation Plots
ACF — Autocorrelation Function
PACF — Partial Autocorrelation Function
Significant Lags
| Lag | ACF | PACF | Signal | Strength |
|---|---|---|---|---|
| 1 | +0.2686 | +0.2686 | Momentum | Strong |
| 5 | -0.1017 | -0.1151 | Mean-Reversion | Moderate |
What This Means
Lag 1 ACF = +0.269 (strong). This series exhibits momentum — yesterday's direction predicts today's.
A trend-following strategy would exploit this structure. Positive autocorrelation means up days tend to follow up days, and down days follow down days.