Forecasting Prices with Lagged Autocorrelation and Return Regression

Summary

The document describes an indicator that compares closing prices with prices from a chosen lag over a rolling window, using Pearson correlation to identify recurring patterns. When correlation clears a user-set threshold, it applies linear regression to percentage returns and uses the result to project a future price. The chart displays that estimate as a direction-colored line, alongside a hypothetical price difference. Settings control the lag, analysis window, and minimum correlation level.

The explanation presents the tool as potentially useful in range-bound or seasonal markets and suggests tuning its settings by asset and timeframe. It gives no backtest, forecast horizon evaluation, or evidence that the projections are predictive. The author cautions that threshold choices affect signal frequency and noise, and that the indicator may be less useful during breakouts or strong trends. It should be treated as an analytical aid rather than a complete trading system. The supplied implementation also retains the last regression value when no new cycle is detected, so forecasts may persist; readers should inspect implementation details before relying on the plotted output.

Key ideas

• The indicator calculates correlation between current prices and prices separated by a configurable lag.
• A threshold on that correlation determines whether the method treats the data as cyclical.
• When the threshold is met, regression on percentage returns informs a projected price.
• The lag, analysis window, and correlation threshold affect the indicator's responsiveness and signal count.
• The document offers no performance evidence and advises caution in strongly trending or breakout markets.

Tags

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