Cointegration Traffic Light

"These two symbols are cointegrated" is a claim that needs an actual test, not just a chart that looks like it reverts. This indicator builds a regression spread the same way a pairs-trading z-score tool does (an OLS hedge ratio between the chart symbol and InpSecondSymbol over InpRegressionWindow bars), then runs a Dickey-Fuller test on the spread's own trailing InpDfWindow bars: regress the spread's bar-to-bar change on its own lagged level, and take the t-statistic of that slope. A more negative t-statistic is stronger evidence the level series pulls back toward its own mean rather than wandering like a random walk - the standard unit-root logic behind every ADF-based cointegration test. HONEST ABOUT WHAT IS SIMPLIFIED: this is a plain Dickey-Fuller regression, not an Augmented one - it has no lagged-difference terms added to whiten serial correlation in the residuals, which is what the "Augmented" in ADF normally buys you on data with its own autocorrelation structure. The critical values used (InpCritical1pct / InpCritical10pct, -3.5 and -2.6 by default) are the commonly published approximate Dickey-Fuller thresholds for this sample-size range, not a table looked up for the exact window length. Treat the traffic light as a genuine statistical test with an honestly simplified specification - useful for tracking whether a pair's mean-reversion behaviour is currently holding up - not as a precise MacKinnon p-value suitable for academic citation. The light is recomputed on every closed bar and shown both as a plotted t-statistic (coloured by tier via the split-buffer trick, since MQL4 has no DRAW_COLOR_* line types) and as a plain-language panel label, with an optional alert when the light changes colour. Notes and limitations This is a plain (non-augmented) Dickey-Fuller test - see the full description above for exactly what that omits versus a true ADF test. The hedge ratio is a simple OLS beta with no intercept, recomputed fresh over each InpRegressionWindow rather than tracked incrementally; a spread built from a genuinely unstable hedge ratio can flip the light frequently even when the underlying relationship is real, since the regression window itself is reacting to the same price action. As with any pairs tool in this catalogue, both symbols are read at the SAME shift index, which is accurate when both share the chart's timeframe and closely aligned session hours, and only approximate otherwise.

Indicator · MT5 · v1.0 · Free

Tags: Hedging · Trailing Stop & Break-even · Alerts · MT5 Indicator · MQL5 Source Code

MetaTrader 4 & 5 indicators · PolyMaria