SuperTrend AI (Clustering)

SuperTrend AI (Clustering) lets performance data choose the settings. The indicator evaluates the SuperTrend trailing stop across a user-defined factor range, scores how each instance has performed, and applies k-means clustering with three groups — below-average, average, and exceptional — to that performance set. The factor driving the on-chart stop is drawn from the chosen cluster's average, with centroids initialized from performance quartiles to speed convergence.

How to Trade the SuperTrend AI (Clustering)?

• Trailing stop flips: read them like the classic — the stop below price marks an uptrend, above price a downtrend.
• Signal performance metric: each signal prints a value; higher readings tend to accompany stronger trends, lower ones warn of possible retracements — informative, not foolproof.
• Stop plus adaptive average as S/R: both can serve as support and resistance; raising Performance Memory produces a longer-term adaptive moving average.
• Factor range choice: higher minimum and maximum factors favor signals suited to longer-term positioning.

SuperTrend AI (Clustering) Settings

• ATR Length: period of the ATR in the SuperTrend calculation.
• Factor Range: minimum and maximum factors evaluated.
• Step: increment between tested factors.
• Performance Memory: how much history weighs on performance scores; higher values emphasize the long run.
• From Cluster: which cluster — best, average, or worst — supplies the final factor.
• Maximum Iteration Steps / Historical Bars Calculation: caps on centroid iterations and the calculation window.

Frequently Asked Questions

How is the final factor actually selected?

Each factor's performance is tracked bar by bar, the results are clustered with k-means (K=3), and the factors in your selected cluster are averaged into the one that computes the plotted SuperTrend. By default that is the top-performing cluster, though From Cluster invites experimentation.

Is there an oscillator version of this idea?

Yes — the AI SuperTrend Clustering Oscillator carries the same clustering approach into a separate-pane format, if you prefer momentum-style reads over an on-chart stop.

What does SuperTrend AI (Clustering) cost?

Nothing — it is free on the LuxAlgo Library, and Quant, LuxAlgo's AI, can backtest its signals directly from this page.