Statistical Asset Drift Classification with Robustness Checks

Summary

This indicator assesses whether an asset has positive or negative drift over a fixed horizon. It collects non-overlapping log returns, calculates their mean, median, dispersion, hit rate, and an annualized drift estimate, then checks whether the evidence clears statistical and economic thresholds. Its classification also depends on sign consistency across two sample halves, agreement between mean and median direction, a power check, and a variance-ratio test for mean reversion.

The chart display reports the classification and supporting statistics, with optional bar coloring, background shading, dashboard, and alerts. The script presents a retrospective diagnostic rather than a trading entry or exit system. Its inference requires a minimum sample and labels smaller samples as heuristic; results also depend on fixed thresholds and assumptions in the tests. The code provides no asset-specific performance study, out-of-sample validation, or evidence that classified drift persists.

Key ideas

• The indicator uses non-overlapping log returns to estimate drift over a fixed horizon.
• It combines significance, estimated power, economic magnitude, and direction checks to classify drift.
• Split-sample consistency and a variance-ratio test screen for unstable direction and mean reversion.
• Small samples are flagged as heuristic, and the output is a retrospective classification rather than a forecast.

Tags

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