Standard Error Bands: what it's actually best at

291 of 1,520 out-of-sample tests beat simply buying and holding the same asset — 19.1%. On the other 1,229 it did not. That is indistinguishable from the 20.1% rate across all 382 indicators we test (one pooled rate over all 660,005 tests we have run, not a mean of the per-indicator rates) — the difference is inside the margin this many tests can resolve, so read it as ordinary, not better or worse.

Each test is one asset on one timeframe: 1,520 of them, drawn from 885 assets across up to 4 timeframes. Not every asset has usable history on every timeframe, so that total is the grid we could actually run — it is not 885 × 4, and we do not pad it with tests we did not do. "Beat" means a higher return than holding that same asset over that same window. Measured out-of-sample — on data the setup was not chosen on.

Picking the single best timeframe for each asset after the fact raises it to 24.9% (220/885 assets). That number is the one worth distrusting: choosing the timeframe once you already know the answer is how backtests flatter themselves. Every indicator, ranked by this number