What is the per-ticker price-range check (CI90 coverage)?

Per-ticker band coverage is how often a specific ticker's own historical 90% confidence bands actually contained the realized price — distinct from the platform-wide calibration average.

Per-ticker vs. platform-wide calibration

The platform-wide calibration page aggregates coverage across the committed TOP-20 backtest by market cap. That aggregate can look well-calibrated (85–95% range) while individual names remain poorly covered — and on the current record it does exactly that, with per-ticker coverage running from the mid-30s to nearly 100%. Structural-break tickers, like a company mid-acquisition, are the classic case, but they are not the only one: several large, stable names are also covered far worse than the aggregate suggests.

We publish no coverage percentage for an individual ticker. One ticker's backtest is a single hold-out window — the model fitted up to one past date and scored against the one stretch of prices that followed — so the figure moves on nothing but which date the test began. It would also read backwards: a 90% range that held 99% of prices is too WIDE to tell you anything, yet 99% looks like the better score. A ticker page reports which way the range missed instead. The measured record is the universe table on the calibration page, reported separately at each quarterly forecast start date. See the full AAPL forecast for the current band.

Reading a reading near 90%

A reading near 90% is well-calibrated for that name, and both directions of miss are defects: well below it the model was overconfident (bands too narrow), well above it the bands were too wide to tell you anything — 99% is not a better score. We publish no single per-ticker percentage, because one ticker's backtest is a single hold-out window; the ticker page reports which way the range missed, and the calibration page holds the measured universe record.

Frequently asked questions

Why can a ticker have worse band coverage than the platform average?

Individual names, especially those going through a structural business change, can have materially worse per-ticker coverage than the platform-wide aggregate.

Is band coverage the same as a win rate?

No. It measures how often the confidence band contained the realized price — a coverage statistic — not whether a directional call on the ticker was correct or profitable.

Where can I see the platform-wide version of this figure?

The live calibration page reports aggregate CI90 coverage per horizon across the committed TOP-20 backtest.