Historical example (analog)
A past stock observation selected for comparison with the chosen stock and date. A similar shape does not guarantee the same cause or outcome.
Sample (cohort or comparison set)
The group of selected historical examples. Its count can include repeated stocks or nearby dates, so it is not necessarily a count of independent events.
Trading day (session)
A day when the relevant market is open. Five trading days is different from five calendar days. The simple tool starts from a completed trading day.
Excess return
A stock's return minus its date-matched liquid-stock baseline, measured in percentage points. A +2-point excess return means 2 points above that baseline.
Middle result (median)
The central historical outcome: roughly half the observations are on each side. It can sit near zero even when individual outcomes differ widely.
Lower and upper ends (percentiles)
The simple tool uses the 10th and 90th percentiles to describe the middle 80% of historical outcomes. Some observations fall outside that range.
Coverage
The fraction of evaluated outcomes that landed inside a particular published range. The intended fraction and the observed fraction can differ.
Calibration
Checking or adjusting a method so its claimed range coverage agrees with observed behavior under a defined evaluation. It does not establish trading profitability or apply automatically to another method.
Comparison check
A method-specific check attached to a historical sample. An informative flag is not a directional accuracy score. A failed or missing check must remain visible.
Recorded example
A previously captured response, labeled with its stock, trading date, and capture time. Running the same query later can return different data or methods.