Concept
How confidence and baselines work
What a "typical night" is built from, the floors that hold a comparison back until there is enough history, and why a short record is treated as noise.
What "usual" is built from
When RollCall says a night, or a moment, is compared with "usual," it means one specific thing: the venue's own history on the same weekday. A Friday is read against recent Fridays, a Tuesday against Tuesdays — never against a different weekday, and never against another venue. Crowd baselines are built the same way, half-hour by half-hour, from the venue's own same-weekday scan history.
Where a typical figure is a single number, it is the median of those past nights, not the average. A median ignores the one freak night — a private buyout, a washout — that would drag a mean around, so "usual" reflects the ordinary night rather than the outlier.
Floors: why a comparison sometimes waits
A comparison only appears once there is enough history to mean something. Completed nights vs. usual needs four same-weekday nights before it renders; the recent-nights strip needs four plottable points; crowd baselines need a floor of measured nights at that half-hour before a room is judged off its pattern. Below the floor, RollCall shows the raw fact and says it cannot yet compare — it does not offer a comparison it cannot stand behind.
These floors exist because a short record invites you to read noise as a pattern — the documented human tendency to over-trust small samples and to see streaks in randomness.[1][2] One to three points look like a trend to the eye; they are usually just variance. The floor is the guardrail against that.
What confidence is, and isn't
"Confident" here means the inputs are proven and the arithmetic is exact — not that the number is a verdict. A figure can be perfectly certain and still carry no judgment about whether it is good: higher or lower than usual is context, never automatically better, worse, or a flagged exception.
- Belief in the Law of Small Numbers ✓ verified
Read the source ↗ - The Hot Hand in Basketball: On the Misperception of Random Sequences ✓ verified
Read the source ↗