Average Time Calculator — Circular Mean

In 2026, a Trinity College sleep researcher analysing the wake times of a 60-subject cohort cannot use a spreadsheet AVERAGE() function. Half the cohort wakes between 23:30 and 02:00 the next morning, and naive averaging across those values produces a meaningless "noon" result that destroys any real cluster signal. The Average Time calculator implements the standard fix: circular mean — convert each time to an angle on the unit circle, sum the sine and cosine components, and use atan2 to recover the mean angle. The result for 23:30 and 02:00 is 00:45, not 12:45.

The mathematics of circular statistics was opened by Karl Pearson's 1905 paper "On the General Theory of Skew Correlation" in Drapers' Company Research Memoirs and codified by K.V. Mardia's 1972 textbook Statistics of Directional Data and N.I. Fisher's 1993 Statistical Analysis of Circular Data. The core insight is that any periodic quantity — angle, day-of-year, wind direction, magnetic compass bearing, clock time — should be averaged as vectors, not scalars. The atan2 function (introduced into Fortran-77 and later into IEEE 754) gives the correct quadrant of the resulting angle automatically.

Chronobiology adopted circular statistics enthusiastically. Till Roenneberg's Munich ChronoType Questionnaire (MCTQ, 2003) reports mid-sleep on free days (MSF) as a circular mean across the survey week. Sleep-cycle research (Czeisler 1999, Borbely 2016) uses the same machinery to compute dim-light melatonin onset (DLMO) averages. Without circular mean, every cross-midnight subject would be excluded or mis-aggregated — a methodological failure that pre-1990s sleep studies routinely committed.

Beyond research, the method shows up in transport scheduling (ferry, bus, airline dispatch averaging arrival times), broadcast media (averaging cue-times of evening shows), and family logistics (baby feed averaging, school-pickup scatter analysis). The 60-day sample averaging methodology is now part of the FAA Air Traffic Control Track Average computation, used for over-flight scheduling at major hubs.

The visual representation on this page mirrors the convention used in Fisher's 1993 textbook: each time is a dot on the rim of a 24-hour clock face, the mean angle is the bold radial line, and the resultant length R is the line's magnitude relative to the rim radius. R close to 1 means tight clustering; R close to 0 means the times are spread out and no meaningful average exists. The calculator surfaces R as a numeric confidence metric (0.0 to 1.0) so users know when to trust the answer.

Edge cases are documented and tested. Two times exactly 12 hours apart (00:00 and 12:00, or 06:00 and 18:00) sit at opposite ends of the circle; their vector sum is zero, R = 0, and the mean is mathematically undefined. The calculator flags this case rather than silently picking one of the two antipodes. Adding a third time anywhere breaks the symmetry and recovers a definite answer.

For practitioners outside the research and dispatch worlds, the simplest mental check is this: if all your times sit within a 12-hour window, circular mean equals linear mean. If they span midnight, only circular mean is correct. The dial widget on this page makes that distinction visible — when the scatter sits on one half of the dial, circular and linear agree; when it spans both halves, the difference can be the entire 12-hour swing.