How every number is measured

Your trip

How long from arriving at the stop to stepping off at your destination?

We simulate a rider arriving at the stop at every second of the chosen window (say, AM rush on a weekday). For each second, we measure the wait until the next bus that actually reaches the destination, plus that bus’s travel time to the destination stop. The typical (median), average, and 1-in-10-worst figures summarize all those simulated riders.

The scheduled version uses PRT’s published schedule. The observed version replaces both pieces with reality: bus arrivals detected from GPS, and ride times measured from the same GPS tracks, pooled across up to 42 recent matching days (same day type and time window).

The wait and ride medians shown under the total are each typical on their own, but they don’t sum to the total median — a long wait and a long ride rarely land on the same trip.

With the schedule-timed plan selected, the same total is rebuilt from the timed simulation instead: each simulated rider’s wait plus the measured ride of the bus they actually boarded — a true distribution, not a sum of medians.

The wait

If you walk up without checking an app, how long until a bus that takes you there?

Waits are measured the way riders experience them, not as half the average gap: we sample every second of the window and take the time to the next qualifying bus. Long gaps count for more because more riders are standing inside them (the “inspection paradox”) — this is why the average wait can be worse than half the scheduled frequency suggests.

A bus qualifies by geometry, not by name: it must stop at your origin and later in its run reach a stop in the destination zone. Short-turn trips that pass your stop but end before your destination are excluded automatically.

The scheduled histogram comes from the current published schedule for the chosen day type and window. The observed histogram comes from GPS-detected arrivals across up to 42 recent matching settled days, restricted to days covered by the current schedule so buses are matched against the schedule that was actually in effect.

Spacing regularity

Are the gaps even, or do buses clump?

Headway CV is the coefficient of variation of the gaps between consecutive qualifying buses — 0 means perfectly even; higher means clumpier.

The extra-wait figure compares the average wait against what a perfectly even schedule with the same number of buses would deliver: uneven spacing alone adds that many minutes to the typical wait.

Punctuality

Does the bus come when the schedule says?

Arrivals are not self-reported by PRT — they are derived from each bus’s GPS track: the track is projected onto the route’s path, and the moment it crosses each stop is recorded as an observed arrival.

Each observed arrival is graded against its scheduled time using PRT’s own on-time definition: up to 1 minute early through 5 minutes late counts as on time; earlier is “early”, later is “late”.

GPS tracking cannot time arrivals at the very start or end of a bus’s trip, so a stop close to where trips begin or end is only partly visible. Where fewer than half the buses that demonstrably ran were caught at a stop, no figure is shown for that stop — the sliver that was caught describes the tracking, not the service. Route-wide and system-wide figures are still shown, since they are dominated by mid-route stops where tracking is near-complete.

The record covers settled service days only (a day settles about three hours after service ends) and accumulates from June 29, 2026. Days that ran under a different schedule than the currently published one are excluded rather than guessed.

Headway reality

Does the bus come as often as promised?

For each route and direction at this stop, the scheduled gap is the average time the schedule promises between consecutive buses; the observed gap is the same figure measured from GPS-detected arrivals over the settled record.

“Effectively missed” counts observed buses that ran late by at least half their promised gap and by more than PRT’s 5-minute lateness threshold — for a rider who timed the promised slot, that bus is functionally a no-show. A day’s first arrival for a route and direction has no preceding gap and is never counted either way.

Two adjustments keep tracking blind spots out of these numbers. A gap between two detected arrivals is dropped when a bus that demonstrably ran elsewhere went undetected in between — that gap is an artifact of the tracking, not a wait anyone had — while a gap widened by a bus that ran nowhere is kept, because that is a real missed bus. And where fewer than half the buses that ran were caught at a stop, no per-stop figure is shown at all.

The record covers settled days from June 29, 2026, under the currently published schedule; days that ran under an older schedule are honest gaps, not guesses.

Countdown reliability

Do the arrival countdowns tell the truth?

We track every arrival countdown PRT displays for this stop. A countdown “blinks out” when it disappears from the feed while still predicting an arrival at least 2 minutes away.

Blink-outs are cross-checked against the bus’s GPS, so feed hiccups where the bus kept driving are excluded, as are normal end-of-run drop-offs near a terminus.

The share shown is vanished countdowns divided by all countdowns displayed, over the settled days on record (accumulating since July 2026).

Timing it to the schedule (Wait panel)

How early should I arrive if I plan to catch a particular scheduled bus?

Most riders on infrequent routes aim for a scheduled bus rather than showing up at random. For each origin/destination query we match every GPS-observed arrival to its own scheduled departure to get that bus’s delay. The advised head start — “arrive ~N min early” — is the smallest whole minute that would have caught at least 90% of the buses that actually ran (the 10th percentile of those delays, never less than 1 minute).

We then replay every recorded day with a simulated rider standing at each scheduled time minus that head start; they board the first bus that really arrived after that moment. A scheduled bus that left before the rider arrived, or was never seen arriving, counts as a miss — the miss rate is the share of scheduled buses missed that way, and the miss cost is the median wait those riders paid for the next real bus.

The trip card shares one toggle between the two plans — driving both the door-to-arrival total and its wait breakdown: when the scheduled average gap is over 12 minutes it defaults to the schedule-timed plan, and on frequent service it defaults to the random-arrival wait instead. The timed door-to-arrival total adds the boarded bus’s own measured ride to each simulated wait.

Which stops can take you there

Why do only some stops light up for a destination?

A stop lights up for a destination when at least one scheduled trip stops there and later reaches the destination zone — the same sequence-geometry rule the wait itself uses.