The Diversion Trap: Why Diversions Tell You Something On-Time Rates Don't

A flight that lands in Providence instead of Boston can still count as on-time. Under 14 CFR § 234.2, the federal on-time definition measures whether a flight arrives less than 15 minutes after its scheduled arrival time — it does not require that the flight arrive at its scheduled airport. If a diverted flight touches down at an alternate within the 15-minute window, or if the airline later gets it to the original destination inside that window, the on-time math doesn't automatically flag the detour. This is a narrow, real gap in the 14 CFR § 234.2 definition, and it's why on-time rate alone can't tell you which airlines quietly reroute passengers versus which ones just run late honestly.

What a diversion actually is

A diversion is a flight that lands somewhere other than its scheduled destination — typically because of weather, mechanical issues, medical emergencies, or airport closures at the original endpoint — before either terminating there or continuing on to the original destination. The DOT's on-time performance filings, drawn from the Marketing Carrier On-Time Performance dataset (table 236), track diversions as a distinct operational outcome alongside on-time rate and cancellation rate. That separation matters: a diversion is not a delay and not a cancellation. It's a third outcome, and it's reported separately for a reason — folding it into a single on-time percentage would hide exactly the information a traveler needs.

Why on-time rate alone misses the diversion signal

On-time rate answers one question: did the flight arrive within 15 minutes of schedule? It does not answer a second, related question: did the flight arrive where you expected it to? A carrier that diverts a flight to a nearby airport and gets passengers there quickly can, in principle, post a perfectly respectable on-time number for that flight while the traveler experiences a bus transfer, a missed connection, or an unplanned overnight. This is the same structural issue explored in why cancellation rate matters more than on-time rate when choosing an airline — a headline reliability number can look strong while a different operational metric, reported separately, tells the real story of what happens when things go wrong.

The fix is the same in both cases: read the numbers side by side rather than in isolation. The Airline Head-to-Head tool does this directly, putting on-time arrival rate, cancellation rate, and diversion rate for any two U.S. airlines next to each other, pulled from the same DOT on-time performance filings. Comparing two carriers this way surfaces a pattern that on-time rate by itself cannot: one airline might accept more delays and hold flights until they can land at the correct airport, while another might divert more readily to protect its on-time number, or vice versa. Neither behavior is visible from on-time rate alone.

What the federal filings measure — and what they scope to

The DOT's carrier-level on-time performance data, covering the trailing 12 months from 2025-07 through 2026-06 in the filings this site serves, reports on-time arrival rate and cancellation rate for each of ten reporting carriers: Hawaiian (81.5% on-time, 0.7% cancellation rate), Delta (81.3%, 1.9%), Alaska (80.1%, 1.4%), United (78.5%, 1.6%), Southwest (76.6%, 0.9%), American (76.3%, 2.7%), Allegiant (74.1%, 0.6%), Frontier (73.2%, 1.9%), Spirit (73.1%, 3.7%), and JetBlue (72.7%, 2.6%). These figures come from the Marketing Carrier On-Time Performance filings (table 236) and cover U.S. domestic operations of reporting carriers only — no international routes, no live tracking, no predictive claim about any specific future flight.

Notice the pattern across cancellation rate: Spirit's 3.7% and American's 2.7% sit well above Allegiant's 0.6% and Hawaiian's 0.7%, even though the on-time rates themselves don't move in lockstep with those cancellation figures. That's the same lesson diversions teach, one level down: two carriers can post similar on-time numbers while handling operational disruption in very different ways — one canceling more, one presumably absorbing more delay-minutes instead. The honest way to compare carriers is to hold all three figures — on-time rate, cancellation rate, and diversion rate — in view at once, which is exactly what a head-to-head comparison is built to do.

A worked example: reading the three numbers together

Suppose you're deciding between two carriers for a route where both operate. Pulling up their records in Airline Head-to-Head for the trailing 12-month window (2025-07 through 2026-06) gives you all three figures at once — arrival on-time rate, cancellation rate, and diversion rate — rather than forcing you to hunt them down separately or, worse, judge a carrier by on-time rate alone. If Carrier A shows a lower on-time rate but a lower diversion rate too, that combination suggests a carrier that tends to run late but still gets you to your actual destination. If Carrier B shows a higher on-time rate paired with a higher diversion rate, that combination suggests a carrier that protects its on-time number in part by rerouting more often when things go wrong. Neither pattern is visible from the on-time percentage by itself — it only shows up when you read the numbers side by side.

This is also connection-sensitive: a flight that diverts and later completes its trip within the 15-minute window shows up as on-time in the record even if a connecting flight was missed in the process. If your itinerary involves a tight layover, it's worth checking the inbound flight's history separately using a connection risk tool, since a diversion or a late-arriving aircraft upstream is exactly the kind of event that on-time rate alone won't flag.

None of this is predictive. These are historical base rates over a stated 12-month window, not a forecast for any specific flight, and they say nothing about non-U.S. routes or carriers outside the ten reporting in this dataset. This guide is informational only and isn't a substitute for professional travel or legal advice.

Frequently asked questions

Does a diverted flight count as delayed under the federal definition?

Not necessarily — the 14 CFR § 234.2 on-time definition only measures whether the flight arrives less than 15 minutes after its scheduled arrival time, so a diverted flight that still lands within that window, at the alternate or eventually at the original destination, can be recorded as on-time.

Where can I find an airline's diversion rate?

The DOT's Marketing Carrier On-Time Performance filings (table 236) report diversions as a distinct category alongside on-time and cancellation figures, and the Airline Head-to-Head tool surfaces diversion rate directly when comparing two carriers.

Is a high on-time rate always a good sign?

Not on its own — on-time rate should be read alongside cancellation rate and diversion rate, since a carrier can post a strong on-time percentage while canceling or diverting more flights than a competitor with a lower on-time number.

Do these figures apply to international flights?

No — the scope here is U.S. domestic operations of reporting carriers only, based on the DOT filings for 2025-07 through 2026-06, with no live tracking and no predictive claim about any specific flight.