Guide
Does flight distance affect delays? Less than it affects cancellations
Across 2,978 US routes, on-time rates vary by just 2.3 points between distance bands, while cancellations fall from 2.38% on the shortest to 1.45% on the longest.
A route under 250 miles and a route across the continent arrive on time at almost exactly the same rate. Across the 2,978 US routes with a page on this site, flights under 250 miles ran 77.8% on time from September 2025 to August 2026; flights of 2,000 to 3,000 miles ran 78.1%. Between the best and worst distance bands, the whole spread is 2.3 points.
Cancellations tell a different story. The shortest routes cancelled 2.38% of their scheduled flights and the longest 1.45%, a ratio of about 1.6 to one. Distance barely changes whether you arrive on time.
It does change whether you leave at all.
On time: a flat line
Show the figures as a table
| Distance | Routes | Flights | On time | Cancelled |
|---|---|---|---|---|
| Under 250 miles | 324 | 703,483 | 77.8% | 2.38% |
| 250 to 500 miles | 717 | 1,434,793 | 77.9% | 1.99% |
| 500 to 750 miles | 600 | 1,217,344 | 77.2% | 2.12% |
| 750 to 1,000 miles | 479 | 1,003,509 | 76.9% | 1.79% |
| 1,000 to 1,500 miles | 456 | 930,281 | 75.8% | 1.86% |
| 1,500 to 2,000 miles | 194 | 402,121 | 76.5% | 1.30% |
| 2,000 to 3,000 miles | 208 | 383,584 | 78.1% | 1.45% |
On time, all route pages, weighted by flights. Distance is miles flown between the two airports. Source: US DOT Bureau of Transportation Statistics, September 2025 to August 2026.
The best band was 2,000 to 3,000 miles at 78.1% and the worst 1,000 to 1,500 miles at 75.8%. Every band lands within a few points of the national 77.3%. I expected a slope here and did not find one. For a traveller that is a non-result, and a useful one: choosing a nonstop over a shorter connecting itinerary, or the reverse, on the belief that long flights are more or less reliable is choosing on nothing.
Two forces pull against each other here. Long flights carry more schedule padding in absolute minutes, because airlines build block times to cover a bad-wind day, and a flight that leaves 25 minutes late on a five-hour sector can often make up enough of it in the air to arrive inside the 15-minute DOT threshold. Transcontinental schedules also include red-eyes, which depart into empty night-time airspace and are usually the aircraft's first flight of a new day.
Long flights also have more to go wrong. They cross more weather and more air traffic sectors, and they tend to link big hubs at both ends. Short flights have less padding and no time to recover, but many of them run between smaller airports or fly early-morning rotations. The two effects roughly cancel. In this data the weakest band is a mid-length one, consistent with the typical hub-to-hub leg: long enough to link two congested airports, too short to carry much slack.
Cancellations: short routes go first
Show the figures as a table
| Distance | Routes | Flights | On time | Cancelled |
|---|---|---|---|---|
| Under 250 miles | 324 | 703,483 | 77.8% | 2.38% |
| 250 to 500 miles | 717 | 1,434,793 | 77.9% | 1.99% |
| 500 to 750 miles | 600 | 1,217,344 | 77.2% | 2.12% |
| 750 to 1,000 miles | 479 | 1,003,509 | 76.9% | 1.79% |
| 1,000 to 1,500 miles | 456 | 930,281 | 75.8% | 1.86% |
| 1,500 to 2,000 miles | 194 | 402,121 | 76.5% | 1.30% |
| 2,000 to 3,000 miles | 208 | 383,584 | 78.1% | 1.45% |
Cancelled, all route pages, weighted by flights. Distance is miles flown between the two airports. Source: US DOT Bureau of Transportation Statistics, September 2025 to August 2026.
The fall is not perfectly smooth from band to band, but the two bands above 1,500 miles sit clearly below everything shorter, and the shortest band is the highest.
The main reason is a decision airlines make on bad days. When a ground delay programme cuts an airport's arrival rate, or a storm leaves crews and aircraft out of position, the airline has to remove flights from the schedule. It removes the ones that are cheapest to lose. A 200-mile hop flown six times a day is the obvious candidate: its passengers can be moved to the next departure, and some can drive. A 2,500-mile flight is not. Its passengers have no alternative that day, often no seats on the later flights, and the aircraft is needed at the far end for the next morning's schedule.
The operator adds to it. A large share of short routes are flown by regional carriers with 50- to 76-seat jets, and regional carriers cancel more than most: PSA cancelled 4.80% of its flights and Republic 4.55%, against 1.89% nationally. Many of those short routes also feed the slot-controlled airports where cancellations are concentrated. LaGuardia (LGA) cancelled 5.04% of its departures and Washington National (DCA) 3.99%.
The busiest congested corridors show the combination. Washington National to Boston, around 400 miles, cancelled 4.34% of its flights, and LaGuardia to O'Hare, about 730, cancelled 5.45%. Frequent service and congested airports at both ends make those flights the first to go on a bad day.
Distance tells you little about a particular route
One routeMedian of routes at that distance
2,978 routes. Rates below 40% are drawn on the bottom edge. Source: US DOT Bureau of Transportation Statistics, September 2025 to August 2026.
Each dot is one route, and the spread inside every band dwarfs the difference between bands. Short routes include some of the best and some of the worst records in the country. Honolulu to Kahului, about 100 miles, ran 84.1% on time. Los Angeles to San Francisco, a bit over 300, ran 60.1%. The highlighted transcontinental pair, JFK to Los Angeles and the return, ran 75.4% and 77.5%, sitting inside a long-distance cloud that is as wide as any other band's. They are two dots, not one, because the pipeline keeps each direction of a route as a separate record, and the two directions of the same pair can differ by ten points.
What sets a route's record is what sits at either end of it: the airports, the carrier, the departure times on offer and the local weather. Hawaiian inter-island routes run well because they are short hops in stable air between uncongested fields. Routes into San Francisco run badly, likely because of the airport's arrival limits in low cloud, and worse still from April to July 2026, when several of them fell to roughly 35 to 50% on time. The data does not say why. Neither has anything to do with miles.
The same goes for cancellations. JFK to Los Angeles cancelled 2.78% of its flights against 1.45% for its distance band, because one end is JFK, which cancelled 3.98% of all its departures.
What the bands leave out
The bands average routes with a page on this site, each weighted by its flights, which means routes with at least 365 flights in the year. Thin seasonal routes are not in them, and neither are the few routes over 3,000 miles, which are too few to form a band. Each band also mixes very different operations: a 1,200-mile hop between two hubs and a 1,200-mile leisure flight to Florida land in the same column. One year of data is one set of storms. A winter that hits the Northeast shuttles harder, or spares them, would move the short-route cancellation figure more than any other.
What I take from the bands when booking
Do not choose a route for its length. A long nonstop is not riskier than a short one, and a short one is not safer. The mileage is the last thing I would look at. Look up the route itself: every route page gives its own on-time and cancellation rate.
On short routes, plan for cancellation rather than delay. If a short hop is cancelled, there is usually a later flight; check how many daily departures the route has before relying on that. The short-haul shuttle guide covers the corridors where frequency makes this work.
On short routes into LaGuardia or Washington National, keep a ground option in mind. Between Washington, New York and Boston, the train is a real alternative when a ground delay programme starts thinning the schedule, and airlines know it, which is likely part of why those flights get cut.
On long routes, the risk is a late departure, not a cancellation. That makes the departure time the lever. Red-eyes and early-morning departures leave before the day's delays build up. The transcontinental routes guide compares the coast-to-coast options.
Read a cancellation rate with the route's length in mind. A short route at 2.38% is ordinary for its band. A long route at the same rate is not. The cancellation rate guide explains how to read the figure.
Check the routes: JFK to LAX · HNL to OGG · LAX to SFO · DCA to BOS · Compare any US route