Coast to coast: how the transcontinental routes actually perform

Routes of 2,000 to 3,000 miles ran 78.1% on time from September 2025 to August 2026. Routes under 250 miles ran 77.8%. On paper the coast-to-coast flights keep pace with the shuttles, and part of that is not operational. It is arithmetic, and understanding it changes how you read every long-haul number on this site.

Padding scales with distance

An on-time rate is measured against the schedule the airline filed. A five-hour flight carries more scheduled block time than a one-hour flight, and more of that time is buffer rather than flying.

The practical effect: a transcontinental departure that pushes back twenty-five minutes late has several hours of cruise in which to recover some of it, plus a schedule that already assumed some headwind. A 55-minute shuttle that pushes back twenty-five minutes late arrives twenty-five minutes late, because there is nowhere to put the time.

So when a coast-to-coast route reads a few points better than a regional hop out of the same airport, at least part of that gap is the clock rather than the operation. The long routes cross more weather and link busier airports; the padding is what lets them finish level.

What that means for reading the figures

Compare like with like. A transcontinental route's number is meaningful against other transcontinental routes, and misleading against a short-haul one.

The comparison that survives is between carriers on the same route. Two airlines flying the same coast-to-coast city pair are padding against the same taxi times, the same airspace and the same headwinds, so a gap between them is real. That is the table at the top of every route page here.

The failure mode is different too

Long routes fail less often and worse.

Because they are flown fewer times a day, often two or three rather than eight, a cancellation has thin recovery options. Nationally 1.89% of flights are cancelled; on a route with three daily departures, a cancellation on the last one means tomorrow.

The other asymmetry is direction. Routes on this site are directional for exactly this reason: my pipeline keeps each direction of a city pair as its own record, and the two legs of the same pair can differ by ten points.

Look up both directions before booking. The page for each names the other's figure.

The red-eye question

Overnight transcontinental departures perform unusually well, and the reason is structural rather than lucky.

A red-eye departs into empty airspace, lands into an empty arrival bank, and is frequently the first leg of the aircraft's new duty day rather than the last of its old one. On the flagship route the numbers bear it out: Los Angeles to JFK departures between 8pm and 5am ran 79.8% on time, while JFK to Los Angeles departures between 5pm and 8pm ran 69.6%.

The cost is obvious and personal. But if the question is purely which departure is most likely to arrive when promised, the overnight is a serious answer on these routes.

What I would book coast to coast

Take the morning departure or the overnight, not the late afternoon one. I would not book an evening westbound out of New York if the trip had any slack to lose. Nationally the 5am–8am window runs 89.3% against 68.0% for 5pm–8pm, and on long routes the evening slot compounds a full day of inherited delay.

Fly non-stop. A connection on a coast-to-coast itinerary adds a failure point to a journey that already has the fewest of any long trip.

And weight cancellation rate above on-time rate here, because the recovery options are thin and the padding has already flattered the punctuality figure.

The routes themselves are not alike

"Transcontinental" covers several quite different products, and lumping them together hides most of what matters.

The premium business routes between the New York and California metros are flown many times daily by several carriers on large aircraft with generous schedules. They have depth: a cancellation finds you another seat within hours, and the competition between carriers on identical city pairs makes the comparison table on those route pages unusually informative.

The thinner long routes, connecting a coastal hub to a mid-sized city three time zones away, are flown once or twice daily by one carrier. They post similar on-time figures and carry a completely different risk, because there is nothing behind them.

The on-time rate does not distinguish these two cases at all. Frequency does, and you have to count it yourself.

Why the eastbound leg usually reads better

I expected the westbound leg to win, on the theory that it chases the clock into a quieter afternoon. In this data it mostly does not.

JFK to San Francisco ran 71.4% on time and San Francisco to JFK 78.0%. Newark to San Francisco ran 67.9% against 76.8% the other way, and Boston to San Francisco 69.2% against 74.5%.

Three mechanisms are consistent with that. The eastbound schedule includes the red-eyes, which fly empty airspace as the aircraft's first leg of a new day. The jet stream gives eastbound flights a tailwind, so a late departure has more room to recover in the air. And the westbound legs leave the congested Northeast airports, several of them bound for San Francisco, whose arrival rate drops in low cloud.

The result is that the two directions of the same city pair are genuinely different products, and this site scores them separately for that reason.

{
 "flights": 10285,
 "completed": 9985,
 "onTimePct": 75.4,
 "veryLatePct": 9.5,
 "avgDelayMin": 77.5,
 "medianArrDelay": -7,
 "cancelPct": 2.78,
 "divertPct": 0.14,
 "causes": {
  "carrier": 0.3597,
  "weather": 0.04,
  "nas": 0.2632,
  "security": 0.0018,
  "lateAircraft": 0.3353
 },
 "byHourBlock": [84, 77.9, 79.8, 74.2, 69.6, 65.8 ],
 "byMonth": [71.9, 74.8, 87, 80.8, 80.3, 70.3, 70.5, 63.9, 83.2, 74.7, 77.3, 69.7 ],
 "byDow": [75.5, 79.7, 79, 74.1, 77.6, 73.2, 68.3 ]
}

This is the file my pipeline writes for the route each month; every JFK to Los Angeles figure above is recomputed from it, and its return leg has a file of its own.

Finally, weigh the arrival time rather than the departure time on these routes. A three-hour time difference means an eastbound red-eye lands at breakfast and an afternoon westbound lands in time for dinner, and the arrival hour determines how much of a delay you can absorb before it costs you the day. The on-time rate is the same number either way; its consequences are not.

Check specific long routes: Kennedy (JFK) · Los Angeles (LAX) · Compare any US route