FlyerIntel Research
The Worst Departure Hours at the Biggest Hubs
Sunday 22:00 at MCO is the weakest slot at the 10 largest airports.
Published October 18, 2026 · Source period July 2023 – July 2026
Summary
51.9% of flights scheduled to leave MCO in the Sunday 22:00 hour landed on time, the lowest of any slot at the 10 largest airports. The worst slot fell in the evening at 10 of them.
Key findings
- MCO Sunday 22:00: 51.9% on time over 1,706 scheduled flights, against 74.8% for the airport’s slots as a whole.
- MCO has the widest spread: 94.6% on time in its best slot (Tuesday 05:00) and 51.9% in its worst, 42.7 points apart.
- The worst slot starts at 17:00 or later at 10 of 10 airports; Sunday is the weekday of the worst slot at 8.
- At 10 of 10 airports the worst slot is more than 2 standard errors below the airport’s other slots.
One slot out of every week
An airport’s on-time rate averages over every hour of every day of the week. This report takes the 10 covered airports with the most scheduled departures and, for each, ranks every weekday-and-hour slot that has at least 500 scheduled flights. The least reliable slot is named beside the most reliable one. A slot is the day of the week and the scheduled local hour a flight was due to leave, and the rate says whether flights booked into it landed within 15 minutes of schedule. MCO has the weakest single slot, Sunday 22:00. MCO has the widest distance between its best and worst slot.
The worst and best slot at each large airport
| # | Airport | City | On time, all slots | Worst slot | On time in worst slot | Flights in worst slot | Cancelled in worst slot | Best slot | On time in best slot | Spread, points |
|---|---|---|---|---|---|---|---|---|---|---|
| 1 | MCO | Orlando | 74.8% | Sunday 22:00 | 51.9% | 1,706 | 2.4% | Tuesday 05:00 | 94.6% | 42.7 |
| 2 | DFW | Dallas–Fort Worth | 73.1% | Sunday 19:00 | 56.9% | 10,416 | 3.6% | Thursday 05:00 | 91.0% | 34.1 |
| 3 | CLT | Charlotte | 74.8% | Sunday 19:00 | 57.8% | 2,738 | 5.0% | Thursday 06:00 | 93.4% | 35.6 |
| 4 | LAS | Las Vegas | 78.0% | Thursday 18:00 | 58.3% | 4,345 | 1.5% | Thursday 05:00 | 95.6% | 37.3 |
| 5 | DEN | Denver | 76.0% | Thursday 22:00 | 58.5% | 2,326 | 2.9% | Wednesday 05:00 | 93.4% | 34.9 |
| 6 | ORD | Chicago | 74.5% | Sunday 22:00 | 59.8% | 1,945 | 4.9% | Wednesday 05:00 | 91.2% | 31.4 |
| 7 | ATL | Atlanta | 78.7% | Sunday 20:00 | 63.4% | 8,536 | 2.7% | Wednesday 06:00 | 91.3% | 27.9 |
| 8 | PHX | Phoenix | 80.3% | Sunday 19:00 | 67.0% | 4,464 | 0.9% | Monday 05:00 | 95.3% | 28.3 |
| 9 | SEA | Seattle | 79.1% | Sunday 22:00 | 68.3% | 2,676 | 1.4% | Wednesday 05:00 | 94.2% | 25.9 |
| 10 | LAX | Los Angeles | 81.4% | Sunday 20:00 | 70.2% | 2,784 | 1.1% | Sunday 05:00 | 95.9% | 25.7 |
Rates are flights that landed within 15 minutes of schedule over flights that completed, summed from slot counts. Cancelled is cancelled flights over scheduled flights in the same slot. The all-slot rate counts only slots above the floor. Ordered by the rate in the worst slot.
The weakest slots across all of them
| # | Airport | Slot | On time | Completed flights | Cancelled |
|---|---|---|---|---|---|
| 1 | MCO | Sunday 22:00 | 51.9% | 1,662 | 2.4% |
| 2 | MCO | Sunday 20:00 | 52.4% | 4,070 | 3.6% |
| 3 | MCO | Sunday 18:00 | 53.3% | 4,258 | 3.1% |
| 4 | MCO | Sunday 19:00 | 53.5% | 4,897 | 3.6% |
| 5 | MCO | Sunday 21:00 | 54.5% | 2,687 | 3.7% |
| 6 | MCO | Sunday 17:00 | 54.8% | 4,632 | 2.9% |
| 7 | MCO | Sunday 16:00 | 56.4% | 4,159 | 2.4% |
| 8 | DFW | Sunday 19:00 | 56.9% | 10,000 | 3.6% |
| 9 | DFW | Sunday 20:00 | 57.5% | 7,935 | 4.1% |
| 10 | CLT | Sunday 19:00 | 57.8% | 2,596 | 5.0% |
| 11 | MCO | Monday 22:00 | 57.9% | 1,375 | 1.2% |
| 12 | MCO | Saturday 20:00 | 58.0% | 3,587 | 1.9% |
The lowest on-time rates among every slot above the floor at these airports.
What this does not show
The least reliable of more than a hundred slots is partly the noisiest, which is why a slot needs 500 flights to be named and why the key findings say how many worst slots stand clear of their airport by 2 standard errors. Even then the flights in a slot share aircraft and weather, so the test is a guard against naming a tie, not an interval. The grid has no date, so a slot that looks bad because of a few storms or a holiday weekend cannot be separated from one that is bad every week. Rates describe flights that landed, grouped by the airport where they left; a late arrival can come from the destination as much as from the hub. Every airport among the largest cleared the floor.
Pages for what this report covers
Airports: Orlando (MCO) · Dallas–Fort Worth (DFW) · Charlotte (CLT) · Las Vegas (LAS) · Denver (DEN) · Chicago (ORD) · Atlanta (ATL) · Phoenix (PHX) · Seattle (SEA) · Los Angeles (LAX)
Methodology
Calculated by FlyerIntel from 21,708,321 individual flight records in the US Department of Transportation Bureau of Transportation Statistics Reporting Carrier On-Time Performance dataset, covering July 2023 through July 2026. Following BTS methodology, a flight counts as delayed when it arrives 15 or more minutes after its scheduled arrival time. Cancelled and diverted flights have no arrival time and are excluded from on-time rates rather than counted as late; cancellation rates are reported separately. Full definitions are published at /methodology/. Airports are the 10 covered airports with the most scheduled departures in the weekday-by-hour grid, which places each departure by day of week and scheduled local hour at its origin. Slots under 500 scheduled flights are dropped before anything is ranked or summed, and an airport needs 60 slots above the floor. A worst slot is called clear when its rate sits 2 or more standard errors below the rate of the airport’s other slots above the floor (a two-proportion z).
Limitations
The source dataset covers scheduled domestic flights operated by carriers above the DOT reporting threshold; smaller carriers and international flights are not included. Flights sold under a mainline brand but flown by a regional carrier are attributed to the operating carrier, which is how the DOT records them. Past performance describes what has happened and is not a prediction about any individual flight.
Sources
- FlyerIntel analysis — FlyerIntel · source
- Reporting Carrier On-Time Performance — US Department of Transportation, Bureau of Transportation Statistics · source
Full FlyerIntel methodology · Report a problem with this analysis