FlyerIntel Research
The Day After the Holiday
Every day around the five holiday windows, set against ordinary days of the same weekday in the same season.
Published October 12, 2026 · Source period July 2023 – July 2026
Summary
Thanksgiving Day was the easiest day: 93.2% of completed flights arrived on time against 80.8% on ordinary days of the same weekday, 12.4 points above the expected rate, and it beat expectation in 3 of 3 years. Dec 28 in the Christmas window was the hardest day: 66.2% of completed flights arrived on time against 78.0% expected, 11.8 points below it, and the day was below expectation in 3 of 3 years. Thanksgiving and Christmas each had a worst day below expectation in every year of the data.
Key findings
- 2,484,712 departures across 5 holidays were compared, day by day, with ordinary days of the same weekday in the same season.
- Thanksgiving Day was the easiest day: 93.2% of completed flights arrived on time against 80.8% on ordinary days of the same weekday, 12.4 points above the expected rate, and it beat expectation in 3 of 3 years.
- Dec 28 in the Christmas window was the hardest day: 66.2% of completed flights arrived on time against 78.0% expected, 11.8 points below it, and the day was below expectation in 3 of 3 years.
- Thanksgiving and Christmas each had a worst day (the Sunday after Thanksgiving and Dec 28 in the Christmas window) more than 3 standard errors below expectation in every year of the data, and in each the worst day came after the holiday.
- Of the 5 holidays, 4 had a lower average on-time rate against expectation in the days after the day itself than in the days before.
- The largest excess of cancellations was the Saturday after Thanksgiving (2.68% against 0.73% expected), above expectation in 1 of 3 years.
What was compared
The usual warning is about the day before a holiday. This report follows every day of the 5 holiday windows, 2,484,712 departures in all, and sets each one against ordinary days of the same weekday in the same season. That matters because a holiday falls on a different weekday from year to year, and because the ordinary pace differs from one season to the next. Each day’s expectation is worked out airport by airport, so a day is not blamed for the airports it happens to be busy at.
The days before, the day itself and the days after
| Holiday | Days before: average points | The day itself: points | Days after: average points | Worst day | Points | Years below expected |
|---|---|---|---|---|---|---|
| Thanksgiving | -2.2 | +12.4 | -5.4 | Sunday after | -11.6 | 3 of 3 |
| Christmas | -2.2 | +7.8 | -4.6 | Dec 28 | -11.8 | 3 of 3 |
| Fourth of July | +3.4 | +12.2 | +0.8 | Jul 7 | -0.6 | 2 of 4 |
| Memorial Day | -0.1 | -3.1 | -3.2 | Tuesday after | -3.2 | 2 of 3 |
| Labor Day | +1.3 | +4.5 | +3.8 | Friday before | -2.5 | 2 of 3 |
Points are the on-time arrival rate minus the expected rate, weighted by completed flights. A day before or after is any day in the holiday window other than the day itself. The worst day is the day furthest below its expected rate.
The holiday itself
Thanksgiving Day, Christmas Day, Fourth of July, and Labor Day beat the expected on-time rate in every year of the data (4 of 5 holidays). Thanksgiving Day was the easiest day: 93.2% of completed flights arrived on time against 80.8% on ordinary days of the same weekday, 12.4 points above the expected rate, and it beat expectation in 3 of 3 years. The holiday itself was the best day of its window for 3 of 5 holidays. A year of Thanksgiving Day carried 9,337 departures at these airports, against 16,444 on each of the other days of its window.
The way home
Thanksgiving and Christmas each had a worst day (the Sunday after Thanksgiving and Dec 28 in the Christmas window) more than 3 standard errors below expectation in every year of the data. In each the worst day came after the holiday. Dec 28 in the Christmas window was the hardest day: 66.2% of completed flights arrived on time against 78.0% expected, 11.8 points below it, and the day was below expectation in 3 of 3 years. Of the 5 holidays, 4 had a lower average on-time rate against expectation in the days after the day itself than in the days before it (Thanksgiving, Christmas, Fourth of July, and Memorial Day).
Thanksgiving: each day against ordinary days
| Day | Years | Flights | Arrived on time | Expected | Points | Years below expected | Cancelled | Expected |
|---|---|---|---|---|---|---|---|---|
| Monday before | 3 | 50,149 | 82.1% | 81.6% | +0.5 | 1 of 3 | 0.42% | 1.00% |
| Tuesday before | 3 | 50,817 | 81.9% | 86.1% | -4.2 | 3 of 3 | 0.19% | 0.43% |
| Wednesday before | 3 | 51,180 | 82.4% | 85.2% | -2.8 | 2 of 3 | 0.12% | 0.47% |
| Thanksgiving Day | 3 | 28,011 | 93.2% | 80.8% | +12.4 | 0 of 3 | 0.05% | 0.78% |
| Friday after | 3 | 44,123 | 86.7% | 80.6% | +6.1 | 0 of 3 | 0.14% | 1.03% |
| Saturday after | 3 | 49,243 | 76.3% | 83.5% | -7.2 | 2 of 3 | 2.68% | 0.73% |
| Sunday after | 3 | 52,565 | 67.1% | 78.7% | -11.6 | 3 of 3 | 1.24% | 1.25% |
| Monday after | 3 | 51,670 | 75.5% | 81.7% | -6.2 | 3 of 3 | 0.77% | 1.00% |
| Tuesday after | 3 | 44,920 | 79.6% | 86.0% | -6.4 | 2 of 3 | 0.34% | 0.43% |
Departures from the covered airports in the holiday window. Expected is the airports’ own rate on ordinary days of the same weekday within 28 days of the holiday, applied to each airport’s flights that day. Years below expected counts the years in which the day’s on-time rate was under expectation.
Christmas: each day against ordinary days
| Day | Years | Flights | Arrived on time | Expected | Points | Years below expected | Cancelled | Expected |
|---|---|---|---|---|---|---|---|---|
| Dec 19 | 3 | 48,681 | 74.0% | 80.3% | -6.3 | 2 of 3 | 1.25% | 1.74% |
| Dec 20 | 3 | 48,314 | 72.3% | 79.5% | -7.2 | 2 of 3 | 1.20% | 1.59% |
| Dec 21 | 3 | 48,558 | 74.8% | 77.9% | -3.1 | 1 of 3 | 0.84% | 1.56% |
| Dec 22 | 3 | 49,449 | 79.2% | 76.2% | +3.0 | 0 of 3 | 0.32% | 2.22% |
| Dec 23 | 3 | 48,369 | 79.9% | 79.1% | +0.8 | 1 of 3 | 0.52% | 1.86% |
| Dec 24 | 3 | 38,711 | 79.6% | 80.1% | -0.4 | 1 of 3 | 1.32% | 1.75% |
| Christmas Day | 3 | 39,258 | 88.5% | 80.7% | +7.8 | 0 of 3 | 0.64% | 1.54% |
| Dec 26 | 3 | 48,812 | 69.7% | 80.4% | -10.7 | 3 of 3 | 3.36% | 1.72% |
| Dec 27 | 3 | 48,000 | 68.0% | 79.5% | -11.5 | 3 of 3 | 1.54% | 1.58% |
| Dec 28 | 3 | 48,015 | 66.2% | 78.0% | -11.8 | 3 of 3 | 1.63% | 1.57% |
| Dec 29 | 3 | 48,860 | 67.3% | 76.2% | -9.0 | 2 of 3 | 1.43% | 2.21% |
| Dec 30 | 3 | 48,139 | 78.3% | 79.1% | -0.8 | 1 of 3 | 0.43% | 1.85% |
| Dec 31 | 3 | 40,989 | 82.2% | 80.0% | +2.3 | 1 of 3 | 0.36% | 1.75% |
| Jan 1 | 3 | 42,653 | 81.6% | 80.7% | +0.9 | 1 of 3 | 0.43% | 1.53% |
| Jan 2 | 3 | 48,187 | 79.9% | 80.3% | -0.4 | 2 of 3 | 0.58% | 1.73% |
| Jan 3 | 3 | 46,778 | 74.6% | 79.5% | -5.0 | 3 of 3 | 1.45% | 1.57% |
| Jan 4 | 3 | 46,467 | 75.3% | 77.9% | -2.5 | 2 of 3 | 1.01% | 1.58% |
| Jan 5 | 3 | 47,073 | 75.3% | 76.2% | -0.9 | 2 of 3 | 2.90% | 2.21% |
Departures from the covered airports in the holiday window. Expected is the airports’ own rate on ordinary days of the same weekday within 28 days of the holiday, applied to each airport’s flights that day. Years below expected counts the years in which the day’s on-time rate was under expectation.
Cancellations are one storm at a time
The day with the largest excess of cancellations was the Saturday after Thanksgiving: 2.68% of scheduled departures cancelled against 0.73% expected. It was above expectation in 1 of 3 years. The on-time shortfalls above recur year after year; this cancellation excess does not, and a total over the years hides which year it came from.
What this does not show
The data holds 3 to 4 years of each holiday, a small number, and a holiday that fell in a storm in one of them leaves a mark on its average. The years-below column is there so a day that was bad once can be told from one that was bad every time. The expectation uses ordinary days within 28 days of the holiday, and an ordinary day that was itself a bad storm day lowers it. The comparison covers departures from covered airports and says nothing about how full the flights were.
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/. The holiday windows are the five the site defines (Thanksgiving, Christmas to New Year, the Fourth of July, Memorial Day weekend and Labor Day weekend); the holiday itself is the reference day, and days before and after are counted from it. Each day is read from the day-by-airport holiday table. The expectation is, for each airport, its on-time arrival rate on ordinary days of the same weekday within 28 days either side of the holiday (days outside every holiday window) in the years loaded, applied to the airport’s completed flights on the day; the same is done for cancellations over scheduled departures. Points are the observed on-time rate minus the expected rate. A standard error is the binomial variance of the expected on-time count.
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. Years of data differ by holiday, and the Years column shows how many each day has. The standard error assumes independent flights, which a storm is not, so it is a floor.
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