Guide
How to book a trip that survives one bad flight
Every piece of advice on this site moves the odds a few points. This is how to arrange a trip so that the flight which goes wrong anyway does not cost you the trip.
Nationally 77.4% of flights arrive on time, 1.80% are cancelled, and 8.2% of arrivals land an hour or more behind schedule. Nothing on this site changes those numbers. What it can change is how much a bad outcome costs you when one lands on your itinerary.
This guide is about structure rather than selection: how to arrange a trip so that a single failure is absorbed instead of amplified.
Everything on one ticket
The single most expensive mistake in US air travel is booking two legs separately.
If the first airline delivers you late and you miss a flight on a separate ticket, the second airline owes you nothing. Not a rebooking, not a refund, not a discounted seat. It will sell you the next available seat at the walk-up fare, which on the day is the most expensive fare that exists.
On one ticket, a missed connection is the airline's problem to solve. That difference is worth far more than any few points of on-time rate, and it costs nothing to arrange.
Buffer where it does the most work
Ninety minutes rather than sixty, and early in the day rather than late.
The reason ninety is the number: nationally 8.2% of arrivals land an hour or more late, and a sixty-minute connection does not survive that distribution while a ninety-minute one mostly does. Going from ninety to three hours buys very little more, because what remains is cancellations, which no layover survives.
The reason early matters: a missed morning connection has a whole day of alternatives behind it. A missed evening connection has a hotel.
A spare day for anything that cannot slip
A wedding, a cruise departure, a job interview, a non-refundable onward booking. Fly the day before.
This is the only advice on the site that removes risk rather than reducing it. No route in this data reaches 95% on time. The best you can buy through good selection is better odds, and better odds are not a guarantee. A spare day is.
Frequency is your insurance
Before booking, count the daily departures your carrier operates on the route. That number is your recovery capacity and it appears in no DOT statistic.
Six or more: a cancellation costs hours. Three: same-day recovery only if you are early in the queue. One or two: a cancellation can cost a day.
Then check whether a second carrier flies the route at all. Two airlines flying it four times each is structurally safer than one flying it eight times, because a problem grounding one rarely grounds both.
Know the two rules before you need them
A cancelled flight entitles you to an automatic refund to your original payment method if you choose not to travel, regardless of fare type, and the same applies to a domestic delay of three hours or more. That is law, not courtesy.
A hotel, a meal or rebooking on a competitor is not required by any rule. Those come from each airline's own published customer service commitment, which is binding once published and differs meaningfully between carriers. The rights guide covers both in detail.
The order that matters
Non-stop over connecting. One ticket over two. Morning over evening. Ninety minutes over sixty. High-frequency route over thin. Spare day over everything, when something genuinely cannot slip.
Carrier choice and airport choice come after all of those, and the route pages here are where you make them.
What to do on the day it goes wrong
The preparation above matters most in the twenty minutes after something breaks, so it is worth knowing what those twenty minutes should look like.
Rebook before you move. Open the app or call while everyone else is walking toward the service desk. Seats are allocated in the order they are claimed, not in the order people reach the counter, and the queue you can see is the competition.
Decide in advance what your fallback is. Driving, waiting, or abandoning the trip are all reasonable answers, and choosing between them under pressure at eleven at night produces worse decisions than choosing between them at home. If you know that a six-hour delay means you drive, the decision takes no time at all.
Keep the receipts if your carrier's published commitment covers a meal or a hotel for a delay within its control. And check your credit card's trip delay coverage before travelling: for many people it is worth more than anything the airline is obliged to provide, and it is the most commonly forgotten protection there is.
What this cannot do
None of this makes a flight more likely to be on time. The base rates are the base rates, and a site built on historical statistics can only ever move odds.
What the structure above does is bound the downside. A trip built on one ticket, with ninety-minute connections early in the day, on a high-frequency route, with a spare day before anything that cannot move, absorbs a failure that would otherwise end it.
That is the honest ceiling of what planning achieves, and it is a great deal more than most itineraries are built to survive.
A worked example
Two ways to book the same trip, to make the difference concrete.
The first: two separate tickets, a 55-minute connection at a congested hub, the last departure of the evening, on a route flown twice daily, arriving the morning of the event. Every one of those five choices is individually defensible on price. Together they mean a single delayed inbound ends the trip, and nobody owes you anything.
The second: one ticket, a 95-minute connection, a mid-morning departure, on a route flown six times daily, arriving the day before. The fare is higher. The itinerary survives a cancellation, a two-hour delay, a missed connection, or a mechanical problem, because each of those has somewhere to go.
The gap between those two bookings is larger than the gap between the best and worst airline in this data, and it costs nothing to understand.
Check the route you are considering: Compare any US route · Atlanta (ATL) · Delta's record