Almost everyone assumes a longer trip costs more. On most of the nonstop routes where we can check it properly, it's the other way round: staying longer is cheaper — the same flights, the same route, just a different return date.
"Properly" is doing real work in that sentence. We only count a route when we hold enough priced dates in two different trip-length bands — a long weekend against a week, say — because comparing seven nights to eight isn't comparing trip lengths at all, it's measuring noise. That leaves 262 routes of the 439 we can price across dates.
Airlines don't price the number of nights you stay. They price the seat on each flight, and the two flights that make up a short weekend trip are the two most contested departures of the week. Extend past them and you're buying a Tuesday seat instead of a Sunday one.
Why nobody publishes this
Because measuring it properly is annoying. If you simply compare the cheapest fare on a route to the dearest, you get an enormous number — and it's wrong. That comparison quietly changes two things at once: the dates and the length of the trip. A three-night trip in October and a sixteen-night trip in September are not "the same trip on different dates."
So we hold the trip length fixed. For each route we take the best-sampled stay length, and only then look at what moving the dates does. That one correction cut our own headline figure roughly in half — which is exactly why it's worth doing.
What flexible dates are actually worth
Same route, same trip length, different dates. Across 439 routes with at least 8 priced departure dates at one stay length (a median of 24 dates each):
The first row is the one to act on. It isn't the best case against the worst — that flatters us and it isn't a choice anyone makes. It's what picking the cheapest date is worth against picking an ordinary one, which is the actual decision.
Measured on August 30, 2026 against departures inside a normal booking window. Each date contributes its typical price rather than its single cheapest flight, so one odd 5am departure can't stand in for a date.
What this doesn't say
- Not "always book longer". It's a tendency across routes, and yours may be one of the third that runs the other way. The point is that the length of your trip is a price lever, not a fixed input — and almost nobody treats it as one.
- Not a claim about the cheapest month of the year. We price a rolling window a few months ahead, so we can tell you what moves inside it, not what December will do in August.
- Not every stay length equally. We price the stay lengths people actually search — weekends and week-long trips get far more coverage than three-week ones, so the mix reflects our sampling, not travel in general.
How to use it
- Give the search a range of nights, not a number. "6 to 9 nights" prices every length in that band instead of assuming seven.
- Widen the window before you widen the budget. Dates move the price more than almost anything else you control.
- Check the length you'd rejected. If ten nights is possible, price it — on most routes that's the cheaper side of the trade.
- Search a flexible window from your airport — every date combination in your window, priced nonstop and ranked cheapest-first.
- See what your route usually costs — the typical price by month and by trip length, route by route.
Related: what the cheapest day of the week is worth, and why searching multiple weekends finds cheaper flights.
The cheapest version of your trip usually isn't the one you typed into the box. It's the same route, a few days over, and a stay that's longer than you assumed you could afford.