How far in advance should you book a flight? It depends on the route

Ask how far ahead to book a flight and you will usually get a number: some count of days or weeks, delivered with the confidence of a law of physics. We don't publish one, and it is worth explaining why, because the reason is also the most useful thing to know about timing a booking.

Why one number can't be the answer

A single "best time to book" figure is an average across an enormous number of routes, months and fares. Averages like that are real, but they describe no route in particular. A short domestic hop that a dozen flights a day serve behaves differently from a long-haul leisure route with one daily departure, and both behave differently in July than in February.

It is the same problem we described in what "26% below typical" means: compare a fare against flights nothing like it and you get a confident number that answers a question you didn't ask. A booking rule built from every route at once is that comparison, made about timing instead of price.

What actually happens to a fare as departure gets closer

Airlines sell seats in fare buckets — a handful at the lowest advertised price, more at the next level up, and so on. Far from departure, the cheap buckets are mostly open and fares tend to wander up and down without much direction. As the date approaches and the cheaper buckets sell, the lowest available fare starts to step upward, and closer still it tends to rise faster. The mechanics are in the "searching raises the price" myth.

The part that varies is where that turn happens. On one route it might come well before departure; on another, cheap seats might last much longer. That turn — the point where a route's fares stop drifting and start climbing — is the thing actually worth knowing, and it is a property of the route, not of air travel in general.

How we find that point for a route

We call it the route's rise point, and it is computed from fares we have recorded ourselves, by code, the same way every time:

1. Every recorded fare is filed by how far ahead of departure it was seen, in windows of 0–3 days, 4–7, 8–14, 15–21, 22–30, 31–45, 46–60, 61–90, 91–120, and more than 120.
2. We take the median fare in each window.
3. We find the cheapest window, then walk from it toward departure.
4. The first window whose median sits 10% or more above that cheapest one is where we say this route's fares have started to climb.

Two conditions stop us from claiming a rise point at all: each window needs at least 8 recorded fares before its median counts, and the route needs at least 3 windows that qualify. Below that, the honest answer is that we don't know yet — so we say nothing, and the timing part of the score is held at neutral rather than filled with a guess. That is the same instinct behind why we refuse to score some routes.

What to do with that

Book before the rise point, and inside that window let the fare decide, not the calendar. A genuinely good fare well before the rise point is a better buy than an ordinary one on whatever day a rule told you to wait for. That is why timing is only 30% of the Price Hash Score and where a fare sits against the route's own history is the other 70%: timing tells you whether the window is still open, but only the price tells you whether this particular fare is worth taking.

Two edges worth knowing:

• Very early is not automatically best. More than about six months out, a route's schedule is barely priced and fares can still fall, so we never give a fare that early full marks for timing.
• Past the rise point, waiting rarely helps. Fares can and do drop occasionally after the turn, but you would be betting on an exception. We look at that bet in do flight prices drop closer to departure?

The short answer

There is no universal number of days. There is a point on each route where fares start to climb, it can be measured once enough fares have been recorded, and the goal is to book a good fare before it. If you'd rather not track that yourself, give us the route and we'll watch it, and email you only when a fare is worth it.