The obvious trade is money against time: a connection costs less and takes longer. That framing is incomplete, because it leaves out the part that actually ruins trips.
The third variable
A nonstop flight has one chance to go wrong. A connection has several, and they compound — the inbound can be late, the connection can be tight, the gate can move, and a missed connection puts you at the mercy of seat availability on the next departure.
You are not trading three hours for $120. You are trading three hours and a meaningful chance of a much worse day for $120. Sometimes that’s still the right trade.
When a connection is genuinely fine
- Generous layovers. A longer connection absorbs a delayed inbound. Very short ones look efficient and leave nothing spare.
- Same airline, one ticket. If the airline causes you to miss the connection, rebooking is normally their problem. Separate tickets on separate carriers make it yours.
- Not the last flight. A missed connection at midday is an inconvenience. At night it’s a hotel.
- Hub weather. Connecting through an airport prone to weather disruption in the relevant season raises the risk independently of the schedule.
When to pay for nonstop
- You have something fixed and unmissable at the other end.
- The itinerary is already long enough that a bad day would be genuinely costly.
- The saving is small. Below some threshold — everyone’s differs — the risk isn’t worth it.
The part search engines handle poorly
Price sorting treats a tight connection and a comfortable one as equivalent so long as the fare matches. The layover duration is displayed, but nothing weighs whether it is sensible for that airport at that time.
Constraints like “nonstop” also tend to get lost between searches, because a form throws away everything you didn’t re-enter. That’s why keeping constraints like nonstop as part of the plan matters more than it sounds — the requirement should survive the conversation rather than being restated every time.