Guide
How often a connection fails at the major US hubs
Across the 20 busiest US hubs, the median chance of missing a 60-minute connection is 15.5%. At New York (LGA) it is 24.4%. Here is the hub-by-hub table for 45, 60 and 90 minutes.
A connection fails in one of two ways. The inbound flight is cancelled or diverted, or it lands too late to make the next departure. Airlines publish minimum connection times, but those say nothing about how often the first flight is late. The DOT arrival data does.
This guide uses the same method as the site's connection risk calculator. For every route into a hub, it takes the real arrival delays of every flight that landed from September 2025 to August 2026, counts how many arrived too late for a given layover, adds the flights that were cancelled or diverted, and weights each route by how many flights it carries. I count diversions as failures on purpose. The on-time rate leaves them out entirely, and in the DOT file a flight that turns back to its departure airport is a diversion too: for a connecting passenger, that is a missed connection whatever the layover. The result is the chance that a connection at that hub fails because of the inbound flight.
Two assumptions are built in. It takes 30 minutes to get from one gate to the next, and the onward flight leaves on time. Both are cautious, and both can be changed in the calculator for your own trip.
The hub table
| Hub | 45 min | 60 min | 90 min | 2 hours | Try your own |
|---|---|---|---|---|---|
| New York (LGA) | 31.5% | 24.4% | 16.8% | 12.7% | Calculator |
| San Francisco (SFO) | 33.7% | 23.7% | 12.7% | 7.7% | Calculator |
| Washington (DCA) | 31.7% | 23.4% | 15.0% | 11.1% | Calculator |
| Boston (BOS) | 29.3% | 22.2% | 14.3% | 9.9% | Calculator |
| Chicago (ORD) | 28.6% | 21.0% | 14.2% | 10.9% | Calculator |
| Newark (EWR) | 25.0% | 19.1% | 13.2% | 10.0% | Calculator |
| Orlando (MCO) | 25.7% | 17.9% | 11.0% | 7.7% | Calculator |
| Dallas-Fort Worth (DFW) | 23.8% | 17.7% | 11.9% | 9.0% | Calculator |
| Miami (MIA) | 23.4% | 16.4% | 9.9% | 7.0% | Calculator |
| Detroit (DTW) | 21.9% | 16.1% | 10.5% | 7.6% | Calculator |
| Las Vegas (LAS) | 22.6% | 15.0% | 8.2% | 5.1% | Calculator |
| Charlotte (CLT) | 19.9% | 14.6% | 9.7% | 7.2% | Calculator |
| Denver (DEN) | 21.2% | 14.5% | 8.6% | 5.9% | Calculator |
| Phoenix (PHX) | 21.0% | 13.9% | 7.9% | 5.2% | Calculator |
| Atlanta (ATL) | 19.2% | 13.8% | 8.8% | 6.4% | Calculator |
| Los Angeles (LAX) | 20.8% | 13.5% | 7.5% | 5.0% | Calculator |
| Minneapolis (MSP) | 19.3% | 13.5% | 8.4% | 6.1% | Calculator |
| Houston (IAH) | 19.5% | 13.3% | 8.1% | 5.8% | Calculator |
| Seattle (SEA) | 20.7% | 12.7% | 6.7% | 4.4% | Calculator |
| Salt Lake City (SLC) | 15.8% | 10.5% | 6.0% | 4.0% | Calculator |
Each hub links to its airport page. The last column opens the calculator with that hub already selected, so you can pick your actual inbound and onward routes.
Across the 20 hubs, the median chance of a missed connection is 22.3% for a 45-minute layover, 15.5% for 60 minutes, 9.8% for 90 minutes and 7.1% for two hours. 5 of the hubs are at 20% or more for a one-hour connection.
The riskiest hub for a 60-minute connection is New York (LGA) at 24.4%. The safest is Salt Lake City (SLC) at 10.5%.
At 24.4%, I would not book a one-hour connection at New York (LGA) on a trip that matters.
Why the curve flattens
Look along any row and the risk falls steeply between 45 and 90 minutes, then slowly after that.
That shape is the most useful thing in the table.
Most late arrivals are a little late. A flight 20 or 30 minutes behind schedule breaks a 45-minute connection and is absorbed by a 90-minute one, so adding that time removes a large share of the risk. What is left after two hours is mostly cancellations and multi-hour disruptions, and no reasonable layover survives those. That is why the two-hour column never reaches zero.
The practical reading: going from 45 to 90 minutes buys a lot. Going from two hours to three buys very little, and the extra hour would be better spent on an earlier departure or a more reliable inbound route.
Why hubs differ
Hub-to-hub differences come from a few sources the airport pages break down:
- Airspace and runways. Hubs where arrivals are regularly held by air traffic control, charged to the National Airspace System in the delay causes, carry more late arrivals for every inbound route.
- Weather. Hubs exposed to thunderstorms or winter storms have lumpy records, with a few very bad days doing most of the damage. The month-by-month pages show when.
- Inbound mix. A hub fed mostly by short regional flights has different arrival patterns from one fed by long-haul flights with padded schedules.
- Cancellation habits. A hub where carriers cancel rather than delay produces fewer late arrivals and more outright failures.
How I check a connection before booking
- Open the calculator, choose the hub, your inbound route and your onward route.
- Try the layover your ticket gives you, then 30 minutes longer. If the risk drops sharply, the longer connection is worth it. If the figure for my layover came out at one in five or worse, I would move to a later onward flight.
- Check the inbound route's own page. A route with a strong morning record and a weak evening one will look very different depending on when you land.
- If the connection is on separate tickets, treat the calculator's figure as the chance of having to buy a new ticket.
Airline minimum connection times vary by terminal and by whether you change terminals, so replace the 30-minute default with the figure your airline gives for your airport.
What the table cannot capture
These are averages over a year and over every inbound route into a hub. A particular route can be far better or far worse than its hub, and a single storm can make any hub's figure irrelevant for a day.
The onward flight is assumed to leave on time. In practice a late onward departure sometimes rescues a tight connection, especially when both flights are with the same airline at its own hub, and airlines sometimes hold a departure for connecting passengers. Neither is guaranteed, so the calculator does not count on them.
Common questions
Why does a two-hour connection still show some risk? Because cancellations and diversions break a connection whatever the layover. The two-hour column is close to the share of inbound flights that did not arrive at all, plus the small number of very long delays.
Is it safer to connect at the airline's own hub? Often, because the airline has more flights to rebook onto and sometimes holds a departure for connecting passengers. Neither shows up in this table, which assumes the onward flight leaves on time and no one waits for you.
Does the time of day change the risk? Yes, substantially. These figures average every inbound flight across the day. An early-morning arrival usually carries far less risk than an evening one, and the inbound route's page shows the on-time rate for each departure window.
What about international connections? This data covers US domestic flights only. International connections add immigration and customs time, so they need longer layovers than anything in this table suggests.
How is the 30-minute connection time chosen? It is a typical figure for a domestic transfer in the same terminal. The calculator lets you change it to your airline's published minimum for your airport.