The name is wrong, and that is the first useful thing to know. Rush hour is not an hour, and it is not one block of time.
Morning and evening rush hours have different shapes, behave differently, and respond to different tactics. Treating them as the same thing is why most people plan their commute badly.
What traffic costs, in hours
The 2025 INRIX Global Traffic Scorecard puts the average US driver's loss at 49 hours a year — roughly $894 per driver, $86 billion nationwide. The Texas A&M Transportation Institute's Urban Mobility Report is starker for commuters specifically: the average American car commuter now loses a record 63 hours a year to congestion — the highest since the data began in 1982 (Axios). In the worst metros it's far more — Chicago drivers lose 112 hours.
The two peaks are not mirror images
The morning peak is compressed. People are aiming at fixed start times, so departures cluster tightly. The build is sharp, the peak is intense, and it clears reasonably quickly once the start times have passed.
The evening peak is spread out. People leave when their work finishes, and that happens across several hours. It builds earlier, lasts longer, and drains slowly.
The practical consequence: the evening bad window is roughly twice as wide as the morning one. Most people plan for a two-hour evening peak and get caught by the extra two.
In most US cities the clock windows look roughly like this — your metro may start earlier or run longer:
- Morning: heaviest roughly 7:00–9:00 a.m. — leave before 6:30–7:00 or after 9:00–9:30.
- Evening: heaviest roughly 4:00–6:30 p.m., usually the worse of the two — leave before 3:30 or after 6:30–7:00.
Why the evening is usually slower
Three things stack, and none of them apply in the morning:
The longer window means more of your realistic departure options fall inside it.
Discretionary trips join the commute. Shopping, school runs, gym trips and social travel all happen in the evening on the same roads. The morning has almost none of this.
Events load the evening. Stadiums, arenas and venues start in the evening and load traffic two to three hours before start time — directly on top of the commute.
Direction matters more than the hour
This is the factor most people never account for, and it is larger than anything else on this page except departure timing.
Employment concentrates at one end of most corridors. That means traffic loads inbound in the morning and outbound in the evening.
If you travel against that flow — living near the employment centre and working outward — you have a materially easier commute on the same road at the same hour. The difference is frequently twenty minutes or more.
That is why two colleagues with identical mileage can describe completely different commutes. One is going with the flow and one against it.
When does rush hour end?
Later than you think in the evening, and it does not stop at a particular minute.
The evening peak drains rather than ending. Traffic thins gradually as the discretionary trips finish and the last commuters clear. There is no moment when the road becomes free.
The morning ends more sharply, because it was driven by fixed start times in the first place. Once those have passed, the reason for the peak is gone.
Practical upshot: if you are waiting out an evening peak, waiting an extra twenty minutes usually keeps paying. If you are waiting out a morning one, the return drops off quickly.
Friday is a different animal
On most corridors Friday breaks the pattern three ways:
- The evening peak starts earlier
- It lasts longer
- It stacks weekend departure traffic on top of the commuter load
If you have one flexible day a week, make it Friday. That is where flexibility buys the most.
Sunday evening is the return version of the same problem on intercity routes, as weekend travellers come home.
The day of the week matters more broadly too. TTI's national data shows Monday is the lightest weekday — flexible workers stay home — while Thursday now matches Friday as the heaviest, and Inrix finds Friday afternoon the worst window in about three-quarters of US metros. Some cities flip the pattern: in Phoenix, Thursday has overtaken Friday outright. If you're hybrid, which day you commute is a lever as big as the clock. See which day has the worst traffic — and how much time traffic actually adds.
The thing that actually changes your commute
Not the route. The departure time.
On a typical metro corridor:
| Variable | Typical spread |
|---|---|
| Best vs worst route at a given hour | 10–15 minutes |
| Best vs worst departure time on the same route | ~30 minutes |
Departure timing is worth roughly twice what route choice is worth. If you are going to optimise one thing, optimise that one.
The curve is steepest at the edges
This is the detail that makes small changes worth making.
The delay curve climbs fastest at the shoulders of the peak, not in the middle. Shifting a departure from the steep part of the climb saves considerably more than shifting one that is already near the flat top.
Which means a 15-minute change often saves more than 15 minutes — you clear a bottleneck before it saturates rather than joining the back of it.
The three-day test: leave fifteen minutes earlier than usual for three days and record your arrival time. On most corridors the saving exceeds fifteen minutes. If it does not, you are past the steep part, and the answer is a different route, a different mode, or a different work pattern rather than a different alarm.
What makes any given day worse
Weather. A corridor running near capacity has no spare headroom, so disruption compounds non-linearly. The same rain that costs five minutes at 11 a.m. can cost twenty-five at 5 p.m. The rain did not change; the road's ability to absorb it did.
The rule that follows: if rain is forecast to land on your peak, leaving early beats any route choice you can make. Getting ahead of weather is easy. Getting out of it is not. See how weather affects traffic and check weather along your route before leaving.
Incidents. Unpredictable, and worst on corridors with no alternative — a bridge, a tunnel, an elevated section with no shoulder. There, an incident does not slow the trip, it stops it.
Events. Predictable and almost universally ignored. If your route passes a venue, checking the schedule once a week is the highest-return thirty seconds available to you.
Construction. Frequently a permanent feature rather than an occasional one.
Route still matters when something breaks — see the fastest route to work and how to find alternate routes. And if a jam is building ahead, a coffee or food stop on your route that lets it clear can beat sitting in it.
The number worth tracking
If you track one thing about your own commute, do not track the average. Track the spread.
A commute that ranges from 30 to 40 minutes is a completely different life from one that ranges from 25 to 70 — even though the second has a better best case. The unpredictable one forces you to plan for 70 minutes every day, so you pay the worst case even on good days.
How to measure it: note your departure and arrival times for one week. Two columns, five rows. The gap between your best and worst day is the number worth attacking, and it is usually larger than people expect.
When timing cannot fix it
Some commutes are structural. The signs:
- Leaving thirty minutes earlier saves under ten
- Your route has a single chokepoint with no alternative
- The gap between good and bad days exceeds forty minutes
At that point the options are a different route, a different mode, or a different work pattern.
On congested urban corridors, rail is often competitive — and the trade is worth stating plainly. Transit is usually not faster than a good drive. It is the same as a bad drive, every single time. You give up the best case in order to eliminate the worst.
For a commuter whose real problem is unpredictability rather than duration, that is frequently the better deal.
Your city's windows: the local guides
National averages hide the differences that decide your drive — DC's peaks start earlier, LA's run longer, Phoenix's start at dawn. These guides cover each metro's actual corridors and windows, with the congestion data behind them:
- Chicago — the most congested US city (112 hours lost)
- New York (102) · Philadelphia (101) · Los Angeles (87)
- Boston (83) · Phoenix (76, record) · Miami (~75) · Atlanta (~75)
- Houston (70) · Washington DC (70) · Seattle (~68) · Nashville (65)
- Baltimore (63, fastest-rising) · Denver (51) · San Francisco (~49) · Charlotte (~41)
- Dallas · Austin · Orlando
- Driving between cities? Baltimore to DC: best time to drive
Before you leave
Rush hour is a pattern. Your Tuesday is not.
BoardSpy watches your specific route and direction — not your city — and tells you before you leave whether today is a go-now day or a wait-twenty-minutes day, with live weather on the road ahead. When you set off, it opens your map app for turn-by-turn.
Frequently asked questions
What hours are rush hour?
Morning and evening rush hours have different shapes, which is why a single answer misleads. The morning peak is compressed because people aim at fixed start times. The evening peak is spread across several hours because people leave when their work finishes, so it builds earlier and drains more slowly than most drivers plan for.
When does rush hour end?
Later than most people expect in the evening. The evening peak is typically about twice as wide as the morning one, and the tail clears gradually rather than stopping at a particular minute. The morning peak clears more sharply once the fixed start times have passed.
Why is rush hour worse in one direction?
Employment concentrates at one end of most corridors, so traffic loads inbound in the morning and outbound in the evening. Travelling against that dominant flow is frequently 20 minutes faster on the same road at the same hour, which is why two people with identical commutes can have completely different experiences.
Why does rush hour last longer on Friday?
Friday breaks the pattern in three ways: the evening peak starts earlier, it lasts longer, and it stacks weekend departure traffic on top of the normal commuter load. If you have one flexible day a week, Friday is where flexibility buys the most.
Does leaving 15 minutes earlier actually help?
Usually more than 15 minutes' worth, because the delay curve is steepest at the shoulders of the peak rather than in the middle. A small shift at the right moment beats a large shift at the wrong one. If a 15-minute change saves you less than 15 minutes, you are already past the steep part of the curve.