Every rideshare driver asks the same question on every ping: is this Uber ride actually worth it? Here are three real offers from one Friday night in Houston — one driver, one shift — and what each one actually paid.

17:52 — $7.37
5 min (0.8 mi) to pickup · 23 mins (5.3 mi) trip

20:21 — $10.03
14 min (6.7 mi) to pickup · 14 mins (11.3 mi) trip

21:30 — $17.52
3 min (0.5 mi) to pickup · 27 mins (23.2 mi) trip

Which do you take? Don’t scroll. Decide first.

Let’s do the math the right way

I’m not going to pretend you only look at the trip line. You add the legs together — pickup plus trip, both miles and both minutes — because you’re not new at this. So here’s both common rules, computed the right way, on committed miles and committed time:

Nothing wrong with that math. But look at what happens when you check it against what the rides actually earned.

What did these rides actually pay per hour?

Look at those two. One pays $1.21 a mile. The other pays 56 cents a mile — less than half the rate. And they earn the same thing — about thirteen an hour, thirty-four cents apart.

A ride paying less than half the rate per mile, earning exactly as much per hour. How?

The number that the offer includes, but is hidden: speed

17:52 is a crawl — 23 minutes to cover 5.3 miles. That’s 14 mph, stop-and-go. 20:21 is highway; it moves nearly four times as fast.

That’s the whole difference. The slow ride’s high rate per mile gets eaten alive by the clock; the fast ride’s low rate per mile is rescued by how quickly the miles go by. Same money per hour, opposite-looking cards.

Dollars per mile has no slot for speed. Neither does the fare on the screen. Two rides at the same rate per mile pay wildly different amounts depending on how fast you cover the distance — and nothing on the card tells you which is which until you’ve already driven it.

It’s one equation, and every driver already knows it:

dollars per hour = dollars per mile × how fast you’re going

The dollar-a-mile crowd fixes the first term and lets speed float — so their hourly swings and they can’t say why. The twenty-an-hour crowd fixes the result and lets distance float — same blind spot, other end. Both rules are right. Both go blind in the exact place the other one sees, and the thing moving underneath both of them is speed: traffic, time of night, which direction the ride runs.

You can’t solve that at the wheel. The speed on this specific offer, both legs, isn’t something you work out in your head in fifteen seconds while merging, sixty-three times a night.

“So you just sit there turning down work?”

Fair question — it’s the one that kills most advice like this.

I declined the 20:21 offer at 20:21:20. Sixteen seconds later — 20:21:36 — a $26.05 request came in. Fifty-six minutes, taken. Had I accepted the $10.03, I’d have been committed until 20:49, and that $26 ride would’ve landed while I was mid-trip. $17.52 of profit instead of $6.25, for the same stretch of the evening.

One story proves nothing, so here’s the whole log. Across 1,785 real declines:

Declining doesn’t work because the next offer is better. It works because the next one shows up in under three minutes.

When should you take the slow ride anyway?

If the replacement takes more than ten minutes, declining was the wrong call about 70% of the time. Past 45 minutes, it’s wrong essentially always — you turned down thirteen an hour to earn four.

Dead market, bad corner, Tuesday at 2pm: take the thirteen. The thing that makes declining smart is how fast the next ride arrives, and when that dries up, so does the logic. Any tool that tells you to decline without knowing how busy your market is right now is guessing.

The honest version

The rules aren’t wrong. They’re just half of one equation. Dollars per mile and dollars per hour both leave speed floating, and speed is what decides whether a ride that looks good on the card is actually good by the time you drop off.

You can work that out once, at the kitchen table. You can’t work it out sixty-three times a night, in fifteen seconds, with the car moving. That’s the whole job.

For the skeptics: all three cards are real screenshots from Friday 2026-10-30, same driver, same shift. The rule figures use committed miles and committed time — pickup plus trip — not just the trip leg. Committed miles and minutes are Uber’s offer-time estimates, which run 16–29% high against my odometer. Net is after my measured per-mile cost. The “5% Pro Perk” on these cards is already in the quoted fare. The 21:30 offer came through Trip Radar; the other two were standard requests.

One number, every offer, before you tap — [puddlejumper.io]. First 20 Android drivers in the free beta keep it for life.