Hacker News Reader

Comment by throwawayboise | Hacker News Reader

Because Uber drivers are not always transporting people. About half the time, they are driving their empty car ("deadheading") on their way to pick someone up. This still contributes to traffic congestion.

The study also found a corresponding 9% drop in public transport ridership.

View on HN

Why half the time? I'd expect that Uber tries to allocate a nearby driver, so that most pickups are near the last drop-off.

And of course with Uber Pool and Lyft Line, there may be only a slight detour for pickups.

This situation seems better than taxis, which drive an empty car around when they're NOT EVEN on their way to pick someone up. They'll just drive around and around waiting for someone to stick their arm out.

I think Boise was making a general point.

Actually if you look at the complete research report they quote a study which says 40% not 50%:

"due to the substantial deadheading miles (miles travelled without a passenger), which account for at least 40.8% of TNC miles"

They also just drive around "hot" spots with the hope that they will get a customer. Realistically Uber will not match a customer with a driver who is 45' away. They also cannot park like taxis do in designated locations (they are very rare for uber), so they have to keep driving.

45' means 45 minutes, I guess?

It does. Personally, as an amateur astronomer, I just read it without thinking.

However, I can see how many people may take several " to work it out.

' is also the symbol for foot. Which does not make a lot of sense because 45 feet is right there for this purpose but you never know.

45 feet could be across the street, and thus by the time they safely turn around a lot farther than it looks.

> This situation seems better than taxis, which drive an empty car around when they're NOT EVEN on their way to pick someone up.

This is true, but mainly for dense cities prior to Uber/Lyft, and even then cabs driving looking for fares would be in the central part of the city. Here in NYC back in the day I’d rarely find a cab just driving around in Brooklyn, you’d typically find them on main roads driving back to Manhattan after dropping off their last fare if you did.

Now there are just a lot more for hire vehicles driving around the city in general. There was an artificial shortage of medallions, with about 14000 of them back before the apps, plus a bunch of for hire car services (and dollar vans on Flatbush). Now there’s about 100,000 cars for hire in NYC.

Certainly, there is spare capacity of empty ride-sharing cars; it's not just-in-time inventory where they drop someone off and immediately pick someone up.

Also, in many jurisdictions, capacity for taxis is limited; that's what the medallions are for. Ride-sharing capacity is not limited.

I haven't seen data on whether taxis or ride-sharing have more spare capacity in terms of miles, but I have seen data that ride-sharing vehicles significantly increase traffic.

Things may have changed, but last time I read about this topic, Uber is doing a lot less to help the efficiency of the drivers than you’d think. Making real money requires the driver’s themselves to learn their local area’s hotspots and strategies. Which probably makes sense from their incentive perspective.

Drivers ultimately choose to snag rides or not. Many a times I see local drivers circling my area, only the one who ends up grabbing my ride many times is the one 10 miles away on the freeway heading the opposite direction. Now I need to wait 15 minutes while they exit and double back, and if I cancel I might get ranked as a bad passenger.

It's not just deadheading that is the problem. It's also the traffic caused by them stopping in the middle of lanes for pickups and dropoffs, driving around going 3mph looking for their pickup, and various other behaviors.

None of those things are very problematic on very small scales...an inconvenience at most. But just a tad more than that, and it can create absolute chaos.

FTA: "The researchers think that the substantial deadheading miles (miles traveled without a passenger) by TNCs could contribute to the TNC’s negative impact on road congestion. According to some other studies, approximately 40.8 percent of TNC miles are deadheading miles."

Uber could optimize this by predicting where demand will be and telling a subset of drivers to move there. Right now drivers predict and move to hotspots in herds.

Reply on news.ycombinator.com