←back to thread

Tesla has the highest fatal accident rate of all auto brands, study finds

(www.roadandtrack.com)
307 points MBCook | 9 comments | 15 Nov 24 20:15 UTC | HN request time: 1.023s | source | bottom
Show context
TrainedMonkey ◴[15 Nov 24 21:10 UTC] No.42151114[source]▶
> The study was conducted on model year 2018–2022 vehicles, and focused on crashes between 2017 and 2022 that resulted in occupant fatalities.

Teslas can go fast real fast, so naively this is the result I would expect given how they have filtered the data. In other words, unless they controlled for this, this would be biased by natural selection playing out.

Having said that, as someone who had a couple of close calls with the autopilot. I would love to know what percent of those crashes was with autopilot enabled.

replies(2): >>42151164 #>>42151683 #
1. phony-account ◴[15 Nov 24 21:14 UTC] No.42151164[source]▶
> Teslas can go fast real fast, so naively this is the result I would expect given how they have filtered the data. In other words, unless they controlled for this…

Explain to me why you would want to filter out fatalities caused by going “real fast”?

replies(2): >>42151199 #>>42151235 #
2. Petersipoi ◴[15 Nov 24 21:17 UTC] No.42151199[source]▶
Because comparing the fatality rate of a Corolla going 50mph and a Tesla going 90mph is useless to a person who wants a car that is safe when driven responsibly.
replies(4): >>42151291 #>>42151324 #>>42151392 #>>42151561 #
3. IncreasePosts ◴[15 Nov 24 21:21 UTC] No.42151235[source]▶
If a bunch of lunatics buy the car because it is fast, and kill themselves, that doesn't necessarily affect my safety in the car if I'm buying it for some other feature and don't intend to drive it dangerously.
replies(1): >>42151838 #
4. ◴[15 Nov 24 21:26 UTC] No.42151291[source]▶
5. janalsncm ◴[15 Nov 24 21:29 UTC] No.42151324[source]▶
Unless the Tesla induces unsafe behavior, of course. Does the car make it easier to break the speed limit, drive distracted, or drive under the influence? I don’t know.
replies(1): >>42151601 #
6. skybrian ◴[15 Nov 24 21:35 UTC] No.42151392[source]▶
It seems fair to say that it's difficult to control for variation due to the drivers being different. But I don't think giving faster cars a better rating is a good way to control for that. Faster seems more dangerous for other reasons.
7. mrguyorama ◴[15 Nov 24 21:49 UTC] No.42151561[source]▶
Corollas can easily go 90mph, so can a Prius, so can a dodge neon.

Even eco-shitboxes in the US have 160 hp. Sure, they get 0-60 times of 10 seconds, but I don't think there's been a car model in the US that cannot reach 100mph in decades

8. bravetraveler ◴[15 Nov 24 21:54 UTC] No.42151601{3}[source]▶
Quite literally selling points
9. piva00 ◴[15 Nov 24 22:16 UTC] No.42151838[source]▶
I think gathering the data to judge who's a lunatic is quite hard and fragile. Finding proxies like past infractions record would be already hard enough to compile at scale, I can't really derive a passable methodology that could tell you what you're asking for.

At least the data informs others that perhaps it's good to be cautious around Teslas, not very much if it's a safe purchase, and they state that it's a safe car so I don't see the hangup you had about it on position of a buyer.