Blog Tesla Says Q1 was Autopilot’s Safest to Date

Tesla’s Autopilot experienced its safest quarter since release, according to the company’s quarterly safety report.

The report showed that vehicles operating on Autopilot were involved in an accident one time for every 4.68 million miles driven. That performance bested Q4 2019, where Tesla logged one accident for every 3.07 million miles.

Tesla said owners who drove without Autopilot but use Tesla’s active safety features experienced an accident every 1.99 million miles. Accidents with no Autopilot or active safety features occurred every 1.42 million miles.

Tesla notes that National Highway Transportation Safety Administration data suggests crashes occur once every 479,000 miles on average.

 

[cucubits]

an accident one time for every 4.68 million miles driven

It would be interesting to see how many of these accidents are actually APs fault vs the driver not using it properly (not paying attention).

 

[diplomat33]

Since AP is a L2 driver assist and the driver is still considered to be responsible, it would probably be more appropriate to understand these statistics as "human drivers assisted by Autopilot only had 1 accident per 4.68 million miles."

 

[SOULPEDL]

Since AP is a L2 driver assist and the driver is still considered to be responsible, it would probably be more appropriate to understand these statistics as "human drivers assisted by Autopilot only had 1 accident per 4.68 million miles."

Agreed. But those numbers make a compelling case because it's almost exactly 10x safer than the national average. Wasn't that Elon's goal? I've always thought it's like 2 people driving. And once you master the hand-offs, this is safer, therefore get Autopilot.

 

[Ken Cole]

Once again apples to oranges. What are the numbers on the same types of roads/conditions? How about at least just comparing highway miles?

 

[Mardak]

How about at least just comparing highway miles?

Yeah, assuming people have been using Autopilot on similar types of roads year to year, here's trying to look at the numbers over a year instead of quarter by quarter as driving behavior changes with seasons.

So 2019 Q1 to 2020 Q1 had 63% improvement in Autopilot miles per accident. And the average Autopilot miles per accident for the previous 4 quarters since 2019 Q2 have always been improving but not as much for the 12 months ending 2019 Q4.

Unclear if the big increase the past quarter were from Autopilot improvements or perhaps a less wet winter?

 

[Ande]

And yet, the corrupt EU is crippling the AP and nagging us to boredom!

 

[Dutchie]

In the 1st quarter, we registered one accident for every 4.68 million miles driven in which drivers had Autopilot engaged. For those driving without Autopilot but with our active safety features, we registered one accident for every 1.99 million miles driven. For those driving without Autopilot and without our active safety features, we registered one accident for every 1.42 million miles driven. By comparison, NHTSA’s most recent data shows that in the United States there is an automobile crash every 479,000 miles.

Total overall miles and crashes were significantly reduced in this quarter.

So let's assume AP engaged is only highway
All other situations is driving on all roads.

Conclusion:

1. Tesla drivers are very safe drivers in general 1.42 million Tesla w/o safety, versus 479,000 NHTSA's;
2. Safety features increases safety of Tesla's by about 40% (1.42 ->1.99);
3. Highway driving is safer than general driving.

 

[AdamVIP]

I think these results can at least somewhat be attributed to the fact that significantly less cars were on the road for the second half of the quarter. Less things to accident into means less accidents.

 

[Mardak]

Tesla Vehicle Safety Report

Looks like 2020 Q1 keeps the highest Autopilot engaged miles per accident with 4.68 million as Q2 dropped QoQ to 4.53 but was an increase YoY from 3.27. As before looking at the trailing 12 months, assuming similar number of miles driven per quarter, yearly average Autopilot engaged miles has been steadily increasing.

At least comparing Autopilot engaged miles relative to without active safety miles, the ratio has increased from 2.1x from a year ago to 2.6x now. It'll be interesting to see if Tesla provides more details for these numbers such as the split between highway vs city street miles especially as Autopilot now has stop control or if Tesla considers Level 3 for highways.

 

[Soda Popinski]

It'll be interesting to see what happens now that more people will use AP in surface streets.