Tesla crashes while driving on Autopilot increased lots quarter-to-quarter, based on Tesla’s own Vehicle Safety Report, however it was to be expected with the tougher driving conditions within the winter and crashes are still less frequent than without Autopilot.
For over a 12 months now, Tesla has been releasing a voluntarily quarterly safety data report “in an effort to provide critical safety details about our vehicles to the general public.”
The information shouldn’t be perfect, however it’s the most effective we’ve at once to statistically study the protection of Autopilot inside a big fleet.
Today, Tesla released the report for the fourth quarter of 2019:
“Within the 4th quarter, we registered one accident for each 3.07 million miles driven during which drivers had Autopilot engaged. For those driving without Autopilot but with our lively safety features, we registered one accident for each 2.10 million miles driven. For those driving without Autopilot and without our lively safety features, we registered one accident for each 1.64 million miles driven. By comparison, NHTSA’s most up-to-date data shows that in the US there may be an automobile crash every 479,000 miles.”
It compares to the next stats from the previous quarter:
- one accident for each 4.34 million miles driven during which drivers had Autopilot engaged
- one accident for each 2.70 million miles driven during which drivers didn’t have Autopilot engaged but with lively safety features
- one accident for each 1.82 million miles driven during which drivers didn’t have Autopilot engaged nor any lively safety feature
It represents a big increase within the variety of accidents per mile for Autopilot.
The rise may be partly explained by the tougher road conditions attributable to the winter starting in some markets.
The information from the identical quarter for the previous 12 months show a really slight improvement of accidents per mile for Autopilot.
Nonetheless, the info still looks good for Autopilot when put next to your entire auto market. With a crash every 479,000 miles for vehicles overall, Autopilot is about 6 times safer with a crash every 3.07 million miles based on this very simplistic option to have a look at the situation. Here’s an awesome visualization from Hypercharts:
The foremost grievance about Tesla’s set of Autopilot data is that the system is usually used on the highway versus city driving, where accidents are more common.
Subsequently, it’s probably not useful to check those two datasets.
Also, the comparison with the general NHTSA data also includes older vehicles, which usually tend to be involved in accidents than Tesla’s far more recent vehicles on average.
Nonetheless, the info with Tesla vehicles without Autopilot and without lively safety features is a more interesting comparison point in my view and it’s promising.
It’s the most effective we’ve at once, but there’s definitely room for improvement.