Tesla has released its latest ‘Vehicle Safety Report’, which looks at different accident rates per mile for vehicles on and off the Aut Topilot.
The latest data shows a slight improvement.
Since 2018, Tesla has been trying to set a benchmark for improving autopilot safety by releasing a quarterly report comparing the number of miles per accident on the quarter topilot vs. accident topilot.
Today, Tesla released its report for Q3 2020:
In the third quarter, we reported one accident for every 59.5959 million miles in which drivers engaged an autopilot. Op For those driving without a topilot but with our active safety features, we’ve registered an accident every 2.42 million miles. Op For those driving without a topilot and without our active safety features, we’ve registered one accident every 1.79 million miles. By comparison, the most recent data from the NHTSA shows that a tomobile crashes every 479,000 miles in the United States.
Here’s how to compare all the different metrics with data from the previous quarter and the same period last year:
- Op Topilot Engaged: It’s an accident for every 4..5959 million miles running in Q32020. That’s one accident for every 3.33 million miles in Q22020 and one accident for every 34.3434 million miles in Q2019201.
- Op topilot is detached but with active safety features: it’s an accident for every 2.42 million miles driven in the Q32020. It has had one accident for every 2.27 million miles in Q22020 and one for every 2.70 million miles in Q3.
- Op topilot detached and without active safety features: It’s an accident for every 1.79 million miles running in the Q32020. The Q22020 had one accident for every 1.56 million miles and the Q3 had one accident for every 1.82 million miles.
Therefore, the data show a slight reduction in accidents from Q2 to Q3 on the Q topilot and a proper improvement year-on-year.
With the exception of the Aut Topilot, but with the safety features operated by the Aut Topilot, it improved quarter-to-quarter, but went down year after year.
Take the electric
As always, please keep in mind that the comparisons here are not exhaustive.
Aut topilot is currently mainly used on highways where it is easy to collect many miles without accidents and non-city driving is getting non-op topilot mileage, where accidents are more likely, so the two datasets cannot be compared.
However, it is still useful to compare op topilot mileage per accident during periods like year-to-year. The data is far from perfect, but we have the best for now.
Tesla will need a lot more data if they want to persuade officials to allow them to configure their entire self-driving system.
Now that the FSD beta has come out, it will be interesting to see how Tesla integrates the data into its security report.
FTC: We use revenue generating auto to affiliate links. More
Subscribe to Electric on YouTube and subscribe to podcasts for exclusive videos.
