For the sake of transparency, and so that each accident that occurs with Autopilot activated is not a new opportunity to denigrate its production, Tesla has made a habit of posting quarterly reports on this subject.
The last document reports a slight decline without calling into question the usefulness of the system. It has been calculated an average of one accident for 7.29 million kilometers traveled in Tesla with the Autopilot in service.
To be compared with the figures communicated by the American motor vehicle safety control authority NHTSA. The latter figure for the United States 1 collision for 771,000 km. That is almost 9.5 times more accidents than with electric plugs from the Palo Alto manufacturer.
–
In the previous quarter, the performance was even better with 1 damage for 7.53 million km. A decrease of 3.2% which should not erase the very clear increase of 40% between the last quarter of 2019 (1 accident for 5.38 million km) and the first of 2020.
Nice improvement (+ 18% additional distance traveled) in the figures for Tesla drivers who do not use Autopilot but still benefit from active safety devices: 1 collision for 3.65 million km swallowed in the second quarter 2020, compared to 1 for 3.09 million km during the first 3 months of the year.
Progress also (+ 9.6%) for those who do without both Autopilot and active safety systems on board Tesla. In the previous quarter, the manufacturer reported an accident for 2.29 million km. 2Q2020, this average drops to 1 loss record for 2.51 million km.
Author’s opinion
The effectiveness of Autopilot is well established. Even less the one that combines this device and the maximum attention of the driver. The decrease observed is minor a priori. It is still a little troubling that the numbers are improving for Tesla drivers who do without Autopilot, whether or not they are using the other active safety systems.
Why these reverse curves in a period when travel has been particularly reduced in different corners of the planet due to Covid-19? Slackening of attention or excessive speeds when the police are mobilized elsewhere? Or would it be due to the release last April of the version including the operation of Autopilot in town?
Either way, the number comparisons made by Tesla suffer from at least 2 criticisms. First it is compared an average of accidents observed in the United States for all cars, with another which includes Tesla used throughout the world. Then, Autopilot is mainly activated on motorways and expressways where its use can be justified the most. But it is in urban areas that the number of collisions is highest, where Tesla’s autonomous driving device was still absent until last quarter.
We can easily imagine that the difference would be much smaller, between Tesla with Autopilot and the average of all vehicles, if the comparison were made on exclusively identical journeys. Without this erasing the testimonies attesting to the interest of the system.
A figure is also missing: the one that would account for the number of times drivers have had to correct the trajectory controlled by the Autopilot.
–
