Interestingly, Q1 2023, which Tesla is only releasing now, showed a significant decrease in miles driven between accidents compared to the same period for the year prior, which might explain why Tesla stopped releasing the data at that time.

[–] threelonmusketeers@sh.itjust.works 2 points 4 months ago* (last edited 4 months ago) (1 children)

I wonder what caused the temporary regressions in Q3 and Q4? I assume the Q1 improvement is due to the majority of the fleet switching to the FSD v12 stack.

[–] mortalic@lemmy.world 2 points 4 months ago

When did they switch to vision only and stop using lidar or whatever it was?

[–] Everythingispenguins@lemmy.world 2 points 4 months ago (1 children)

I wonder what they're defining an accident as?

[–] polygon6121@lemmy.world 4 points 4 months ago* (last edited 4 months ago) (1 children)

Tesla only counts accidents where the airbag have been deployed. This makes the stats very misleading as the 'United states average' count any accident reported to either police or insurance.

Also it is likely statistically biased, for example one issue is that the different drive assist systems are more likely to be used under the best driving conditions, so accidents are less likely to happen when it is turned on, because it is mostly used in the easiest of driving conditions.

[–] Nogami@lemmy.world 2 points 4 months ago* (last edited 4 months ago)

Sounds like the US averages are pretty useless then. Everyone around here tries to game reports like crazy if there’s any potential benefit for them.

I’d actually prefer to see stats that generate a documented injury that was treated somewhere, rather than someone’s paint getting scraped by a poor parking job.

[–] MeanEYE@lemmy.world 1 points 4 months ago

Heavily biased numbers. Had it been this safe EU wouldn't blink and eye.