Tesla appears to have an impressive history of automatically disengaging FSD immediately before a collision, and this page is quite light on details of what they’re comparing to what, so I would take it with a large grain of salt.

> and this page is quite light on details of what they’re comparing to what

In the page:

"If FSD (Supervised) was active at any point within five seconds leading up to a collision event, Tesla considers the collision to have occurred with FSD (Supervised) engaged for purposes of calculating collision rates for the Vehicle Safety Report."

They are pretty open about how the stats are reported.

For how many seconds after FSD is disengaged do accidents still get attributed to FSD?

True, we need to see the details of how they compute all this data. I do remember reading that they include crash data up to like a minute after FSD has been disengaged so these types of crashes should be included.

Nevertheless, the latest version of FSD I can certainly believe is seven times less likely to get into an accident than the average driver. I experienced it daily.