I presume you know of this web site. Interesting how many FSD/Autopilot versus non-Autopiplot Every Tesla Accident Resulting in Death | Tesla Deaths
I was not aware of that web site but would like know more their methodology. We have similar goals. BTW, I scanned a sample of their cited sources and found all but one predated my March 2019, Model 3 purchase. A rich collection of skeptical articles about Tesla safety, these predate my 'hands on' ownership. The first version AutoPilot had some rough spots. It took three weeks before I could trust it and identify patterns of reproducible failures. But over the +4 years of ownership, I have seen these early defects either disappear or significantly moderate. Today: multi-lane, left turns sometimes include unnecessary lane changes during the turn; poor lane choice approaching a known left or right turn, and; sometimes treating an exit lane as part of the route to the destination. There are still phantom brake events but they coincide with Google Maps defects when the survey car recorded a temporary construction zone. Bob Wilson
Comparing "Tesla Deaths" for 2021, 41, to the 2021 FARS database, 69, there is a significant difference. Although the four, January deaths align, there are 28 missing. I could do a correlation study to identify the data gaps between the two but I'm not sure their other data can be trusted. The FARS data has enforcement by Federal law. The "Tesla Deaths" uses other sources and apparently, does not include NHTSA FARS. Bob Wilson
