On: Tesla Was Kicked Off Fatal Crash Probe by NTSB

[valuearb]

- The average US fatality is not in a ten year old car on a country road. - US crash rates are lower than worldwide rates, so Tesla’s estimate of lives saved worldwide is low. - If autopilot is only used part of the time, that implies its safety benefit is significantly more than 3.7x. A model S is not 3.7x better at passive safety than other cars.

Of course none of this data is very meaningful. You have to compare the rates of autopilot users to non-autopilot users of similar demographics, and have all crashes recorded, not just fatalities, to get close to a comparative samples.

[idoubtit → valuearb]

> US crash rates are lower than worldwide rates, so Tesla’s estimate of lives saved worldwide is low.

Road fatalities per 100,000 inhabitants per year (2013, WHO):

    UK      2.9
    Spain   3.7
    Germany 4.3
    France  5.1
    USA    10.6

Among the rich countries, the US crash rate is one of the worst. With a similar set of vehicles, this rate is very fluctuent, per country and per year. Hint: the problem may not be purely mechanical.

Seriously, would Tesla's autopilot lower the crash rate in the UK? In Thailand? I guess it wouldn't change much in European countries, and any good car would do the same in poor countries.

[kps → idoubtit]

> I guess it wouldn't change much in European countries

Why not?

Via https://en.wikipedia.org/wiki/List_of_countries_by_traffic-r..., here is another table (same countries, same order) giving road fatalities per 10⁹km in 2015:

    UK      3.6
    Spain   7.8
    Germany 4.9
    France  5.8
    USA     7.1