Jason71
Well-Known Member
I am pretty sure they work this out based on the number of miles per accident that FSD covers vs the average number of miles per accident for a human driver. I have never thought this was a very fair statistic.FSD Beta is now safer than the average US driver, but not the average UK driver, so that criterion isn’t yet met. The regulatory one neither.
Historically all FSD miles were on highways and even now in the US I would bet that the vast vast majority of FSD miles are on multi lane highways.
So if we want to compare the relative safety we need to compare the number of accidents per highway mile on FSD vs Human.
If that is what they are saying then fair enough it may be safer on a highway I honestly don't know. Highways are one of those places where a human can loose concentration or even fall asleep when a computer will not so it is probably the place where at present FSD will have the most chance of beating a human. It is also probably better at keeping a safe distance from the car in front which humans seem to be really really bad at.
however I would also ( for interest sake) like to see the stats broken down by Day/Night/Rain/Fog/Snow to know how FDS performs against a human in ALL circumstances not just sunny days and dry nights which is what you get most of the time in California.
Also if FSD gets into trouble and hands back to the driver seconds before a crash which side are they scoring that on?