Welcome to Tesla Motors Club
Discuss Tesla's Model S, Model 3, Model X, Model Y, Cybertruck, Roadster and More.
Register

Any Actual Crashes with FSD?

S4WRXTTCS

Well-Known Member
May 3, 2015
6,029
7,286
Snohomish, WA
Hopefully someone who has time can look at it from a statistical standpoint.

I think roughly 15K people have it, and if you assume the average miles driven is still the same then what are the odds there was a reportable crash during the time frame? Assuming only 10.3.1 (or whatever people are referring to the fixed one as).

I can't see the FSD Beta adding much safety or reducing safety that much.

The reason for that is there are things that contribute to non-safety like the desire a driver has to see if it will pull off something. Some drivers are also likely allowing differences in how it drives versus how they normally drive. Like it's been following more closely than I would (in city driving), and its closing speed is much higher than I would have when it comes up to stopped traffic (its totally inefficient). There is also likely distraction like talking to an audience or pushing the report button.

Then there is the fact that its just too terrible for anyone to let their guard down. It's far more exhausting to drive with FSD Beta than without it. So I think its safe to assume it cuts down on some accidents simply because people are so tuned into it. Heck I'm so tuned into it that I haven't had any music on in the car during the start of the testing.

So my assumption is it averages out. So the question is will we hear about it? I think it depends on who the accident happens with. Some people would be far too embarrassed to admit. Or they'll be the person who has to blame someone else.

A lot of this also comes down to whether people use it. I've been rather selective in where to use it. Like today I was hoping to take a left with it from a stop light, but then a car got behind me. Now sure I could have just used it, and accepted any embarrassment. But, I was more like "drats, maybe next time". So I don't know if its really being tested that extensively by the 15K people when I myself have used it a very limited manner.
 

mtndrew1

Active Member
May 12, 2015
1,388
3,971
Gardena, CA
After a few hundred miles using the system I’d wager that there have been zero incidents with pedestrians or cyclists. FSD is just insanely cautious around any of them.

Likewise I doubt there have been any additional highway accidents (above conventional autopilot or human driving) because it’s largely the same system as the public build.

FSD stumbles with braking abruptly or misjudging moderate-speed cross traffic. I would bet that if there have been any collisions they’ve been fender-benders from being rear ended or from turning into moderate speed traffic at the wrong time.

I’d be surprised if there have been any remotely life-threatening instances with FSD beta. The nature of the system really has the user hovering over the pedals and watching everything like a hawk. The stuff FSD is bad at is either so obvious as to generate an immediate correction from the driver or so slow that it would merely result in some bent sheet metal.

I’m so engaged with FSD while it’s activated that I’m probably safer using the system than not. It naturally requires driver engagement to the point that it’s fatiguing. Nobody in practice is gonna turn it on and think that they can just let it roll, at least in high traffic environments.
 

JulienW

Active Member
Jul 7, 2018
3,167
4,203
Atlanta
The maim problem will be statistical. There have only be a couple of thousand until 3 weeks ago. Looks like it is up to over 10K now and next week several thousand more will be added. Sooner or later there will be accidents and the more people added with the lower scores will add to the numbers when they start.
 

EVNow

Well-Known Member
Sep 5, 2009
12,584
33,635
Seattle, WA
I think roughly 15K people have it, and if you assume the average miles driven is still the same then what are the odds there was a reportable crash during the time frame? Assuming only 10.3.1 (or whatever people are referring to the fixed one as).
Odds of small accidents (L4+) are about 1 in 10k. Reportable accident (L2+) is 1 in 100k.

If FSD testers drover 10 miles daily, on average we could have one reportable accident per day ! That has definitely not happened. If its more like 5 miles daily, we should still see an accident every other day. That hasn't happened either. So - I'd say FSD + cautious drivers are safer than average drivers in US.

1635980410064.png
 
Hopefully someone who has time can look at it from a statistical standpoint.

I think roughly 15K people have it, and if you assume the average miles driven is still the same then what are the odds there was a reportable crash during the time frame? Assuming only 10.3.1 (or whatever people are referring to the fixed one as).

I can't see the FSD Beta adding much safety or reducing safety that much.

The reason for that is there are things that contribute to non-safety like the desire a driver has to see if it will pull off something. Some drivers are also likely allowing differences in how it drives versus how they normally drive. Like it's been following more closely than I would (in city driving), and its closing speed is much higher than I would have when it comes up to stopped traffic (its totally inefficient). There is also likely distraction like talking to an audience or pushing the report button.

Then there is the fact that its just too terrible for anyone to let their guard down. It's far more exhausting to drive with FSD Beta than without it. So I think its safe to assume it cuts down on some accidents simply because people are so tuned into it. Heck I'm so tuned into it that I haven't had any music on in the car during the start of the testing.

So my assumption is it averages out. So the question is will we hear about it? I think it depends on who the accident happens with. Some people would be far too embarrassed to admit. Or they'll be the person who has to blame someone else.

A lot of this also comes down to whether people use it. I've been rather selective in where to use it. Like today I was hoping to take a left with it from a stop light, but then a car got behind me. Now sure I could have just used it, and accepted any embarrassment. But, I was more like "drats, maybe next time". So I don't know if its really being tested that extensively by the 15K people when I myself have used it a very limited manner.
I appreciate your saying that using the FSD BEta is exhausting. We have a score of 97 after driving 1.000 miles to the Grand Canyon and back and experienced lots of glitches--esp. unreliable driving though construction zones---and this is on the Cruise Control/autopilot. I haven't driven the car all that much so have been afraid to try the FSD--so feel reinforced by what you have said. We do regret shelling out 10 grand on the FSD and feell like fools in that regard. I initially was a fan-girl but now would NOT recommend Tesla to anyone.
 

EVNow

Well-Known Member
Sep 5, 2009
12,584
33,635
Seattle, WA
But this begs the question. How many and how low of a score are going to be allowed on the Beta? If everyone that pays for FSD gets it in a few months that seems to be a recipe for LOTS of problems.
Elon was saying something about 70 (above 70th percentile ?). They may just have a cut-off at say 95 and let anyone who has the determination & patience to get 95 in.

If they are already at 12k testers (~90 percentile) - I'm not sure they need more at this point. Letting more people in now is probably a non-technical decision.
 

Daniel in SD

Well-Known Member
Jan 25, 2018
7,369
10,908
San Diego
Odds of small accidents (L4+) are about 1 in 10k. Reportable accident (L2+) is 1 in 100k.

If FSD testers drover 10 miles daily, on average we could have one reportable accident per day ! That has definitely not happened. If its more like 5 miles daily, we should still see an accident every other day. That hasn't happened either. So - I'd say FSD + cautious drivers are safer than average drivers in US.

View attachment 729055
Really curious about the source of this data and definition of the collision severity.
Anyways, it does seem like FSD Beta does not have any issues with inducing automation complacency!
I think when there is a severe collision it will be because the car moves from a stop while the driver isn’t paying attention.
 
  • Like
Reactions: Terminator857

Terminator857

Member
Aug 5, 2019
839
1,014
Ca
I doubt if there is an FSD accident it will be reported. Most users are fans of FSD and know it is there fault if an accident happens. They will report it to their insurance if over deductible and not much more. Is Tesla required to report? Would Tesla know of an accident that broke a headlight or ran over someone's toe?
 
  • Like
Reactions: Dan D.

novox77

Active Member
Nov 25, 2017
2,371
5,114
NH, MA
I initially was a fan-girl but now would NOT recommend Tesla to anyone.

You can still recommend Tesla if you think the non FSD aspects of the car are good. Just explain FSD is a long work in progress and don't assume the car can drive you anywhere without driver attentiveness. I'm hoping we'll see a significant FSD offering in a few years.
 

Dan D.

Member
Dec 7, 2020
879
1,097
Vancouver, BC
Several recent crashes. Are any related to AP or FSD? No definitive answer to that question. If the driver didn't admit it how would the police know? Yes, Tesla must report AP/FSD crashes to the NHTSA, is that database publicly searchable?

Oct 28, 2021
a Nissan Sentra ... incurred a flat tire, causing it to veer off the road, strike the guardrail and come to a stop in the roadway ...Then a 58-year-old Mulberry, Fla., man crashed his Tesla into the Nissan Sentra

Oct 19, 2021
A Tesla crashed into the wall of an Olathe business... It is not known if the Tesla was on Autopilot at the time of the crash

Oct 30, 2021
CHP said the rider of a black 2016 Harley Davidson Road Glide motorcycle and a Tesla collided... The two crashed for an unknown reason

Nov 1, 2021
a black Tesla car veered off the roadway while traveling on the highway in Sacramento

Based on these types of videos, do we really believe there are no crashes due to FSD?
FSD BETA 10.3.1 Fails with Traffic!!

 
Last edited:
One thing that has positively impressed me with FSD beta 10.3.1 (my first FSD experience, and for under a week still) is how well the car protects itself. Almost as well as it avoids hitting pedestrians and bicycles. I'll be cruising along a narrow street with cars parked on both sides, and never feel like I'm about to scrape one of the parked cars. It seems very reliable that way.
 

Products we're discussing on TMC...

About Us

Formed in 2006, Tesla Motors Club (TMC) was the first independent online Tesla community. Today it remains the largest and most dynamic community of Tesla enthusiasts. Learn more.

Do you value your experience at TMC? Consider becoming a Supporting Member of Tesla Motors Club. As a thank you for your contribution, you'll get nearly no ads in the Community and Groups sections. Additional perks are available depending on the level of contribution. Please visit the Account Upgrades page for more details.


SUPPORT TMC
Top