4 ms·

All of your arguments are expounded upon in the article itself, and their conclusions still hold, based on the publicly available data. The 3x figure in the ti

All of your arguments are expounded upon in the article itself, and their conclusions still hold, based on the publicly available data. The 3x figure in the title is based on a comparison of the Tesla reports with estimated average human driver miles without an incident, not based on police report data. The comparison with police-report data would lead to a 9x figure instead, which the article presents but quickly dismisses. The denominator problem is made up. Tesla Robotaxi has only been launched in one location, Austin, and only since July (well, 28th June, so maybe there is a few days discrepancy?). So the crash data and the miles data can only refer to this same period. Furthermore, if the miles driven are actually based on some additional length of time, then the picture gets even worse for Tesla, as the denominator for those 9 incidents gets smaller. The analysis indeed doesn't distinguish between the types of accidents, but this is irrelevant. The human driver estimates for miles driven without incident also don't distinguish between the types of incidents, so the comparison is still very fair (unless you believe people intentionally tried to get the Tesla cars to crash, which makes little sense). The comparison to Waymo is also done based on incidents reported by both companies under the same reporting requirements, to the same federal agency. The crash definitions and reporting practices are already harmonized, at least to a good extent, through this. Overall there is no way to look at this data and draw a conclusion that is significantly different from the article: Tesla is bad at autonomous driving, and has a long way to go until it can be considered safe on public roads. We should also remember that robotaxis are not even autonomous, in fact! Each car has a human safety monitor that is ready to step in and take control of the vehicle at any time to avoid incidents - so the real incident rate, if the safety monitor weren't there, would certainly be even worse than this. I'd also mention that 5 months of data is not that small a sample size, despite you trying to make it sound so (only 9 crashes).

parineum 9mo ago
> The 3x figure in the title is based on a comparison of the Tesla reports with estimated average human driver miles without an incident, not based on police report data. The comparison with police-report data would lead to a 9x figure instead, which the article presents but quickly dismisses. I think OP's point still stands here. Who are people reporting minor incidents to that would be publicly available that isn't the police? This data had to come from somewhere and police reports is the only thing that makes sense to me. If I bump my car into a post, I'm not telling any government office about it.
- estearum 9mo ago
  Insurers? I can't be certain about auto insurers, but healthcare insurers just straight up sell the insurance claims data. I would be surprised if auto insurers haven't found that same "innovation."
  - fc417fc802 9mo ago
    That's a fair point, but I'll note that the one time I hit an inanimate object with my car I wasn't about to needlessly involve anyone. Fixed the damage to the vehicle myself and got on with life. So I think it's reasonable to wonder about the accuracy of estimates for humans. We (ie society) could really use a rigorous dataset for this.
    - estearum 9mo ago
      Tesla could just share their datasets with researchers and NHTSA and the researchers can do all the variable controls necessary to make it apples to apples. Tesla doesn't because presumably the data is bad.
    - eszed 9mo ago
      Was going to say the same thing. I've had three of these sorts of "crashes": once was my fault, but no damage, so I apologized and we shook hands and drove away; once was their fault, but same result; third was unambiguously his fault, but I laughed and told him it was his lucky day that the scratch down the side of my car was only maybe the third worst on the beater I was driving at the time. That guy's bumper was all but ripped off, but (though we exchanged details) I never heard anything more about it, so I'm damn sure he never went to insurance. If you count only the (my fault) first of those I'm well ahead of the once every 200k miles average someone proposed up-thread - though if you count the time I ripped off the underlip of my bumper parking in front of a too-high curb, I'm probably back at par.
- YeGoblynQueenne 9mo ago
  FTA: >> However, that figure doesn’t include non-police-reported incidents. When adding those, or rather an estimate of those, humans are closer to 200,000 miles between crashes, which is still a lot better than Tesla’s robotaxi in Austin.
YeGoblynQueenne 9mo ago
To add to this, more data from more regions means the estimate of average human miles without an incident is more accurate, simply because it is estimated from a larger sample, so more likely to be representative.
brookst 9mo ago
I agree with most of your points and your conclusion, but to be fair OP was asserting that human drivers under-report incidents, which I believe. Super minor bumps where the drivers get out, determine there’s barely a scratch, and go on. Or solo low speed collisions with walls in garage or trees. I don’t think it invalidates the conclusion, but it seems like one fair point in an otherwise off-target defense.
- tsimionescu 9mo ago
  Sure, but the 3x comparison is not based on reported incidents, it's based on estimates of incidents that occur. I think it's fair to assume such estimates are based on data about repairs and other such market stats, that don't necessarily depend on reporting. We also have no reason a priori to believe the Tesla reports include every single incident either, especially given their history from FSD incident disclosures.
  - jimmydddd 9mo ago
    "estimates" (with air quotes)
cameldrv 9mo ago
Statistically 9 crashes is enough to draw reasonable inferences from. If they had the expected human rate of 3 over the period in question, the chance that they would actually get into 9 accidents is about 0.4%. And mind you, that’s with a safety driver. It would probably be much worse without one.
fragmede 9mo ago
Tesla Robotaxi service is also available in the San Francisco Bay Area, with an area serviced greater than Waymo. Since at least September, 2025, but probably earlier.

parineum 9mo ago

> The 3x figure in the title is based on a comparison of the Tesla reports with estimated average human driver miles without an incident, not based on police report data. The comparison with police-report data would lead to a 9x figure instead, which the article presents but quickly dismisses. I think OP's point still stands here. Who are people reporting minor incidents to that would be publicly available that isn't the police? This data had to come from somewhere and police reports is the only thing that makes sense to me. If I bump my car into a post, I'm not telling any government office about it.

estearum 9mo ago
Insurers? I can't be certain about auto insurers, but healthcare insurers just straight up sell the insurance claims data. I would be surprised if auto insurers haven't found that same "innovation."
- fc417fc802 9mo ago
  That's a fair point, but I'll note that the one time I hit an inanimate object with my car I wasn't about to needlessly involve anyone. Fixed the damage to the vehicle myself and got on with life. So I think it's reasonable to wonder about the accuracy of estimates for humans. We (ie society) could really use a rigorous dataset for this.
  - estearum 9mo ago
    Tesla could just share their datasets with researchers and NHTSA and the researchers can do all the variable controls necessary to make it apples to apples. Tesla doesn't because presumably the data is bad.
  - eszed 9mo ago
    Was going to say the same thing. I've had three of these sorts of "crashes": once was my fault, but no damage, so I apologized and we shook hands and drove away; once was their fault, but same result; third was unambiguously his fault, but I laughed and told him it was his lucky day that the scratch down the side of my car was only maybe the third worst on the beater I was driving at the time. That guy's bumper was all but ripped off, but (though we exchanged details) I never heard anything more about it, so I'm damn sure he never went to insurance. If you count only the (my fault) first of those I'm well ahead of the once every 200k miles average someone proposed up-thread - though if you count the time I ripped off the underlip of my bumper parking in front of a too-high curb, I'm probably back at par.
YeGoblynQueenne 9mo ago
FTA: >> However, that figure doesn’t include non-police-reported incidents. When adding those, or rather an estimate of those, humans are closer to 200,000 miles between crashes, which is still a lot better than Tesla’s robotaxi in Austin.

YeGoblynQueenne 9mo ago

To add to this, more data from more regions means the estimate of average human miles without an incident is more accurate, simply because it is estimated from a larger sample, so more likely to be representative.

brookst 9mo ago

I agree with most of your points and your conclusion, but to be fair OP was asserting that human drivers under-report incidents, which I believe. Super minor bumps where the drivers get out, determine there’s barely a scratch, and go on. Or solo low speed collisions with walls in garage or trees. I don’t think it invalidates the conclusion, but it seems like one fair point in an otherwise off-target defense.

tsimionescu 9mo ago
Sure, but the 3x comparison is not based on reported incidents, it's based on estimates of incidents that occur. I think it's fair to assume such estimates are based on data about repairs and other such market stats, that don't necessarily depend on reporting. We also have no reason a priori to believe the Tesla reports include every single incident either, especially given their history from FSD incident disclosures.
- jimmydddd 9mo ago
  "estimates" (with air quotes)

cameldrv 9mo ago

Statistically 9 crashes is enough to draw reasonable inferences from. If they had the expected human rate of 3 over the period in question, the chance that they would actually get into 9 accidents is about 0.4%. And mind you, that’s with a safety driver. It would probably be much worse without one.

fragmede 9mo ago

Tesla Robotaxi service is also available in the San Francisco Bay Area, with an area serviced greater than Waymo. Since at least September, 2025, but probably earlier.