If you follow the link in the article, "iSeeCars" has a table of "Fatal Accident Rate (Cars per Billion Vehicle Miles)". I don't understand why they count fatal accident rates in CARS per billion miles rather than FATALITIES per billion miles. A car that reliably kills everyone in it every crash but almost never crashes would be "More safe" by this metric.
But it's very weird. The "HIGH" list is full of normal cars like "Prius: 5.9" and "Kia Forte: 8.1" and "Infiniti Q50: 5.8"
But the stated average of the whole set is 2.8! That means there are tons of cars that are extremely safe by this metric. Why don't they show anything?
They don't show any of their data at all really. I'm sad, I really want to understand what the LOW outliers are. I think there's a chance that not showing any of the LOW side of the ranking is (purposely or accidentally) hiding some weirdness with this metric or data.
Then you see the Rate per Metro chart and realize the only thing this is tracking is drivers, and some drivers are 40x more likely to get themselves into a deadly wreck than good drivers. Same exact car (both the "dangerous" and "safe" ones) in different metros can swing by an entire magnitude. Apparently there really ARE some cities where drivers are just trying to kill each other.
If you follow the link in the article, "iSeeCars" has a table of "Fatal Accident Rate (Cars per Billion Vehicle Miles)". I don't understand why they count fatal accident rates in CARS per billion miles rather than FATALITIES per billion miles. A car that reliably kills everyone in it every crash but almost never crashes would be "More safe" by this metric.
But it's very weird. The "HIGH" list is full of normal cars like "Prius: 5.9" and "Kia Forte: 8.1" and "Infiniti Q50: 5.8"
But the stated average of the whole set is 2.8! That means there are tons of cars that are extremely safe by this metric. Why don't they show anything?
They don't show any of their data at all really. I'm sad, I really want to understand what the LOW outliers are. I think there's a chance that not showing any of the LOW side of the ranking is (purposely or accidentally) hiding some weirdness with this metric or data.
Then you see the Rate per Metro chart and realize the only thing this is tracking is drivers, and some drivers are 40x more likely to get themselves into a deadly wreck than good drivers. Same exact car (both the "dangerous" and "safe" ones) in different metros can swing by an entire magnitude. Apparently there really ARE some cities where drivers are just trying to kill each other.
Discussion (293 points, 5 days ago, 436 comments) https://news.ycombinator.com/item?id=42150443
Can't wait into Elon is in office and shuts down the NHTSA for their insolence and nerve to publish these reports, which are obviously a lie.