3 ms·

It sounds a bit like the pre-crime angle from Minority Report. Because aggressive drivers with accidents and citations already pay higher rates. This is, I as

by tyingq 3y ago

It sounds a bit like the pre-crime angle from Minority Report. Because aggressive drivers with accidents and citations already pay higher rates. This is, I assume, penalizing aggressive driving that hasn't yet resulted in a citation or accident. Or, maybe, penalizing those that drive in riskier areas , regardless of habits? Might be interesting if someone could push for whether anyone's rates went down as a result of data sharing.

mjbeswick 3y ago
That's how insurance underwriting works anyway, as gender, age, car model ect are all used to calculate a premium based on risk. In the UK you can have a "black box" fitted to your car by insurance companies, where low risk driving get you a discount on your renewal.
- tyingq 3y ago
  Right. It would be interesting, though, if this new source of data for them only resulted in increases for certain people...and no decreases for anyone.
- mfer 3y ago
  You can get the "black box" in the US, too. A big difference is that with the "black box" you opt-in. It's a choice. What's being reported on happens without people even having knowledge of the situation.
nickjj 3y ago
> This is, I assume, penalizing aggressive driving that hasn't yet resulted in a citation or accident. If there's no context about the situation then how do they differentiate aggressive driving vs reasonable driving? For example if you're in a 2 lane highway and the left lane is clear but you're behind someone doing 40 mph in a 60 mph on the right lane, so you quickly accelerate and switch lanes to pass them. Super defensive drivers who happen to also go unreasonably slow cause a lot of problems and traffic. Also, what about temporarily speeding to get through a yellow light because the alternative is to slam on your breaks and potentially get rear ended. Or coasting a bit above the speed limit because you're "keeping up with traffic" on a major highway with no lights.
- tyingq 3y ago
  I assume they just don't care about exceptions, and just correlate past data to actual claims to adjust their actuarial tables. But to your point, maybe that also penalizes non-aggressive things that result in accidents.
  - nickjj 3y ago
    > I assume they just don't care about exceptions, and just correlate past data to actual claims to adjust their actuarial tables Yep, IMO that's a problem in general. So much is based on case studies in aggregate instead of individuals. The same thing happens with medicine (at least in the US). Just because 200 people in 1 study demonstrated that X year old Y gender people who live in Z have 7.8917% more likelihood to be in an accident vs the average population which was determined by 20 other studies doesn't mean you or I have the same percent.
    - falcolas 3y ago
      > So much is based on case studies in aggregate instead of individuals. Isn't this a good thing? Insurance is about spreading individual risk across a group. Hyper targeting the premiums of individuals according to their individual risk is the exact opposite of spreading risk.
      - nickjj 3y ago
        It depends on who you're optimizing for, insurance companies or individual drivers.
- Jimmc414 3y ago
  > Super defensive drivers who happen to also go unreasonably slow cause a lot of problems and traffic. Unfortunately I think the entities driving this only care about the specific car they are insuring, not the traffic problems indirectly attributable to the insured.
mint2 3y ago
Which sounds more fair to use in rating insurance: A. Their credit related data B. Their age C. Their marital status D. Actual data on how they drive I feel like D, actual data on how they drive is the most fair. I think there needs to be mechanisms to contest driving events and also regulations on types of driving data allowed, but I think overall it’s by far the most fair and directly related But actually A B and C are used in most but not all states.
- kelnos 3y ago
  > I feel like D, actual data on how they drive is the most fair. Only if you can actually prove a correlation. I think a lot of people, when asked, "do you people who brake and accelerate hard are likely to be in more crashes than people who don't?", most people might think for a second and say, "yeah, sure, that sounds like it would be the case". But is it, actually? The funny thing is that this data apparently does exist, so it shouldn't be too hard to figure out. But I suspect no one's actually done that.

by tyingq 3y ago

mjbeswick 3y ago

That's how insurance underwriting works anyway, as gender, age, car model ect are all used to calculate a premium based on risk. In the UK you can have a "black box" fitted to your car by insurance companies, where low risk driving get you a discount on your renewal.

tyingq 3y ago
Right. It would be interesting, though, if this new source of data for them only resulted in increases for certain people...and no decreases for anyone.
mfer 3y ago
You can get the "black box" in the US, too. A big difference is that with the "black box" you opt-in. It's a choice. What's being reported on happens without people even having knowledge of the situation.

nickjj 3y ago

> This is, I assume, penalizing aggressive driving that hasn't yet resulted in a citation or accident. If there's no context about the situation then how do they differentiate aggressive driving vs reasonable driving? For example if you're in a 2 lane highway and the left lane is clear but you're behind someone doing 40 mph in a 60 mph on the right lane, so you quickly accelerate and switch lanes to pass them. Super defensive drivers who happen to also go unreasonably slow cause a lot of problems and traffic. Also, what about temporarily speeding to get through a yellow light because the alternative is to slam on your breaks and potentially get rear ended. Or coasting a bit above the speed limit because you're "keeping up with traffic" on a major highway with no lights.

tyingq 3y ago
I assume they just don't care about exceptions, and just correlate past data to actual claims to adjust their actuarial tables. But to your point, maybe that also penalizes non-aggressive things that result in accidents.
- nickjj 3y ago
  > I assume they just don't care about exceptions, and just correlate past data to actual claims to adjust their actuarial tables Yep, IMO that's a problem in general. So much is based on case studies in aggregate instead of individuals. The same thing happens with medicine (at least in the US). Just because 200 people in 1 study demonstrated that X year old Y gender people who live in Z have 7.8917% more likelihood to be in an accident vs the average population which was determined by 20 other studies doesn't mean you or I have the same percent.
  - falcolas 3y ago
    > So much is based on case studies in aggregate instead of individuals. Isn't this a good thing? Insurance is about spreading individual risk across a group. Hyper targeting the premiums of individuals according to their individual risk is the exact opposite of spreading risk.
    - nickjj 3y ago
      It depends on who you're optimizing for, insurance companies or individual drivers.
Jimmc414 3y ago
> Super defensive drivers who happen to also go unreasonably slow cause a lot of problems and traffic. Unfortunately I think the entities driving this only care about the specific car they are insuring, not the traffic problems indirectly attributable to the insured.

mint2 3y ago

Which sounds more fair to use in rating insurance: A. Their credit related data B. Their age C. Their marital status D. Actual data on how they drive I feel like D, actual data on how they drive is the most fair. I think there needs to be mechanisms to contest driving events and also regulations on types of driving data allowed, but I think overall it’s by far the most fair and directly related But actually A B and C are used in most but not all states.

kelnos 3y ago
> I feel like D, actual data on how they drive is the most fair. Only if you can actually prove a correlation. I think a lot of people, when asked, "do you people who brake and accelerate hard are likely to be in more crashes than people who don't?", most people might think for a second and say, "yeah, sure, that sounds like it would be the case". But is it, actually? The funny thing is that this data apparently does exist, so it shouldn't be too hard to figure out. But I suspect no one's actually done that.