Hacker Newsnew | past | comments | ask | show | jobs | submitlogin
Tesla Safety Score Beta (tesla.com)
33 points by sbuttgereit on Sept 25, 2021 | hide | past | favorite | 41 comments


I love this.

There are so many people who truly believe ridiculous bullshit with regard to driving.

People have been telling me that SUVs are safer cars (questionable) or that higher powered cars are safer (highly questionable) or that driving slower than the speed limit increases the risk of accidents (no comment).

If this score provides real time feedback to people based on real world data, maybe it can help change somebodys mind. Maybe it can nudge people to become slightly safer drivers.


>that driving slower than the speed limit increases the risk of accidents

Nit: this is accurate when all others are driving the speed limit (or greater), given the likelihood of accidents increases as speed differential increases.


> higher powered cars are safer (highly questionable)

there's a good correlation (not 1:1 but still, quite good) between car performances and their ability to corner and their braking distance. a fast car driven by an idiot is going to make him more dangerous, but in the hand of an average driver will serve him well in all the emergency maneuver he'll find himself doing.


The second day with my Camaro I had to full on hard brake to avoid a deer. My previous car would have smashed through it.


higher powered cards tend to be high-end and thus have better safety systems built in. that‘s why you often hear of people going 250km/h in some top of the line mercedes, crashing and getting out eith barely a scratch.

and how are SUVs not safer, given they have more mass and a larger crunch zone? also, since the driver is seated higher he can see more what is going on in front of him (a fee small cars ahead)


> and how are SUVs not safer, given they have more mass and a larger crunch zone?

That helps in a head on collission with a smaller car. But it's even safer if you don't collide with anything in the first place.

SUVs have longer breaking distance because of their mass, and because they are higher they don't corner as well and are at higher risk of tipping over.

These factors increase the risk of dangerous accidents versus a normal car.

Also, around 50% of lethal accidents don't involve a second car at all. If you crash into a concrete wall, or if you run off the road, it doesn't matter that your car is bigger, you are still screwed.

The safest car is the one that doesn't get into an accident in the first place, and bigger cars are not necessarily better in that regard.


He might mean not safer for society since more pedestrians and kids are killed by SUVs.


One of the safety factors used here is "Forced Autopilot Disengagement" which I've experienced a bug with that can trigger if you take control of the car using the steering wheel without disengaging via the control stalk. I have a habit of taking over control of the car on a freeway offramp by just turning the wheel, and around 1-2% of the time the car freaks out and gives the disengagement alert thinking that I'm not controlling the car. I suspect that bug will make my safety score worse.


Insurance companies will love this!


If I’m a safe driver, I don’t want to pay more for poor drivers. If this helps insurers more accurately price auto insurance while providing direct financial incentives to poor drivers to improve, I fully support it. This is Econ 101.


Just wait until something breaks or the system determines that you are a bad driver because instead of looking at the road you looked right of left (in the mirror), just like that Amazon story from a couple of days ago.

I seems that we are heading into this weird world where, like in the movie "Brazil" some weird bug in one of our surveillance and human management systems will end up destroying lives, leaving people completely helpless and plagued for their entire existence.


That’s a regulatory issue, not a technology or implementation issue. You prevent that outcome with laws, not code. Auto insurance is already heavily regulated in the US.


How, system making bad decision, is regulatory issue?

You probably won't be able to prove that decision was incorrect in the first place.


Por qué no los dos?


Train self driving models from safer drivers. Encourage drivers to be safer. Seems smart to me. I'd like to see a vehicle crashes per 100k miles before and after the safety score, and by safety score quartiles.


Not a fan but seems inevitable. My personal feeling is that this won’t lead to decreases in rates but just increases in rates on those who are deemed unsafe. I’d want to disable this if possible.


GM did something just like this years ago and in fact looking around they expanded it into an entire insurance program with specific discounts for their vehicles:

https://www.onstarinsurance.com/

OnStar equipped vehicles generate a safety score, but your score will only be shared with the insurance company if it was high enough to earn a deduction.

And since it's an opt-in program, the lack of a score provided doesn't give them a datapoint either.


This is a great move as these sort of metrics will unconsciously push individuals to become better drivers in the same fashion weight balances help you lose weight.


Surprised no one has made the connection that this score will be used to determine the extent of FSD is available to you, which seems a great solution to the scary mass-distribution of the vehicle-as-a-weapon approach.


Between the yoke steering wheel, the farcical promise that is FSD, the lack of physical knobs, the lack of CarPlay/Android Auto, and this new mess of privacy/surveillance issues, I will pass on Tesla. I was ready to make my next car a Tesla despite the other flaws but I think I’m over the tipping point and don’t want to reward those normalizing these things.


Yet another reason not to buy a Tesla, or any car which does this.

Do we really want these kind of scores to determine our insurance rates, outcomes in the case of an accident and so on?

Insurance is obvious, but I won't be surprised if in the case of an accident arguments are made using such a score to determine who was at fault.

OTOH, I wonder if this is a precursor to Tesla launching its own car insurance program.


> Do we really want these kind of scores to determine our insurance rates, outcomes in the case of an accident and so on?

Why wouldn’t I? Seems much more fair than criteria such as age, education and marital status that are currently used to determine rates.


Few points:

1) Assuming you're in the US - Progressive has been offering this for years. Obviously not everyone is enrolled with Progressive, nor do I know how many of Progressive's customers are enrolled in it. But, if you think that it's good, you may want to look into it, again assuming you're in the US.

2) Everyone who drives extremely carefully, say the medical equivalent of young-healthy person, would want to enroll. However, rest of the people won't. Insurance companies want to balance their high-risk with low-risk customers, which is where I think it'll run into issues.

3) The issue comes when your insurance provider has access to detailed data. Say there was an accident where your insurance company would normally pay. But now, they will be able to dig into data and find some clause that they will use to deny your claim and find you at fault. And, therein lies the problem for everyone and not just bad, or less than excellent, drivers.


> The issue comes when your insurance provider has access to detailed data. Say there was an accident where your insurance company would normally pay. But now, they will be able to dig into data and find some clause that they will use to deny your claim and find you at fault.

Buy comp and coll coverage and they won’t deny, they’ll duke it out with the other party’s carrier to put you not at fault if reasonable evidence shows that


> which is where I think it'll run into issues.

Like what? Genuinely curious.


Because your interest and those of the insurance company are not aligned and in case of a claim are adversarial.

Given extra information to your adversary is foolish.


It‘s basically a case of „I swear I am a good driver, even though I am a 22-yo in a sports cars, please why won‘t you just believe me“

now you have a way to prove it


Do you really want some out of touch rich developers and researchers that never interact with most of the real world determining various things that can ruin your life?


Nah, we should just keep charging people insurance rates based on their credit score because that’s more fair… and not use something that actually is more tied to driving and doesn’t arbitrarily penalize all poor people


What a fair reply! Acting like by protesting intrusive always-on driver monitoring... you are automatically defending the use of credit scores to adjust rates!

No one could possibly be against both...

Not at all a straw-man argument of any sort!


Always on driving risk monitoring to price a product that is always on when driving seems fair.

And it’s much more likely to lead to improved driving behavior than charging all 25 year olds more. Or all people with low credit scores more.

What are you going to price on if it’s not directly measuring how one drives? A blanket flat makes no sense.


> A blanket flat makes no sense.

What do you think rates are based on now?

Either this is another straw-man in the making or you genuinely think they were only based on credit scores?

It's not even allowed to be a factor in some states, and it's not allowed to be the only factor in many more...

-

Driver age, home zip code, car driven, miles driven per year, accident history? Do none of those tell anyone anything?

Always-on driver monitoring that's not even adjusted to geography... so someone in midtown Manhattan dodging jaywalking tourists has FCW compared on the same scale as someone on 10 miles from their nearest neighbor in the middle of Wyoming... you think that's more fair?


I specifically called out age in a way that shows it’s unfair, and then one of your example of “fairer than actual driving monitoring” was age?!

And mr Wyoming in your example thinks it’s fair he’s higher for driving more miles than city man? (Assuming we aren’t pricing different based on location bc your comment made that assumption) and how does one verify miles driven? Best way is oem vehicle integration like you object to.

Home zip code is pretty discriminatory despite being allowed and also kinda unreliable.

Accident history is okay but not perfect because fault can be hard to attribute.

Some Speeding Tickets can be gotten out of or erased, and out of state ones are hard to locate

No, all those are very imperfect proxies for driving behavior that are often discriminatory against poor people and minorities. It’s better to monitor ACTUAL driving behavior than penalize minorities because cops target them for tickets. Or poor people who drive okay but don’t have 800 credit so they get charged more. If only there was some way to monitor actual driving instead of these problematic proxies…


It looks like they're already rolling their own insurance program out: https://www.tesla.com/insurance | https://www.tesla.com/support/insurance, currently limited to California.


1) Given Tesla's track record in blaming FSD related accidents on drivers, does one really want them to be the insurance company where their and a driver's incentives are obviously mismatched.

2) One of the issues, once the technical issues related to FSD are resolved, is insurance. So, will/does Tesla's insurance have a clause that says that in cases where FSD is at fault, Tesla is the liable party?


Can you please explain in more details why do you think this is bad?

Right now, both careful drivers and aggressive drivers pay the same amount for insurance, which means that careful subsidize aggressive ones. Tesla scoring separates those groups and promotes more safer on the roads.


That's not true, aggressive drivers' insurance trends upwards and careful drivers' insurance trend downwards.

I have no problem with systems like this but making it always-on and non-optional is insane...

The score isn't even geographically adjusted, they posted the formula right there. Good luck my fellow New Yorkers who have to deal with 100x the pedestrians and random obstructions the average driver has to...

Hard turning is punished so good luck if you're like me and enjoy going on curvy roads where you can easily experience .4gs of lateral force at the speed limit...

Our insurance already has ways to model this stuff in much smarter ways. People in New York pay more for insurance without insurance companies needing every single vehicle's data because driver age and historical data exist...

Their comparable programs are opt-in and often can only reduce your rate (I know my old Volt would only share your score if it'd result in an insurance deduction)


Please see my reply to sibling, @danhak's comment, above.


This is optional and opt in, not sure what your problem is.

Tesla will use it for their own insurance but if you don't have that or want FSD Beta you want have to activate this.

> OTOH, I wonder if this is a precursor to Tesla launching its own car insurance program.

They had this for a while.


Tesla is not alone on this. Other manufacturers are starting to enable similar levels of tracking, often automatically opted in and you have to go through several steps to opt out. In a few years, the only option with be not to drive or buy a classic car.


Do you have more information on which all manufacturers are doing this? It'd be very useful information in making a purchase decision.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: