What's your Tesla FSD Beta Insurance Score

[KaneBlaisdell]

Hi All,
I'm just curious to see how others are doing with their Tesla insurance scores? What's your score? How many miles have you driven? What have you found to be most difficult about maintaining a good score? Any helpful tips you have learned?

So far I have driven 600+ miles and have maintained a score of 99. A lot of my driving has been on Autopilot / Autosteer. One tip that has helped me, is pay attention to your following distance while merging onto a highway and going over 50mph (Before engaging Autopilot). While on Autopilot you won't get penalized for following distance if someone gets too close to you, but you will if you're out of Autopilot and going over 50mph. Also, I have not had any situations where regen breaking penalizes you for stopping too aggressively, so I recommend one pedal driving as much as possible. Just stay far back from other cars to give yourself plenty of time to react to situations.

 

[FalconFour]

Laff. Just stopping by to comment on the "Tesla Insurance Score" bit. I've had Tesla insurance since it first came out, and there's not yet been a single peep about actually using that score, or any score - they just say they're going to, eventually, maybe. The Safety Score, for now, is purely an FSD Beta mini-game. I really wish they would start using some kind of scoring, though!

My tips are your tips, and in a similar position (though not as many miles). Similarly, though, I'd add that many people are missing the idea of "the window" - having to stay INSIDE a 1-3 second following window, braking between 0.1-0.3g, turning 0.2-0.4g... otherwise driving like granny doesn't help you.

 

[gearchruncher]

I've had Tesla insurance since it first came out, and there's not yet been a single peep about actually using that score, or any score - they just say they're going to, eventually, maybe.

That's because using this "score" in CA is illegal, and is the only place Tesla offers insurance right now.
Plus, Tesla is all over the place. When they did propose it as an "insurance" score, it had acceleration in it. Now that it's just a game to get FSD beta, suddenly acceleration is not statistically significant.

 

[FalconFour]

That's because using this "score" in CA is illegal, and is the only place Tesla offers insurance right now.

honestly as a Californian born and raised, and absolutely hates how many drivers flout laws and rules and know there are no consequences, I absolutely loathe that CA has any kind of rules preventing driving behavior from being used for insurance. The one thing that could force self-absorbed A-holes into driving less like the most important person in the universe... and CA doesn't allow it.

I whole-heartedly hope Tesla figures out a way around that, as they've done with many other things.

 

[gearchruncher]

I absolutely loathe that CA has any kind of rules preventing driving behavior from being used for insurance.

Really? You hate that a completely arbitrary, opaque, statistically invalid, non-peer reviewed, full of false positives, completely without context data collection method can't be used to set individual rates?

I assume you voted against Prop 103 in 1988, which actually requires the CA insurance commissioner to approve insurance rates so they aren't arbitrary and capricious, and allows public review of the rates? And that led to:

According to the California Insurance Commissioner, Proposition 103 "has saved consumers billions" since being implemented, specifically a $4.29 billion per year dividend. It also claims "Californians spent 0.3% less on auto insurance in 2010 than they spent in 1989, while the nation spent 43.3% more". The Commissioner quotes a 2013 report of the Consumer Federation that more than $100 billion had been saved by consumers in the 25 years after passage.

Insurance rates are supposed to be set based on actual losses, not just behaviors we don't like but don't actually lead to crashes. If turning quickly actually is a good predictor of higher losses, than so be it. But it's suspicious that Tesla knows if you're speeding or running a stop sign, and doesn't think this is important, and thinks braking "hard" is bad, but accelerating hard is fine. Seems like maybe they'd have a hard time defending the current algorithms as the best way to assess driver risk. There are also huge context issues, such as the fact that taking your car to a driver education event that occurs on private property will count against your rates even though none of those events occurred on a public street.

Tesla can figure it out- they just need to clearly show to the CA legislature that this method is actually functional in estimating losses, and that it is not discriminatory. CA doesn't allow gender for instance, and lots of other programs have used data like "during what time of the day do you drive?" which is actually just a really good proxy for your income. If they really have 6 billion miles of data that estimates loss rates to a statistically valid accuracy, then they should present it. But to do that, they're going to need some statisticians that are better than the ones on the current "Teslas on AP are 10X as safe as other cars!" argument.

The fact they have not, and already say that their current method is expected to be updated with more data, tells us that they are not at a point it should be used to set individual rates.

Here's the law: CA allows CONVICTIONS and at fault accidents, and total mileage driven as the primary rate factors. They allow years of driving, type of vehicle, actual claims, marriage, etc also. Seems like in fact, they actively do allow "driving behavior" to be used. It's just not driving behavior collected by a blind accelerometer.

Section 2632.5 - Rating Factors, Cal. Code Regs. tit. 10 § 2632.5 | Casetext Search + Citator

Read Section 2632.5 - Rating Factors, Cal. Code Regs. tit. 10 § 2632.5, see flags on bad law, and search Casetext’s comprehensive legal database

casetext.com

 

[FalconFour]

Yeah, I don't give an arse about the what-ifs and "but they coulda shoulda mighta, it's so opaque, I want my mommy" garbage. We need enforcement. People need to be held accountable. So freaking sick of people having absolutely no regard for rules and concern for others, absolutely had enough of it. People making excuses for why we can't have accountability are the ones out there hoping accountability never catches them.

Practically every other state has this ability in place. We need it here too, with the number of drivers that don't care about rules and safety in the slightest.

 

[gearchruncher]

We need enforcement.

And you want enforcement to come in the form of a private insurance company charging more to people based on something that monitors how hard they brake? Not from law enforcement officers? Not from actually taking licenses away from people? Just charging them more money to increase profits at a private company?

So freaking sick of people having absolutely no regard for rules and concern for others, absolutely had enough of it. People making excuses for why we can't have accountability are the ones out there hoping accountability never catches them.

It appears you also want this to be REQUIRED, not optional? I mean, every insurer I have ever heard from offers data collection as an option, not a requirement, and they all offer it as a discount not a surcharge. So if you are one of these people you theorize that wants to avoid accountability, you'd just decline signing up or sign up for a different carrier.

Plus, you're asking for "accountability" to someone that supposedly drives dangerously, but isn't getting tickets or getting in accidents. Because when those happen, there IS accountability and higher insurance rates. So what are you asking them to account for?

Hoping to improve driver behavior in the ways you appear to be frustrated about via insurance data collection instead of other methods is just so odd.

Practically every other state has this ability in place. We need it here too, with the number of drivers that don't care about rules and safety in the slightest.

The deepest irony here is that CA is one of the safer states per mile. So apparently this data collection method doesn't really work.

Fatality Facts 2021: State by state

A summary of fatality statistics about state by state compiled by IIHS from 2021 Fatality Analysis Reporting System (FARS) data.

www.iihs.org

 

[nvx1977]

Fatalities are different than collisions. The most congested areas tend to have fewer fatalities because average speeds are down. Look at all the states where traffic is notoriously bad (CA, MA, NY, etc) and you'll see low per-capita death rates.

But a sh*t-ton of collisions. And we are dealing with a metric for collisions.

 

[gearchruncher]

But a sh*t-ton of collisions. And we are dealing with a metric for collisions.

Ok, why is insurance in CA not way more than other states if they have that many more collisions? Insurance pays for collisions.
Gonna need a source that California would have a lot less collisions if they just allowed insurance companies to use data to set rates. I mean other states do, are you saying Miami, Phoenix. Houston have less collisions and better behaved drivers?

Remember, this convo is about the fact that California blocks "accountability" while other states allow it. If that works, statistics should show this quite easily.

 

[alaminsumon]

In Virginia they are considering my score, I have driven 1872 miles so far and my score is 98 (Last month it was 97 and I paid 71.81, if I can keep 98 rest of this month my projected payment on May should be 68.14 according to the app), I mostly got penalized for driving after 10pm, and forward collusion warning (I don't think that was correct, A car was making left turn from a street where I was turning right on a red, and it creates that warning and penalized me), and Closely following distance ( In our area you can't be that far behind if you want to get yourself on time during highly congested traffic hour. I think they needs to work on scoring system. Specially 10PM cut off time is ridiculous. I had to stop taking my car to dinner parties. They should make it at least 12AM.

 

[gearchruncher]

I think they needs to work on scoring system. Specially 10PM cut off time is ridiculous. I had to stop taking my car to dinner parties. They should make it at least 12AM.

Thought experiment:
If 10PM is ridiculous, why is 12AM better? Why is any time cutoff acceptable?

Is it because you think of yourself as a good driver, and you drive between 10-12, so that must be OK, but you don't drive after 12AM, so that must be when the bad drivers are out? Is this a "as long as it doesn't impact me, I'm OK with it" situation?

I know. Why can't Tesla just develop a great safety score, which is meant to accurately represent insurance risk, while also completely ignoring actuarial data? It's not FAIR that they use DATA to say driving after 10PM is unsafe. I'm a safe driver! I want cheap rates!

You can't simultaneously want to use a supposedly data driven algorithm to get cheap insurance rates while also complaining about the logic they use for it, especially when it's working perfectly. I mean look at it- it caused you to change your behavior, exactly what an insurance company wants. Tesla insurance tells you driving after 10PM will cost you X. That cost is higher than you want to pay. So you don't. It's basically a perfect system.

 

[legendsk]

My score is staying at 99 now. Sadly, I can't eat at Burger King anymore because waiting in the drive-thru queue would knock it down 3-5 points. Apparently the actuarial data indicates waiting in line at BK is very unsafe, who would have guessed?

 

[gearchruncher]

Apparently the actuarial data indicates waiting in line at BK is very unsafe, who would have guessed?

Looks like Tesla insurance has partnered with your health insurance company

 

[clydeiii]

I hover between 98-100. 99% of my miles are on AP.

 

[alaminsumon]

I hover between 98-100. 99% of my miles are on AP.

I was like that, however today I returned home after 10pm and on the way I had to take over FSDb by breaking while it was turning right on left in a un safe manner, I thought that won't be count against non autopilot breaking because it was within the 5 second cut off time, but I guess I was wrong, that one hard brake and 15 minutes of driving after 10PM nock off quite a bit of my point, 98 to 54 for the day and 96 from 98 average, now it is showing my May 2023 insurance will be 76.87 it was 72.18 last month (with 97 score) 2 days ago my score became 98 and all my hard earned points went down today .... LOL stat shows over 80% of my drive was on AP/FSD (I have little over 2k within 2 months of delivery).

 

[clydeiii]

I was like that, however today I returned home after 10pm and on the way I had to take over FSDb by breaking while it was turning right on left in a un safe manner, I thought that won't be count against non autopilot breaking because it was within the 5 second cut off time, but I guess I was wrong, that one hard brake and 15 minutes of driving after 10PM nock off quite a bit of my point, 98 to 54 for the day and 96 from 98 average, now it is showing my May 2023 insurance will be 76.87 it was 72.18 last month (with 97 score) 2 days ago my score became 98 and all my hard earned points went down today .... LOL stat shows over 80% of my drive was on AP/FSD (I have little over 2k within 2 months of delivery).

It all averages out over time. Ultimately best to forget the score even exists.