Fun with my car insurance company's OBD II nanny box

[230RN]

No actual clock in the car.  Apparently they're time-stamping time from software on their end, or at the receiving equipment.

I might try noting the actual time on my watch, deliberately slamming on the brakes, then seeing what time the event shows up on my report.

Another thing I'm curious about is on the downloaded report, they include a map of my area, but I don't see any indication on it of the actual route I took.

  

[brimic]

The newer ones are black boxes, and becoming more so every couple of years.

It was at least eight years ago, and maybe closer to ten, that my brother told me about what the "black box" in a then-current BMW could do. Guy came in with a very noisy engine. Mechanics ripped into it and found most of the valves were bent. Brother (who was the service manager) got to explain to the irate customer that the engine would not be replaced under warranty because said irate customer had been racing the car.

Irate customer absolutely swore that he always drove very conservatively and sedately -- right up until brother showed him the printout from the back box, showing that he had been exceeding the engine's redline by at least 1,000 RPM regularly.

If they could do that then, I don't want to find out what they can do now.

I remember fixing some emissions equipment on a 2004 chevy. Part of the fix was that the engine had to go through a self-diagnostic cycle which happened after the engine stabilized at a certain temperature and has been driven at 50-57 mph on level ground for approximately 5 minutes.
With just a cheap ODBII reader bought from Costco, you can glean a lot of information from events that trip the check engine light- example: speed, engine rpm, ambient temperature, barometric pressure, etc... This was just a snapshot of a moment in time, the OBDII system monitors all of this data in real time, newer systems no doubt read even more activities such as time of day, GPS positioning, and likely blue-tooth activity.
The Ins company reader is likely hoovering up all of the data that they can, using the necessary parts for determining your discount (if any) and selling the whatever data they can to whoever might want to buy it.

[KD5NRH]

With just a cheap ODBII reader bought from Costco, you can glean a lot of information from events that trip the check engine light- example: speed, engine rpm, ambient temperature, barometric pressure, etc... This was just a snapshot of a moment in time, the OBDII system monitors all of this data in real time, newer systems no doubt read even more activities such as time of day, GPS positioning, and likely blue-tooth activity.

But dealing with manufacturer-specific (or in some cases model-specific, as there's nothing preventing Chevy from using any unused PID for one thing on one model and something entirely different on another) PIDs is a PITA when you want a standardized log of several thousand cars.  Not only do you have to know where to look across several different configurations, but you have to know how to convert the raw data from each one, then you have to account for all the ones that don't send a particular reading, update it at odd intervals or whatever.

The Ins company reader is likely hoovering up all of the data that they can, using the necessary parts for determining your discount (if any) and selling the whatever data they can to whoever might want to buy it.

Doubtful; remember they're also having to pay to transfer all that data.  Even at bulk cell rates, there's going to be a big difference between a couple PIDs of 2-6 bytes each per car per update and the hundreds just in the official OBDII standard.  Throw in the custom PIDs that could be several dozen bytes each and you're transferring orders of magnitude more data.

[Marnoot]

Regarding the braking. At first glance one would say they shouldn't ding you for braking events that aren't your fault (car cuts you off). From the perspective of an insurer though, if you're carrying un/underinsured motorist coverage or are in a no-fault state, they don't care:

If you have that coverage or are in such a state, if you have to brake sharply alot whether it's your fault or not, you're a higher risk for the insurance company than someone that didn't have to, thus less of a discount.

[zxcvbob]

I had one of those Snapshot sensors from Progressive on my truck a couple of years ago.  Was almost to the end of the 2 or 3 month monitoring period and had a 30% discount just about cinched. 

We had a huge snowstorm.  2WD pickup in deep wet snow.  Twice that weekend, I spun the tires when I got stuck; the sensor said I accelerated from 0 to 80 mph and back to zero in a second or two.  Rather than filter those physically-impossible readings out as noise, their software gave them extra weight and wiped-out most of my discount.

[KD5NRH]

If you have that coverage or are in such a state, if you have to brake sharply alot whether it's your fault or not, you're a higher risk for the insurance company than someone that didn't have to, thus less of a discount.

Much lower risk, though, than the person who should have braked but didn't, thus avoiding the collision purely by luck.  Eventually, luck runs out.

[Firethorn]

Rather than filter those physically-impossible readings out as noise, their software gave them extra weight and wiped-out most of my discount.

Well, to be fair, hooking JATO rockets to your truck then slamming it into a wall to stop(somehow without contacting the insurance company for repairs...) isn't a very safe activity.   

Much lower risk, though, than the person who should have braked but didn't, thus avoiding the collision purely by luck.  Eventually, luck runs out.

Exactly.  They'll get into an accident and have their insurance rates raised that way. 

The measurements aren't perfect.  Doesn't mean that they aren't useful to the insurance companies - through behavior modification if nothing else.

Remember, braking isn't bad, it's hard/extreme braking.  And you can still have quite a few of them and still get a discount.

[lupinus]

I had one of those Snapshot sensors from Progressive on my truck a couple of years ago.  Was almost to the end of the 2 or 3 month monitoring period and had a 30% discount just about cinched.  

We had a huge snowstorm.  2WD pickup in deep wet snow.  Twice that weekend, I spun the tires when I got stuck; the sensor said I accelerated from 0 to 80 mph and back to zero in a second or two.  Rather than filter those physically-impossible readings out as noise, their software gave them extra weight and wiped-out most of my discount.

[Bean counting devils advocate.]

While physically impossible, still higher risk than someone who sat at home sipping hot cocoa instead of venturing forth into said snow. [/Bean counting devils advocate.]

Which, again, why I won't willingly put one of the damn things on my car.

[KD5NRH]

Exactly.  They'll get into an accident and have their insurance rates raised that way.

Raising the rates after the claim sort of defeats the purpose of all the observation and prediction attempts.

[brimic]

I'm not exactly sure why any gun owners would consider progressive or geico- regardless of lower rates or 'money saving gimmicks.'
Both insurance companies have and continue to fund far left and anti-gun causes.

[KD5NRH]

I'm not exactly sure why any gun owners would consider progressive or geico- regardless of lower rates or 'money saving gimmicks.'
Both insurance companies have and continue to fund far left and anti-gun causes.

Most insurance companies have; those two are just more visible because they're bigger.

[birdman]

I should try one of these in my GTR where I can easily get 20mhp/s acceleration events, 25+mph/s braking events, and 1+g laterals.  Do a few of those at 3am, and I bet I could raise some eyebrows :)

[230RN]

zxcvbob remarked,

We had a huge snowstorm.  2WD pickup in deep wet snow.  Twice that weekend, I spun the tires when I got stuck; the sensor said I accelerated from 0 to 80 mph and back to zero in a second or two.  Rather than filter those physically-impossible readings out as noise, their software gave them extra weight and wiped-out most of my discount.

Good point --I have to remember that.  My six-month "monitored" period ends in November, but most of that period is through summer driving.  I can think of a couple of winter instances where I spun all four wheels, and I like to practice braking on snow just for skill enhancement.

Firethorn said,

Exactly.  They'll get into an accident and have their insurance rates raised that way.

Car insurance is nothing but a very expensive savings account anyhow.

brimic noted,

I'm not exactly sure why any gun owners would consider progressive or geico- regardless of lower rates or 'money saving gimmicks.'
Both insurance companies have and continue to fund far left and anti-gun causes.

It's real difficult nowadays to limit your dealing to non-leftist companies, except for the real rabid ones like Ben and Jerry's and Progressive. (See REF.)

And many companies contribute "all around" anyhow,  to both sides of issues and candidacies.  The world ain't perfect.

birdman offered,

I should try one of these in my GTR where I can easily get 20mhp/s acceleration events, 25+mph/s braking events, and 1+g laterals.  Do a few of those at 3am, and I bet I could raise some eyebrows Smiley

D-minus on "Comportment" and "Plays well with others."  

Terry, 230RN

REF (Read 'em and weep):
http://www.2acheck.com/the-boycott-list/nras-list-of-antis/

[230RN]

END OF STORY

I decided to remove the nannybox a couple of days ago.  I was stunned to find that after a shopping trip, which involved maneuvering around four separate parking lots, pulling over to let an emergency vehicle by, and other non-threatening, not-aggressive normal driving actions, I had a bunch of  "aggressive driving" dings in 17 miles of driving.

Hey, honey, guess what?  I'm about the least aggressive driver there is, especially since I got my CCW ten years ago. (I didn't mention this in my e-mail  to the insurance company, but it's true.)

Anyhow, I e-mailed them politely describing the problem and inquiring whether their instrument was out of calibration or their established thresholds were too low to reflect any kind of reality.

They just said to send it back.

Terry

[griz]

Did they say what the triggering events were?

BTW, I can see where monitoring these things could be a useful tool to identify aggressive driving.  My concern is whether or not they check for a correlation between good drivers using the data, and good drivers using the old fashion metrics such as lack of crashes.  As an example, a driver who runs red lights because they would require heavy braking may get away with it for years, but that would never show on the data.

[230RN]

No, not from a technical standpoint, like how much cornering is bad.  I figured out the 7mph/sec braking threshold because they did give the numbers there.  I didn't look in detail to see if there were any numbers on cornering (in parking lots, as I said) because I just looked at the total number of events and decided that was it and pulled the plug, as it were.  On a followup phone call they gave me the phone number of the tech department of the nannybox maker.  I'm thinking how best to approach that from a technical standpoint*.  I.e., how did they arrive at those threshold levels?  By putting the boxes in the cars of the proverbial little old ladies from Pasadena and recording their driving habits?

On Sundays.

To church.  

And how about slowing down on hills?  There's a couple of mild hills around here where my cruise control won't hold the car back from 35 mph by mere compression from my tiny little engine, so I downshift to a lower gear instead of riding the brakes to allow it to control the speed. So there's a slight downshifting transmission braking action event at that point which I spotted by correlation with the time.  And since it's downhill, if they have accelerometers in the box, that would make it look worse by adding the sine of the hill angle times gee to the braking deceleration.

See, another question for their tech department.

No kidding, on testing and correlating incidents like that, I think if you really kept off their "event" thresholds, you'd probably cause accidents.  I'm sure you've seen drivers like that.

Again, no kidding, I'm a pretty smooth driver and most certainly not aggressive, and I don't want those "aggressive driving" dings on anybody's records anywhere.

I didn't think the customer service rep at the insurance company would know anything about the tech details of the box, like about the 7mph/sec braking threshold and she didn't, so she gave me the nannybox company's tech number.

My paranoia neurons are thinking that maybe they set the thresholds so low so the company can deny coverage in case of accident:  "You had 17 'aggressive driving' incidents, so it must have been your fault" or some such.  Look, I laugh at my own statement haha>

Yeah, way out there, but...

Terry

*Or I may just say pisonit and fugedaboudit.

[Tallpine]

I wonder how the nanny box would interpret a 4x4 being driven over something like Black Bear Pass 

 

[KD5NRH]

BTW, I can see where monitoring these things could be a useful tool to identify aggressive driving.  My concern is whether or not they check for a correlation between good drivers using the data, and good drivers using the old fashion metrics such as lack of crashes.  As an example, a driver who runs red lights because they would require heavy braking may get away with it for years, but that would never show on the data.

This.  AFAICT, the only criterion for "aggressive driving" is braking more than 7mi/h/sec.  That's really not all that aggressive, and doesn't appear to consider starting speed.  i.e. stopping in less than one second from an initial speed of 8mi/h because you were driving extremely defensively past the playground when a ball rolled out in the street still gets you dinged.  Running over the kid chasing the ball to avoid nailing the brakes only costs you if you get caught.

[Firethorn]

This.  AFAICT, the only criterion for "aggressive driving" is braking more than 7mi/h/sec.

That's probably why Allstate doesn't call it 'aggressive driving', instead a much more restrained 'hard braking' or 'extreme braking' event.  The more events, the higher your insurance cost.  They also outright rate you on driving distances modified by driving time(really late at night costs more than during the day) and speeds above 80mph.

[230RN]

Well, that 7 MPH/sec was the only one I actually calculated from the numbers given.  

When I got my latest report and saw all the "aggressive driving" dings, I said the heck with it and didn't bother to go into the numerical details.  I just compared their recorded times with what I was actually doing in the car at those times.

They were almost all parking lot maneuvering, where hard turns are necessary (even at low speed) to find and slip into a space.  That particular day all the handicapped spots were taken and I had to (carefully!) cruise around a bit to find a spot closer to the store.  Oh, and I got gas at their station in the same parking lot... which took more maneuvering and circling around until a pump on my gas tank's side of the car opened up.

In addition, I made a few more stops at other parking lots, again with the attendant maneuvering.  

There was also one on the road, where I had to pull over to let an emergency vehicle pass.

As I said, I don't know how they arrived at their criteria for the various "events," but they do not seem to conform to normal safe driving techniques.

Like I've bragged before (knock wood/throw salt over my shoulder) I've gone for much more than 25 years with only one ticket (missing front license plate and it doesn't even appear on my record) and got rear-ended at a stop sign by some yahoo who was on the phone and driving his girl friend's car.  Guess what his excuse to the cops was... "I didn't expect him to actually stop."

 

In addition, I drove a cab for more than a year with only one not-my-fault minor incident.

I also drove motorcycles for a number of years and I sure learned defensive driving from that two-wheeler experience.  (No tickets there, either.)

As I've often said, about 25 years ago I learned to drive, as opposed to merely operate a vehicle.

There's just something really goofy about their criteria, period.

Terry, 230RN

[Tallpine]

They were almost all parking lot maneuvering, where hard turns are necessary (even at low speed) to find and slip into a space.  That particular day all the handicapped spots were taken and I had to (carefully!) cruise around a bit to find a spot closer to the store.  Oh, and I got gas at their station in the same parking lot... which took more maneuvering and circling around until a pump on my gas tank's side of the car opened up.

In addition, I made a few more stops at other parking lots, again with the attendant maneuvering.   

That's why I was wondering how it would do if you were driving "off road."   

I bet they would really have a hissy fit if they saw you backing up on the Black Bear switchbacks 

We only have one OBD2 vehicle (out of five) and I am really thinking about trading it off for something older.  We almost traded it for the jeep, but decided to hang onto it a while longer just for the better gas mileage.  But we don't drive the car anymore unless we're going to leave the local area and get out on the highway.

[Firethorn]

As I said, I don't know how they arrived at their criteria for the various "events," but they do not seem to conform to normal safe driving techniques.

Either they're running their program different than Allstate, or there's something seriously wrong with their monitor, or you're not presenting your story well and braking severely harder than expected.

Because my monitor doesn't trip on turns at all, and at normal parking lot speeds, shouldn't trip at all.

In my truck, I have to brake noticeably hard to trigger an event.  I average ONE event about every 3rd trip.

[Brad Johnson]

Doea anyone make an app that mimics a nannybox so people can "test" their driving before actually committing to one of the damn things?

Brad

[Firethorn]

Doea anyone make an app that mimics a nannybox so people can "test" their driving before actually committing to one of the damn things?

Brad

Why?  I got a flat 7% discount just for signing up, even before they had any driving data on me!

[230RN]

Firethorn offered:

Either they're running their program different than Allstate, or there's something seriously wrong with their monitor, or you're not presenting your story well and braking severely harder than expected.

That's why, in my e-mail to them, I mentioned that it might be out of calibration.  On the "presentation" aspect, I'm telling it like it is... and I fit in well with drivers around me, so maybe, but I don't think so.

Brad Johnson asked,

Does anyone make an app that mimics a nannybox so people can "test" their driving before actually committing to one of the damn things?

KD5NRH had some possible insights on that in an earlier post. *

Firethorn then remarked,

Why?  I got a flat 7% discount just for signing up, even before they had any driving data on me!

Well, I did, too, 10%, but if my normal provably safe driving results in being incorrectly or unfairly tagged as "aggressive," on anybody's records anywhere, I don't like it.

Terry