-
Want to remove ads? Register an account and login to see fewer ads, and become a Supporting Member to remove almost all ads.
Safety Score: parameter dependence
- Thread starter Michael_Vinson
- Start date
It doesn't seem to me we should be penalized in any way by any kind of (manual or regenerative) braking on a steep hill to keep in the speed limit. This is kind of dumb.
Daniel in SD
(supervised)
Only the steepest street in the country would give 0.3g while maintaining speed. That would be a 30% grade.
It's possible that people who drive in areas with hills have more collisions. The Safety Score (beta) does not measure only driver skill, it measures predicted collision frequency which is a function of driver skill and driving environment. Note how the same driver will pay a different insurance premium depending on where they live.
FSD Beta is dumb so using a dumb metric to decide who gets it seems appropriate.
Siddhartha
Member
I was definitely surprised with my score yesterday. Granted I don't drive much (maybe about 10miles/day right now), and my score went from 100 down to 62 with one FCW event. It wasn't my fault-crossed an intersection, traffic, then abrupt slowdown to another light. Happens. I wasn't following closely, and I got the warning. I have been driving like a grandpa, and was surprised when I checked to see that drop in my score.
One more drive, and it's up to 82, which is nice, but that was definitely startling to me.
One more drive, and it's up to 82, which is nice, but that was definitely startling to me.
Fernand
Active Member
I almost got into a crash last night as I was tooling down the freeway in NOA. Someone inattentively swerved into my lane and while NOA was reacting I proactively dropped AP and braked. In these situations AP does very well, but it's 50/50 as to who reacts first. That registered as 1.3% hard braking for the day. It's not a serious ding but I wonder: if I'd let AP complete the evasion it would (obviously) also have had to brake, so would THAT count as hard braking? Is it better for your score to just CYEP? close your eyes and pray?
Last edited:
Cheburashka
Active Member
Bullsh¡t.
I got 100 score every day driving an average of 30 miles a day.
My wife drove the car once for 20 miles and brought my score down.
My unsafe following score is fixed now at 9.3% which was my wife's score during that short drive. No other driving compensates for that even if it's perfect.
The safety score is more rigged than the election.
I got 100 score every day driving an average of 30 miles a day.
My wife drove the car once for 20 miles and brought my score down.
My unsafe following score is fixed now at 9.3% which was my wife's score during that short drive. No other driving compensates for that even if it's perfect.
The safety score is more rigged than the election.
Attachments
They look at their historic fleet data, then find all the accidents, then see what pattern emerges from those accidents.
While it might look like AP disengagement and hard breaking have the greatest impact by the weights, in reality, the front collision warnings (FCW) are what hurts the most. It's really hard to get a AP jail event. You have to purposely ignore all the AP warnings to get that one. Hard braking dings can be nullified same day by some low speed full regen decelerations on a quiet road. FCWs require a ton of same-day mileage to nullify. Avoid those as much as you can (I fully realize that in some locales, this is really hard to do).
Unsafe following doesn't count if you are on autopilot. It is the percent time you are 1-3 seconds behind the lead car minus the time you are within 1 second. Not easy, but can be changed.
I took a long drive to get the denominator up (miles driven) and had to do a 'both scroll wheel' reset (while driving) because my wipers were on for no apparent reason. It fixed that issue but I then noticed it didn't count the drive in the safety score. Is that because I lost connection while reseting the screen?
Cheburashka
Active Member
I use the scroll wheel reset all the time.
Doing it during a drive excludes it from your safety score. Very useful if you get a FCW or AP strikeout.
The point you are making is correct, but your wording isn't exactly accurate: it's the time spent following within one second divided by the time between 1 and 3 seconds. That is, it's the ratio not the difference. As some have noted, the way to improve this score is to spend a lot of time following within 3 seconds but not as close as 1 second, so as to increase the denominator without affecting the numerator of the ratio. (And, in fact, unsafe following is the least impactful to your overall score, so it isn't nearly as important as hard braking. Which is why novox77 correctly advocates for doing "some low speed full regen decelerations on a quiet road" -- this is absolutely a good strategy if you have non-zero hard braking time on your current trip).
Michael
Well given that every study and recount has proven the election was not rigged, then your comment holds no real value. I bet you think vaccines cause you to be magnetic, that it’s the mark of the devil, and that an RFID chip is implanted into all those getting vaccinated.
Fernand
Active Member
Here's a good pair for you. First I got ding'ed for hard braking and turning while driving one of the windiest roads in Northern California today, Highway 29 from Calistoga to Middletown. Too windey to use Autopilot most of the way, with hairpin turns up and down a mountain. There was no way to avoid some hard turns and slowing unless I drove 10 MPH. That sucks. I was already holding up traffic as it was.
But the really bad one is that I left a restaurant in Novato and drove less than 1/4 mile manually with other traffic at under 25 MPH to the 101 freeway, then engaged NOA and drove 28.6 miles over the Richmond bridge to the Emeryville Supercharger, always in autopilot. On that trip I got 60% Unsafe Following. How is that possible?
First of all, I NEVER tailgate, I'm paranoid. No idea what the car logged as unsafe following. But the real bitch is that I was in manual (and subject to scoring) only for less than 1/4 mile. So evidently that 1/4 mile was used to assign 60% Unsafe Following to the whole 28.8 mile trip, of which 28.6 was entirely in NOA. That's not right.
But the really bad one is that I left a restaurant in Novato and drove less than 1/4 mile manually with other traffic at under 25 MPH to the 101 freeway, then engaged NOA and drove 28.6 miles over the Richmond bridge to the Emeryville Supercharger, always in autopilot. On that trip I got 60% Unsafe Following. How is that possible?
First of all, I NEVER tailgate, I'm paranoid. No idea what the car logged as unsafe following. But the real bitch is that I was in manual (and subject to scoring) only for less than 1/4 mile. So evidently that 1/4 mile was used to assign 60% Unsafe Following to the whole 28.8 mile trip, of which 28.6 was entirely in NOA. That's not right.
Just relax guys. This isn’t FSD. It’s a level two update that will be as wonky as NOA. Yes I pushed the button, but if I get it, I won’t use it any more than NOA as NOA is really dangerous on my car. This has little to do with Tesla supplying a self driving system as specified in the contract.
Cheburashka
Active Member
So I need to work on my unsafe following parameter.
Am I correct in understanding that I will only be able to improve it by driving without AP and between 1-3 sec following distance?
Doesn't seem like maintaining a 3+ distance or AP helps this parameter in any way.
Am I correct in understanding that I will only be able to improve it by driving without AP and between 1-3 sec following distance?
Doesn't seem like maintaining a 3+ distance or AP helps this parameter in any way.
n.one.one
Info w/o Drama
Yep...that seems to be the general consensus. And remember, the corrective driving needs to be on the same day as the ding.
Last edited:
Cheburashka
Active Member
Very, very logical....
Seems like I won't be getting FSD for a while then. I'm not sure for how much longer I'll be ae to maintain my 98 score.
Similar threads
- Article
