Safety Score

[AlanSubie4Life]

When I first starting using the app on android (side loaded immediately), there was a fleet average value for each of the five parameters. It was in a light gray that was hard to see. It was a little bar graph and a floating point average value. It was right under the bright green bar graph of each parameter. It's no longer there.

In the screen captures I've seen from Android on the main page, there's a little down arrow you can tap to see that info (just like in the iPhone app). These are fleet averages, not averages amongst those competing, so they're not very informative.

 

[johnmodely]

I really hope the rank/selection system is millage based. For example they add some sort of weighted millage to the calculation. Would not be fair if they let in ALL the people with 100 in first even if a large majority have like 10 miles on the scoring, since only 10 miles of driving does not say much. I also bet a lot of people are keeping their score at 100 and trying to put as little driving as possible to keep the score there.

I think someone with lets say a 95 score that has driven 200 miles should be ranked much higher than a 100 score at 10 miles...

 

[scottf200]

this is just really messed up. For those of us that are driving and monitoring, not the backseat drivers, this new behavior is awful.

Solution would be to have an option of some kind for this, or better on-screen indication of what it's seeing and thinking. Sacrificing driver's comfort for passenger's comfort - without any option otherwise - is a really dumb move. If it were an option, I'd probably pick the middle ground closer to "rigid".

IMO, more people should be pestering Tesla to use Chill or Standard mode for this. I've suggested this in various ways and I've seen a few others.

 

[FalconFour]

agree... with the work we're all putting into this as our only car, I sure as hell hope it's not like... guy with 3 cars drives at night for 10 miles, gets a 100, gets FSD before I do, with daily commuting and wrangling 100 from the grips of doom.

 

[mhan00]

It occurred to me that with some modifications the safety score could be used to measure FSD Beta's safety. All the metrics, except for "forced AP disengagement", are good preventive metrics for safety. They would need to change the "forced AP disengagement" to "manual disengagements". But I could see Tesla using the Safety Score to gauge the safety of the system.

I also wonder if Tesla might start including the average safety score of users in their quarterly AP Safety reports.

If they’re really going to be releasing it to 1k drivers a day, then they’re going to need an automated method of some sort to kick people off for being unsafe, and probably going to have to err on the side of caution to try to do so before a major accident occurs. Having the safety score apply to all driving, human and Beta, might do it. Inform new Beta users that it’s their responsibility to monitor the system and make sure to intervene anytime it’s about to undertake an action that would ding it under the safety score metrics because the scoring system is being used to evaluate Beta use and will be combined with the user’s normal driving score. and if that combined score drops below a certain level then the user is kicked off of Beta. Much howling would ensue, but safety should be prioritized as much as possible.

 

[kbM3]

This.

Currently they are not grading drivers - by figuring out where it would be safer to deploy FSD Beta early access.

But, IMO, unintentionally.

What they’re grading is probability of collision, which measures both driver and environment.

IDEALLY what they should grade would be the following FSD supervisory skills:

1) Attentiveness: How the driver behaves when on AP or FSD. Cabin Cam + AI might be good, but not all Teslas have cabin cam.
2) Effectiveness: Driver intervenes appropriately and effectively when necessary. More difficult to measure as interventions become rarer

Unfortunately those are harder to monitor than their current system. Possibly measuring lateral or longitudinal jerk (sudden braking instead of smooth and swerving instead of smoothly turning) might be effective at determining who’s more likely reacting late or driving in dicier conditions, or following too closely.

 

[EVNow]

IDEALLY what they should grade would be the following FSD supervisory skills:

% of city driving done in the last one year on AP.

 

[dduffey]

55 mph zone, as a safe driver what would you do? You have 1/10th of a second to react.

 

[dckiwi]

If they’re really going to be releasing it to 1k drivers a day, then they’re going to need an automated method of some sort to kick people off for being unsafe, and probably going to have to err on the side of caution to try to do so before a major accident occurs. Having the safety score apply to all driving, human and Beta, might do it. Inform new Beta users that it’s their responsibility to monitor the system and make sure to intervene anytime it’s about to undertake an action that would ding it under the safety score metrics because the scoring system is being used to evaluate Beta use and will be combined with the user’s normal driving score. and if that combined score drops below a certain level then the user is kicked off of Beta. Much howling would ensue, but safety should be prioritized as much as possible.

I think this is a good idea. For example, if a beta user waits too long to intervene in a dangerous maneuver, they will likely swerve or slam on the brakes. This would ding the safety score.

 

[tivoboy]

Has anyone seen their score go UP at all over the past 24 hours? After getting dinged last Saturday with a jack.hole pulling out in front of me and incurring a FCW dropping my score to 54, I drove each day between about 5 and 20 miles for trips, and everyday my score went up. First day, I got it up to 70, Mon, 74, Tues, 78, Weds, 84, Thurs, 88. This morning, it was 92 (I actually think it was 92 last night) and today I did some more drives, but the score has not changed at all. I figured it would go up between maybe 2 and 4/5 more points today.. but alas it hasn’t moved at all - which is odd Since pretty much EVERY drive I have done in the past six days has moved it up 1-2 point PER drive.

 

[FalconFour]

View attachment 71681055 mph zone, as a safe driver what would you do? You have 1/10th of a second to react.

let off the accelerator (remember: you don't have Chill on, so that's full safe braking instantly), lay ON the horn, dip to the side a little. dumbass will react. Looks pretty certain the guy was trying to dip a right turn, not a left.

FCWs are avoided by reacting - the car is looking for you to be responding to a potential collision. If you do, and it's happy with the response, no FCW. If you don't react and it thinks you need a reminder to elicit a pedal-input response, it'll bleeet an FCW at you.

 

[dduffey]

let off the accelerator (remember: you don't have Chill on, so that's full safe braking instantly), lay ON the horn, dip to the side a little. dumbass will react. Looks pretty certain the guy was trying to dip a right turn, not a left.

FCWs are avoided by reacting - the car is looking for you to be responding to a potential collision. If you do, and it's happy with the response, no FCW. If you don't react and it thinks you need a reminder to elicit a pedal-input response, it'll bleeet an FCW at you.

Wrong answer (you can tell by the body roll of the truck, that was at a stop). Truck entered at a high rate of speed to cut into traffic and entered your lane. Regen would cause a collision. I swerved, with an aggressive turn and avoided a FCW.

 

[dduffey]

Here is another one, merge lane and SUV crosses solid line.

 

[nvx1977]

View attachment 71681055 mph zone, as a safe driver what would you do? You have 1/10th of a second to react.

Do the Tesla Wave to the approaching Model 3. duh.

 

[dduffey]

Do the Tesla Wave to the approaching Model 3. duh.

Correct

 

[Sank03]

Anyone who requested FSD, got the fsd download?

 

[Daniel in SD]

View attachment 71681055 mph zone, as a safe driver what would you do? You have 1/10th of a second to react.

From the picture it looks like the truck is turned to stay in right lane. It really crossed into your lane?

 

[Jeff N]

In the last week I've been hyper-sensitive to cars in neighboring lanes and conga lines in adjacent lanes. Previously I was aware of them from a "be ready for evasive manuevers" but with the scoring it's pushed more toward "slow down even more when there's congestion".

In my experience, traffic tends to clump together so if you just drive 1-2 mph slower (or faster, I suppose) then you can often drive within the gap between traffic clumps where there are few cars around you.

Of course, this doesn’t work in congested traffic which many of us have to drive in during work commutes.

 

[Daniel in SD]

Here is another one, merge lane and SUV crosses solid line.

Gently turn right so TACC will see the car and brake. haha.

 

[jebinc]

Where was this reported? Makes me feel a little better about my 94.

Teslatari, I think.