False FSD Strikes?

[AlanSubie4Life]

What software version?

@AlanSubie4Life won’t believe your story unless you ahve a video of it.

I believe it if you keep tugging the same way. Sometimes that does not seem to work. At least on some software versions it did not. Of course, I have no idea exactly how the system is set up, and I could be wrong.

Just the more video evidence there is, the easier it is to figure out a coherent explanation for observed behaviors.

 

[AlanSubie4Life]

I believe it if you keep tugging the same way. Sometimes that does not seem to work

Just want to confirm that this is still the case. Have to apply opposite direction torque in some cases. Same direction torque won’t work.

As described that may well have happened here.

Fortunately it remains quite difficult to get a strike, assuming you are driving the vehicle.

 

[KArnold]

Just want to confirm that this is still the case. Have to apply opposite direction torque in some cases. Same direction torque won’t work.

I'm not so sure. Of course, IMHO.

I'm not a fan of putting torque on the wheel for confirming my attention (no camera in my car) via the torque method. For me it feels much safer to loosely hold my right hand just over the wheel, but letting the car control the turning sliding past my hand. Obviously a needed intervention would be immediate. My left hand is comfortable covering the FSD lever. It's easy to quickly identify the initial "pay attention" toaster message using muscle memory. Of course. If you are distracted the alarms escalate. I believe in more critical traffic situations, the nags escalate too - a nag at every traffic light and yield condition. That's exactly when I want to avoid putting torque on the wheel.

I find this to be a good balance for me - but I suspect most would hate it.

##EDIT## LOL - in rereading my post, I didn't really respond to your post. The torque question may be true. Just wanted to point out another option.

 

[AlanSubie4Life]

I'm not so sure. Of course, IMHO.

You wouldn't notice this with your method (which is also now my method though it has not always been). It's not an issue if you aren't applying torque. I think you can also release and reapply same direction and that will work (not sure).

I'm not a fan of putting torque on the wheel for confirming my attention (no camera in my car) via the torque method

I am not anymore either. It doesn't satisfy the nags to apply torque to the wheel, so I don't bother! What would be the point?

For me it feels much safer to loosely hold my right hand just over the wheel, but letting the car control the turning sliding past my hand

Yes, I just steer the car now, too. Loosely hold it, as though I were steering normally. This is rarely enough torque to satisfy the torque nag, so it works well when the torque nag appears, since you can just keep steering slightly more firmly and then the nag is satisfied. This all happens without the worry about the nag not being satisfied by torque because you were already applying torque!

I assume this is all a side effect of the nag defeat stuff but in the end it works out better since it has got me out of my constant torque habit, which was annoying to do anyway.

I believe in more critical traffic situations, the nags escalate too - a nag at every traffic light and yield condition.

Yes.

That's exactly when I want to avoid putting torque on the wheel.

Yes.

 

[sleepydoc]

Just want to confirm that this is still the case. Have to apply opposite direction torque in some cases. Same direction torque won’t work.

As described that may well have happened here.

Bug

Fortunately it remains quite difficult to get a strike, assuming you are driving the vehicle.

Actually, it's quite easy - have had it happen 3-4 times. and yes, I was driving the vehicle.

 

[YotaMD]

2018 MS. I've received two strikes on fsd. Both out of nowhere. No audio alerts, and no escalating visual queue that I noticed. I got the "telemetry indicated improper use" message.

I have about 10k miles on my car in the last 6 months. Almost all autopilot highway. I use it a TON.

I've never used a defeat device, but this seems to be what my strikes are for. Do these reset at any point? Is there any process to reconcile false alerts? This seems silly to lose this function over a false alert.

 

[KArnold]

What software? The latest version of FSDb is a 2 week suspension. Supposedly. I've not noticed anyone who has done it yet though. You may be first. Let us know if you do

No camera model? You only get 3 strikes. You may want to recalibrate your cameras too - won't hurt.

 

[zroger73]

3 month-old '23 MYLR with 4,400 miles and probably two-thirds of that on FSDb.

Currently on 2023.20.7 with FSDb 11.3.6.

FSDb does "okay" for me when driving straight ahead - the biggest issue is random, unnecessary lane changes. When it comes to stopping and turning, it is often somewhere between useless, frustrating, embarrassing, and dangerous, so lots of interventions are required.

I get nags every several seconds like everyone else and always satisfy them before the "beep stage".

I'd say I'm more cautious than the average driver with good driving skills.

Yesterday, I got my first strike. I have no idea why, when, or where it occurred - I didn't know about it until the next drive when I got the "telemetry data shows improper use" message.

Part of me is frustrated over this - it's like I'm paying to be falsely accused of a crime, but the other part realizes his insignificant this is.

 

[DrChaos]

3 month-old '23 MYLR with 4,400 miles and probably two-thirds of that on FSDb.

Currently on 2023.20.7 with FSDb 11.3.6.

FSDb does "okay" for me when driving straight ahead - the biggest issue is random, unnecessary lane changes. When it comes to stopping and turning, it is often somewhere between useless, frustrating, embarrassing, and dangerous, so lots of interventions are required.

I get nags every several seconds like everyone else and always satisfy them before the "beep stage".

I'd say I'm more cautious than the average driver with good driving skills.

Yesterday, I got my first strike. I have no idea why, when, or where it occurred - I didn't know about it until the next drive when I got the "telemetry data shows improper use" message.

Part of me is frustrated over this - it's like I'm paying to be falsely accused of a crime, but the other part realizes his insignificant this is.

I'm on 11.4.4 FSD, which some people report was a regression from the 11.3 series, but for me on SoCal freeways it is better. Supposedly it's the first version with more heavy neural network based planner instead of rules. For me, it's pretty good, but california is over-represented in the training set.

I would use "minimize lane changes this drive" more with 11.3.6, and still do today, but it's calmer.