Parked Car Losing 20 Miles / Day of Range

[whatthe2]

Please re-read what I wrote about availability heuristic bias.

Oh, I read it. Guess you missed the winky face at the end of my post. Or maybe it just wasn't funny.

 

[datoda]

I'm not convinced of that. Let's do some math. According to the EPA, a Model 3 LR gets 259 Wh/mi, so at 1mph of range loss, that's a consumption of 259 watts. Let's say that you leave Sentry Mode active for 8 hours per day, 365 days a year. That works out to about 756 kWh per year. IIRC, the national average cost of electricity is $0.13/kWh, so that's about $98/year. (I'll round up to $100/year, since that's a nice even number.) That value will obviously go up if you leave Sentry Mode active more, or down if you use it less. (I picked 8 hours a day based on the assumption that you're leaving it active when it's parked in a parking lot at work on a daily basis, and for similar periods when running errands and whatnot on weekends.)

Personally, in 30 years of car ownership, I've had no break-ins and only two incidents that even might have been vandalism or parking-lot accidents. One time my rear window was shattered when I lived in Boulder, Colorado, about 20 years ago. That might have been vandalism (nothing was stolen), or it might have been blown out in a wind storm, which are common in Boulder. Another time, much more recently, I noticed a deep ding that pierced the paint on my car. That might have been deliberate vandalism or a parking-lot accident. I don't recall what it cost to replace the rear window, but the ding cost about $400 to have fixed, which was under my deductible. Let's assume a similar price for the window repair, which means my repairs for these two incidents totaled about $800. If I'd had a feature like Tesla's Sentry Mode for that entire 30-year period, with similar costs, it would have cost me $3,000. The best-case scenario is that Sentry Mode would have saved me $800, but cost $3,000 to operate. The repairs would have been no less hassle, and suing or filing insurance claims against the perpetrators would have been more of a hassle. Overall, this doesn't look like a winning proposition from a selfish economic perspective and based on my personal experience with parking lot accidents or vandalism.

That said, there are caveats and exceptions, such as:

• My experience may not be representative. I don't happen to have statistics handy on how common break-ins, parking-lot accidents, and vandalism are.
  
• Teslas are more expensive cars, and are more expensive to repair, than the cars I've owned in the past. (The blown-out window was on a 1992 Saturn SL, and the door-ding was on a 2016 Chevy Volt.) This will shift the balance in favor of Sentry Mode.
• More judicious use of Sentry Mode may well be worthwhile -- if you're parking the car in a particularly risky area for an hour a week, it may be worth using, for instance; the probabilities shift in this sort of scenario.
• The cost of electricity varies greatly from one area to another. If your electricity is cheap, it'll make Sentry Mode more worthwhile. Of course, it becomes less worthwhile if electricity is expensive where you live.
  
• If/when thieves, vandals, and careless drivers learn that Teslas (and, eventually, other cars) may be recording their mistakes and misdeeds, that may serve as a deterrent. OTOH, this effect has Orwellian overtones that may contribute to an eventual backlash, so I'm not convinced it's 100% positive.

When we see videos of Sentry Mode catching an accident or vandalism, and the owner of the Tesla therefore having an easier time getting redress for the damage, we suffer from various cognitive biases, such as the availability heuristic -- we see, and remember, the one incident of Sentry Mode providing a benefit to the owner. We don't see the thousands of recordings that show absolutely nothing, much less the money spent on keeping those cameras running to record absolutely nothing of interest. If we had to wade through all those recordings, too, we'd be less likely to say that Sentry Mode is such a wonderful feature.

Of course, this analysis could also change if Tesla could bring down the energy costs of running Sentry Mode. Based on my limited understanding, I doubt if this will happen for current cars; but if Tesla (or somebody else) wanted to, a security system with less electricity overhead could almost certainly be built.

Following your analysis, I guess you don't buy auto insurance if your claims never exceed the premiums. To me, Sentry mode, just like insurance, is for capturing major incidents. Minor things like catching door dings and such are just icing on the cake.

 

[Kant.Ing]

Consuming 1mph is like 240W. Three dashcams or surveillance cameras are usually no more than 10W.

If Sentry Mode is going to stay, Tesla will need to dedicate some cheap and small processor(s) to do the job more efficiently, and for the good of engineering.

 

[srs5694]

Following your analysis, I guess you don't buy auto insurance if your claims never exceed the premiums. To me, Sentry mode, just like insurance, is for capturing major incidents. Minor things like catching door dings and such are just icing on the cake.

Insurance already protects against vandalism, theft, and parking lot accidents. Viewed in this way, TeslaCam is, at best, a supplement to insurance, which may, in some cases (if it catches the perpetrator with enough identifying details to track that person down), cover the deductible. You could get the same effect, with more certainty of coverage, by reducing the deductible on your insurance. I haven't checked the pricing on this, so it may be prohibitive, but it would have the advantage of not releasing any additional CO2 into the atmosphere, which Sentry Mode does.

 

[GolanB]

I'm not convinced of that. Let's do some math. According to the EPA, a Model 3 LR gets 259 Wh/mi, so at 1mph of range loss, that's a consumption of 259 watts. Let's say that you leave Sentry Mode active for 8 hours per day, 365 days a year. That works out to about 756 kWh per year. IIRC, the national average cost of electricity is $0.13/kWh, so that's about $98/year. (I'll round up to $100/year, since that's a nice even number.) That value will obviously go up if you leave Sentry Mode active more, or down if you use it less. (I picked 8 hours a day based on the assumption that you're leaving it active when it's parked in a parking lot at work on a daily basis, and for similar periods when running errands and whatnot on weekends.)

Consumption may actually be lower than this if you take into account that the car is already losing power due to phantom drain. We are assuming here all power is lost due to sentry mode which isn't an entirely fair characterization.

Some things that Tesla might consider doing to improve Sentry Mode in future releases:

- If we're going to be leveraging the computer systems while the cameras are running, why not use computer vision to identify events in the video stream. When scrubbing through hours of video, this could be helpful. It could also give us better notifications to our phones in real-time, and it could also be used to filter out recognized faces (already a standard feature with Google's Nest Cams).

- Consider optimizing for a low-power mode. Even if Sentry mode only consumes 150 watts, that still seems very high for motion detection, minimum processing, and writes to storage.

- Better storage management, including cloud-backups.

- Option for Silent Mode operation - in some cases this may be preferred

- If and when Tesla adds cloud capture, we could see the emergence of features such as Neighborhood sharing mode where users could opt-in to share events to other users by region. We've seen similar features on Ring devices. Today, Law Enforcement in some areas already have access to these (opt-in) streams, and it might make sense for mall security and other high traffic areas to have similar access where there may already be a security team in place. The power of video capture really comes together when there are multiple feeds at different locations and angles being shared.

With 500,000 independent recording systems already on the roads, this can be quite an asset to cities to help curb crime should we choose to use it for this. With better power management, and continued drop in network transmission costs it should be feasible to have Sentry Mode running 24x7.

I would, as an example, be willing to pay in the neighborhood of 5$/month if the power consumption costs could be brought down. If this could also bring down insurance costs further, it might be a wash.

 

[srs5694]

Consumption may actually be lower than this if you take into account that the car is already losing power due to phantom drain. We are assuming here all power is lost due to sentry mode which isn't an entirely fair characterization.

True, but by all accounts (as in the title of this thread), Sentry Mode consumes ~1 mile/hour of range, whereas normal Tesla phantom drain accounts for ~2-4 miles per day. That's close to an order of magnitude difference, so the cost comparison I presented won't shift much. Maybe it's $80 or $90 per year rather than $100, and therefore $2400 to $2700 for the 30-year period I used for an example, rather than $3000. That's still much greater than the $800 or so that I paid in that period of my life because of unexplained damages to my cars.

I would, as an example, be willing to pay in the neighborhood of 5$/month if the power consumption costs could be brought down. If this could also bring down insurance costs further, it might be a wash.

$5/month works out to $60/year, which is over half of what Sentry Mode costs to operate, according to this analysis. Unless that payment brought huge power consumption reductions, or maybe if you ran Sentry Mode 24/7 or close to it, it wouldn't be worthwhile.