this post was submitted on 27 Jul 2023
538 points (95.4% liked)
Technology
59370 readers
932 users here now
This is a most excellent place for technology news and articles.
Our Rules
- Follow the lemmy.world rules.
- Only tech related content.
- Be excellent to each another!
- Mod approved content bots can post up to 10 articles per day.
- Threads asking for personal tech support may be deleted.
- Politics threads may be removed.
- No memes allowed as posts, OK to post as comments.
- Only approved bots from the list below, to ask if your bot can be added please contact us.
- Check for duplicates before posting, duplicates may be removed
Approved Bots
founded 1 year ago
MODERATORS
you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
If you know the true answer, but you give your customer a false answer to make your product look better than it is, there's a word for that. It's "fraud".
Which reminds me of Apple and MS and Google and who not too...
Counterpoint: Ive taken numerous road trips in both of our family's Tesla (Tesli?) as well as a couple loaners, and the built in navigation is always spot on with the estimates. Like it's eerie how it can predict within a percentage point on a 2 hour or more drive within the first 10 minutes of a trip.
Range anxiety really is only experienced by those that it doesn't affect (i.e. potential buyers)
It sounds like your talking about you put an address in gps and it gives you an accurate number.
The article is talking about it's version of a gas gauge, where it says X miles remaining, and that is what's inflated.
Trying to lie on the gps would cause more complaints as people got stranded, the fraud was lying on the "gas gauge" where it would be hard for a customer to realize they had less juice than they were being told.
But it's addressing the same thing, no? The number it displays is the epa range and any state of charge. I prefer to just show a percentage but either way it's understood to be an estimate. If you want a true value just enter a destination (you can do a multi leg trip as well)
Also this article is so vague it's almost useless. I highly doubt this team was just straight up closing service tickets; so more than likely they trained a single team on the talking points of the display number vs real world and thus improved efficiency with service tickets. The article even admits the cars didn't need any actual service
I said it in another reply but it's not unlike a phone telling you it has 12 hours remaining, but then you play a graphically intense game and it dies in 2. The margins are much smaller here but the point is still valid
They made the numbers less accurate because people complained real distance per charge isn't what's advertised.
I have no idea why so many keep bending over backwards to make fraud seem normal.
But if you've read this whole thread and still don't get it, I don't think I'm going to keep trying.
Thanks for taking the time to tell everyone you don't know what you're talking about I guess.
And everyone can see your post history and instantly disregard any opinion you have based on your history.
Or just ignore you, which now that I've typed this is probably the smart move here
Edit:
Why it's pointless
https://lemmy.world/comment/473319
The number it gives is based on ideal driving. If it says there's 200 miles left, no one should be surprised they don't get 200 miles when they drive 85 on the highway
Except the article is saying that they purposely inflated the number it gives.
My (diesel) Equinox gives a conservative range based on actual driving conditions. It is slow to increase a range estimate when I get on the freeway; quick to decrease when I get back into the city, and the actual range available is always 20-50 miles more than the gauge indicates. It is consistently and reliably under-promising and over-performing. If it tells me I can just barely make it to my next stop, I know I can make it, with fuel to spare.
I'd be pretty anxious about range if my car consistently overestimated its own capabilities. When I've been in stop-and-go traffic for the last 30 miles, it should not assume that I'll spend the rest of the charge cruising on the freeway at 5 under the speed limit. If a manufacturer were to use such an algorithm to estimate range, I would say that manufacturer is perpetrating a fraud.
No, it should be, but it's not. I'm not going to keep explaining it tho, you should just read the article you're commenting on.
Then you can email the author and explain how they're wrong and Elon is amazing. I'm sure they'd love to hear from you.
I concur, this is also my experience. The car GPS has never directed us to travel further than the charge allows--and it will include stops at superchargers on the way as necessary. It's really not that big an issue.
But, the range that it presents you in the UI is not the actual range that you can travel. The fact that the car won't plan out a route for a location 300 miles away when it claims you can travel 320, but will instead include a stop at a supercharger at around 200, kinda proves they know this.
I think the projected range is basically the platonic ideal if you were traveling in a perfectly flat landscape, with no wind, with an external temperature of 18.2°C, traveling at 37.25 miles per hour or whatever. Every deviation from that ideal will hurt your range. In my experience, I tend to get probably 250-ish miles on a 320 mile charge, depending on the time of year.
Gas vehicles tend, on the other hand, to undersell the range in my experience, and people are used to going further than the car says they can.
The problem is that other vehicles adjust the projection based on current conditions - when I drive up a mountain, my projected range drops like a rock. When I drive back down I can end up with more range than I started. Reporting the “ideal” case during operation is misleading at best.
Yeah, it seems like using 'miles' as an indicator or energy left is the root cause. If they just change the kwh left or similar they'd be more accurate but, ironically, confuse way more people.
Though, ironically a scale of Full - 3/4 - half - 1/4 - empty is perfectly fine for gas. There is usually a visual gauge of % for charge, but it isn't as prominent as the range. Oddly, my car has it divided roughly in thirds.
It's also less accurate. Ever notice your phone sometimes drops from 100% to 80% in only a few minutes, or hangs around at 10% for ages? That's because with batteries it's much less simple than "full, medium, empty". There's actually a bunch of code to improve the estimation specifically for your battery, and still they can behave strangely.
How would anyone actually use that information?
Same as a fuel guage I'd imagine.
I've never seen a fuel gauge marked in any kind of units like liters or gallons, just fractions of a full tank.
Tanks is a unit of volume....
How many liters are there in a tank?
Implementation defined.
But like kWh you can easily learn (either intuitively or by the dark magic of multiplication) the ratio between tanks and ability to get to your destination.
The idea that people can't learn that they get roughly 6 or roughly 3 etc km per kWh with their driving style and thus physical units should be hidden from them is insulting and anti-intellectual. Stop trying to project your stupidity and anti-curiosity onto others.
Says the idiot who doesn't even know what a unit is.
An international standard unit is not the same thing as a unit. A cubit is a unit even though people's arms aren't all the same size.
I said stop trying to project your stupidity onto others, not double down on it.
Bye, asshole.
That's part of my point. kWh isn't very useful to most people. The problem is that 'miles left' is an abstraction from kWh which is more helpful but less accurate.
Now people are complaining that it's not accurate, which it was never going to be in the first place. It's a UX problem. They should probably just change to a percentage based readout with a "Estimated Miles Remaining" option for those that want it.
I found some images of a Tesla's display, and it has a percentage and a bar graph just like a phone. The problem isn't that people can't see roughly how much charge is left, it's that the distance-remaining display is misleading to such a degree that it seems malicious, and it's demonstrably possible to give a much more accurate estimate. They are at the very least guilty of including a defective feature in their cars.
You can display percent instead of rated miles
That's a 28% exaggeration in economy.
You're putting a certain amount of electricity into that car, and you are being told that amount of electricity is good for 320 miles. You're only getting 250 miles from that. Everyone else is only getting 250 miles from that. The only people supposedly getting 320 miles of range from that charge are the salespeople convincing you to buy it. Misrepresentation of the distance you can travel on a battery is no different than misrepresentation of the distance you can travel on a tank. You're just multiplying the MPG by the size of the tank, or the kWh/mile by the battery capacity.
Suppose I buy a sedan with a rating of 32mpg. But I'm only getting 25mpg. I put 10 gallons in the 10-gallon tank, expecting to drive 320 miles, but I have to stop and put another 10 gallons in after 250 miles.
Everyone else who bought the same sedan is getting 25mpg. Meanwhile, my other car, an SUV from a different manufacturer, is also rated 32mpg, and actually gets 32mpg. Everyone else who bought that SUV is reporting pretty damn close to 32mpg.
Clearly, the manufacturer of this sedan is pulling something shady. It doesn't become less shady when the sedan is burning gallons of diesel instead of gallons of gasoline. Or gallons of propane instead of gas or diesel. Or cubic feet of natural gas or hydrogen instead of a liquid fuel. Or pounds of steam instead of a combustible fuel. It doesn't become less shady when the sedan is consuming kWh of electricity instead of a mass or volume of a physical substance.
Counter counter point: if the Tesla is doing fraud with the range estimate there is no need to estimate anything that precisely. Just make the software show the same number as guessed when you arrive let's say you end up with 86 km left as "estimated" at the end of the trip but in reality it's more like 42 km and the Tesla just shows something else.
Except that is really falsifiable. I pull data with TeslaFi and there's definitely no shenanigans being done.
The simplest explanation is that it's impossible to say (miles left) when you consume significantly more power going faster, going uphill, on a very hot day, etc. So they just go with the epa estimate based on your % state of charge and that's it.
If you want detailed info the car will quickly give it to you and consider all those factors if you put in a destination
The gauge shows epa range available given battery SoC. Once a destination is entered it gives you an accurate estimate in the gps directions.
It can’t guess anything until it knows where you’re going.
Not sure why you're being down voted. This has been my experience as well.
The remaining battery estimate given at the beginning of the trip is fairly accurate.
How do you know it's accurate if you don't run it to empty?
I've had many legs of a trip where we get there under 10%
I also pull data from via a third party app and the historical data confirms the numbers aren't just made up
Everyone wants to hate on it but it really is a pleasant experience and the only complaint is it doesn't give you an accurate estimate of miles on the main screen when that is literally impossible without a destination in mind.
Think about it your phone would tell you it had 12 hours of battery left then you played an intense game and it dies in 2 hours... It's a very similar issue
I believe it would be Teslae.
I believe it’s actually Teslapodae since it’s from the Greek
Yes, if the projected range is more optimistic than reality, it’s always because I drive faster than 120-130 kph. Otherwise it’s absolutely spot on or even better than projected, for example if I drive 100-110 kph for a while.
In Europe the manufacturers are legally bound to quote the WLTP range. Which is hopelessly inaccurate.. But nowhere near as bad as the NEDC that preceded it. Of course people still come on forums wondering why they don't get <50% more than actually possible> out of their car, and I don't blame them.. the law is an ass.
TBF to Tesla though the in-car estimate is (I think) EPA and isn't far off.. It's doable in summer, at least. Winter you'll lose 30% but that's normal for all cars.
Many manufacturers get around this by quoting an estimated range and then the WLTP in small print ie. Highway Range: Est 415km in normal conditions (600km on WLTP cycle)
Hasn’t it been proven that the range can go longer but Tesla caps it? I remember during a hurricane Tesla said it was extending the range so drivers can make it out.
That’s not he projected range shown on the display. That’s the actual range the battery will give you.
They temporarily removed a bit of the safety margin built in to reduce battery wear, allowing people to get out of harms way without a stop to charge if they were on the limit. But only for people moving away from The affected area, etc. so smart and helpful.
Ah okay, thank you for the additional info.
I dislike both Tesla and Musk but as an automotive company I'd wager they have good safety factors on at least most of their engineering, so for example in the name of keeping the battery from degrading too fast the tesla probably stops drawing power from the battery when the battery voltage drops to a certian number. The safety factor just means that the number it shuts off at is likely a good bit before the batrery would theoretically start to take damage.
I can't find any other reason they would intentionally keep range off the product, because its not like they selling a solution to it. So I assume they probably just lowered the voltage at which the Tesla stops drawing power because getting stuck in the hurricane was obviously more dangerous. I could be wrong though, this is just from my expirences with bateries in engineering applications.
I think that would be "Teslae" 😉