Tesla rigged the dashboard readouts in its electric cars to provide “rosy” [range] projections

Tesla’s secret team to suppress thousands of driving range complaints

About a decade ago, Tesla rigged the dashboard readouts in its electric cars to provide “rosy” projections of how far owners can drive before needing to recharge, a source told Reuters.

www.reuters.com

Interesting.

TL;DR

1. Above 50% SOC the cars over state range, and under 50% they become more realistic.

2. Tesla had or has a team in Nevada and now moved to Utah solely dedicated to cancelling Service requests for range complaints, as they are deluged with them.

3. EPA ratings: Tesla aggressively uses all options to get the most rosy estimates (we knew that)

4. No comment from Tesla or any mfg on range for the story, some old comments.

So this feels about the same as almost every ICE car I've owned. The fuel gauge drops slower for the first 2/3 of a tank, then drops quicker at the end.

"Supercharging-Batterieanzeige auf dem Display, während der Elektroauto-Ladung an einer Tesla Supercharger Ladestation" by verchmarco is licensed under CC BY 2.0.
Admin note: Image added for Blog Feed thumbnail

 

[strider]

IMO it seems to be getting worse with newer models. In our 2012 and 2018 MS we could regularly get rated range. I cannot get rated range in my 2022 MXP and I drive the same (I have only done 3 full power accelerations in the 15 months I've owned it). I think it's because Tesla actually shrank the battery in the Refresh S/X (no longer 100kWh) and claimed higher efficiency to get the range numbers but I am always woefully below where I should be.

Car and Driver does a real 75mph range test and the results are what you would expect. Tesla radically overstates range (MSLR only went 320 with 402 rated, M3LR only went 240 with 310 rated) while the Germans actually understate range (Taycan went 280 with 225 rated) with other manufacturers closer to their rated. Chart here:

EV Range: Everything You Need to Know

We explain EPA ratings, factors that affect range, how EVs have performed in our testing, and why it's all very complicated.

www.caranddriver.com

 

[MY1stEV]

I believe the EPA range lies estimates are based on highly ideal conditions like driving 45mph on perfectly flat terrain at 65° and no wind.

Maybe someone would like to take one for the team and fully charge in D'Iberville, MS and see if you can make it all the way to Beaumont, TX (330 miles) driving 45mph on I-12/I-10.

 

[Olle]

EPA range is well defined and part of the cycle is very slow and wouldn’t correlate much with highway range overall. As the article says, EPA occasionally do their own to compare with the self reporting and found a 3% discrepancy for some Tesla Model. Not too shabby. As the low speed efficiency increases every year, so does the EPA range.

It‘s too bad that Tesla removed the “how far can I go“ page where you could see range at different speeds And temperatures. They definitely subscribe to the ‘Less is More‘ information model.

 

[wwu123]

Tesla’s secret team to suppress thousands of driving range complaints

About a decade ago, Tesla rigged the dashboard readouts in its electric cars to provide “rosy” projections of how far owners can drive before needing to recharge, a source told Reuters.

www.reuters.com

Interesting.

TL;DR

1. Above 50% SOC the cars over state range, and under 50% they become more realistic.

#1 is about the hidden buffer thing, as explained in the article, and is not well-understood even around these forums. Here is one forum thread that discusses it:
How Tesla Calculates Range (hidden buffer)

It's not that under 50%, it becomes more realistic as in using your driving habits, or factoring in air conditioning and other drain ( it is still using the EPA estimate wh/mi), but rather that it starts removing the 4-5% buffer capacity from the displayed rated range. In order to ever use the buffer range, you have to drive below the displayed 0% - but there's a huge risk in doing so, as you're basically doing it blind without knowing how close you are to the true 0%.

I recall there was a new owner around here several years ago, that was on a drive home and stopped with about 10-15% showing, when they got back into the car after eating, and suddenly the car was showing only a few % left, causing a bit of a range anxiety panic. Most folks chalked it up to a new owner lack of understanding of range estimates, or conflated the issue with other well-beaten range estimation or rated range complaints. He persisted and eventually was one of the first to document in detail the hidden buffer issue, but by that point the thread was already being shouted down, it was already massively confusing to read through. I cannot find that old thread currently, but it's in these forums somewhere.

But certainly the hidden buffer thing will also be lost and misunderstood in this thread as well, as it already is, amongst the other more common range issues.

 

[OxBrew]

The only thing really new to me in the article is overstating the range above 50% SOC, or what you called 5% buffer.

Really weird, I've seen ours drop suddenly right around 50%; a 2 mile round trip I do all the time takes 1%, except if it straddles the 50%, then it takes 5-7%. I wondered what the hell was going on.

 

[MP3Mike]

The only thing really new to me in the article is overstating the range above 50% SOC

And their claim for that was about the Model S in 2012, not in anything current. Back then Tesla had an "Ideal" range you could selected, instead of the "Rated" range, that reported more range available. (But likely only achievable in low speed/hypermiling situations.)

 

[ZenRockGarden]

The EPA range test is completely defined by the EPA and is unfortunately highly unrealistic for highway conditions since it uses an average speed of only 48 mph. Teslas happen to be crazy efficient at that pace - you'll easily pass the EPA range if you drive under 60 MPH on the highway.

Real humans unfortunately often travel at 80 MPH on modern highways. Teslas are much less efficient at that speed. The Germans care more about high-speed performance on their autobahns, so for example the Taycan uses a 2 speed transmission to squeeze a bit more efficiency out of the top end. Tesla went with a no-transmission-fewer-parts compromise.

The biggest single change I think should happen is to get the EPA into the business of making realistic tests that match the typical US American driving experience instead of hiding in their lab with a dyno and a history book.

My personal experience is I can easily match the EPA range on my MYLR around town, I can also match it if I set the cruise-control on 60 MPH on the highway, and every 10 MPH faster eats about 30 miles of range so yeah if I run at 80 MPH I'm in the mid-200's

 

[beatle]

The guess-o-meter range estimate in any vehicle is pretty much useless as the car has no idea where you're going or how fast you plan to get there. Weather, elevation, and speed all play a part. The navigation estimate works pretty well most of the time though which is what really counts. I changed the display to a percentage long ago.

Agreed that the EPA tests are a poor measure of how people will actually use the cars.

 

[GData]

If EPA tests are unrealistic, then all manufacturers range estimates should be equally off, but that's not the case. German car estimates are accurate.

So, why are Tesla's range estimates so bad compared to other makes?

 

[MP3Mike]

If EPA tests are unrealistic, then all manufacturers range estimates should be equally off, but that's not the case. German car estimates are accurate.

No, they aren't. They are often very inaccurate in the opposite direction.

There are two different EPA test procedures, most manufacturers opt for the easy shorter tests.

And it seems for most Tesla vehicles the EPA ran the tests themselves, so it isn't even Tesla that is creating them. (In one case on the Model S, Tesla had to demand the EPA retest it, because they screwed up the testing procedure, resulting in the rating being lower than it should have been.)

 

[MP3Mike]

Edmonds, who doesn't tend to be pro-Tesla, tested actual range, when driven similarly to the EPA test procedures, against the rated EPA range: Edmunds Tested: Electric Car Range and Consumption | Edmunds

The data for the top, and bottom, 10 stolen from @exxxviii's post in another thread:

Rank	Vehicle	EPA Range	Edmunds Range	Range Precision
1	2021 Tesla Model S Plaid	348 miles	345 miles	0.9%
2	2022 Rivian R1T Launch Edition	314 miles	317 miles	1.0%
3	2021 Tesla Model 3 Long Range	353 miles	345 miles	2.3%
4	2020 Tesla Model S Performance	326 miles	318 miles	2.5%
5	2021 Tesla Model Y Long Range	326 miles	317 miles	2.8%
6	2022 Lucid Air Dream Range	520 miles	505 miles	2.9%
7	2022 Kia EV6 GT-Line dual motor	274 miles	283 miles	3.3%
8	2022 Kia EV6 Wind RWD	310 miles	323 miles	4.2%
9	2021 Ford Mustang Mach-E GT Performance	260 miles	272 miles	4.6%
10	2022 Porsche Taycan GTS	246 miles	259 miles	5.3%
22	2021 Volkswagen ID.4 Pro	260 miles	288 miles	10.8%
29	2021 Volkswagen ID.4 Pro S dual motor	240 miles	272 miles	13.3%
35	2021 Volkswagen ID.4 First Edition	250 miles	287 miles	14.8%
44	2020 Hyundai Ioniq Electric	170 miles	202 miles	18.8%
45	2020 Kia Niro EV	239 miles	285 miles	19.2%
46	2022 Hyundai Kona Electric	258 miles	308 miles	19.4%
47	2022 BMW iX xDrive50 (22-in wheels)	315 miles	377 miles	19.7%
48	2022 Mercedes-Benz AMG EQS 53 4Matic	277 miles	332 miles	19.9%
49	2022 Mercedes-Benz EQS 450+	350 miles	422 miles	20.6%
50	2022 Audi RS e-tron GT	232 miles	285 miles	22.8%
51	2022 Porsche Taycan (20-in wheels)	225 miles	286 miles	27.1%
52	2020 MINI Cooper SE	110 miles	150 miles	36.4%
53	2020 Porsche Taycan 4S (20-in wheels)	203 miles	323 miles	59.1%

Notice that Tesla tends to be at the top, being the most accurate, while the least accurate ones tend to be from the German manufacturers. The Porsche Taycan is off by as much as 59%. (Sure, the Taycan went further than they rated, but that isn't an indication of accuracy.)

 

[brkaus]

I believe the EPA specifically allows the range/mpg to be reduced by the manufacturers, if requested.

 

[OxBrew]

And their claim for that was about the Model S in 2012, not in anything current. Back then Tesla had an "Ideal" range you could selected, instead of the "Rated" range, that reported more range available. (But likely only achievable in low speed/hypermiling situations.)

They were unclear about that. You could read it as starting in 2012, but still continuing now. IDK.

 

[ZenRockGarden]

If EPA tests are unrealistic, then all manufacturers range estimates should be equally off, but that's not the case. German car estimates are accurate.

So, why are Tesla's range estimates so bad compared to other makes?

Cars are designed differently and have different optimal operating conditions. As I explained above, an easy example is that the Germans tended to add a transmission to optimize the top end efficiency (at the cost of complexity, weight, and cost). Doing so did not improve their EPA number at 48 MPH, but does help if you plan to run their car at a steady 80 MPH (well above the EPA test params).

If you drive a Tesla at an average of 48 mph you'll get huge range

 

[Loucks Lion]

On my 2020 M3 LR I simply use .8 of Tesla range for actual range. ie 300 x .8 = 240. For me this is OK since 3 hours of driving on Interstate is enough because I need a break. Made several long distance drives and after a few trips rang eanxcity has disappeared. City driving never an issue.

 

[pp61]

Audi shows good estimates. It adjusts the range directly when you start the navigation system. Example, you have 200 miles left when entering the car. You set destination which is up the mountain. It immediately adjust and show 170 miles left. No surprises here from the start.

 

[brkaus]

Audi shows good estimates. It adjusts the range directly when you start the navigation system. Example, you have 200 miles left when entering the car. You set destination which is up the mountain. It immediately adjust and show 170 miles left. No surprises here from the start.

I find the Tesla trip planner to be pretty accurate too. Except when I drive 85.

 

[Dr. J]

Car and Driver does a real 75mph range test and the results are what you would expect. Tesla radically overstates range (MSLR only went 320 with 402 rated, M3LR only went 240 with 310 rated) while the Germans actually understate range (Taycan went 280 with 225 rated) with other manufacturers closer to their rated. Chart here:

EV Range: Everything You Need to Know

We explain EPA ratings, factors that affect range, how EVs have performed in our testing, and why it's all very complicated.

www.caranddriver.com

That is a really useful chart...until you get to this part:

We don't (yet) control the weather, so the worst-performing examples, including a 2018 Model 3 that only managed 65 percent of its range rating, took place with outside temperatures hovering around freezing.

So the 2018 Model 3 Performance went 10% farther than a 2018 Model 3 LR. Reality is probably the opposite. They have the right idea, but this chart is not that useful for comparison purposes. Providing the test ambient temperature would help.

 

[Darmie]

Audi shows good estimates. It adjusts the range directly when you start the navigation system. Example, you have 200 miles left when entering the car. You set destination which is up the mountain. It immediately adjust and show 170 miles left. No surprises here from the start.

Interesting take. Kind of like it...
How does the miles left react when you have entered waypoints in the nav?