Jump to content


Photo
- - - - -

"Real World" Fuel consumption.


  • Please log in to reply
52 replies to this topic

#1 gruntguru

gruntguru
  • Member

  • 7,628 posts
  • Joined: January 09

Posted 01 August 2023 - 21:01

All the way to the highest court in Australia. https://www.drive.co...sumption-claim/

 

Fuel consumption rating labels became mandatory in Australia in 2001 and were designed to provide consumers like-for-like fuel use comparisons across various vehicles. . . . .Fuel rating label figures are derived from laboratory tests (pictured below) so they can be repeated in identical conditions in different countries, rather than based on real-world driving scenarios. However, over time, car companies have developed techniques to optimise fuel consumption results during the laboratory tests. . . . . The difference between real-world fuel consumption and rating label figures was about 10 per cent in 2002, grew to 35 per cent in 2014, and in 2017 was projected to be 49 per cent higher than published claims by 2020, as manufacturers found ways to optimise laboratory tests.

“If this Supreme Court ruling is allowed to stand, then the Australian government will need to update its fuel economy test procedures,” said an automotive industry insider. “The car makers have done nothing wrong, other than conduct laboratory tests according to the current regulations.”
The two Mitsubishi Triton utes at the centre of Australia’s landmark legal dispute – the customer’s original 2008 model and the new 2016 model he bought as a replacement vehicle – displayed consumption rating label figures derived from laboratory tests, as with all new cars sold over the same period and to the present day.
However, Mr Begovic found, in his experience, his 2016 Mitsubishi Triton delivered a 24 per cent increase in real-world fuel consumption compared to his 2008 Mitsubishi Triton. Mr Begovic’s 2008 Mitsubishi Triton fuel rating label displayed an estimated consumption rate of 7.8 litres per 100km in highway use, while his 2016 Mitsubishi Triton fuel rating label displayed an estimated consumption rate of 6.8 litres per 100km in highway use, a 12.8 per cent decrease. However, Mr Begovic found, in his experience, his 2016 Mitsubishi Triton delivered a 24 per cent increase in real-world fuel consumption compared to his 2008 Mitsubishi Triton. Given the reduction in fuel consumption estimates on the 2008 versus 2016 fuel rating labels, Mr Begovic had expected his 2016 Mitsubishi Triton to use about 8.5 litres per 100km in his driving routine.Mr Begovic produced figures that showed his 2016 Mitsubishi Triton used an average of 12.44 litres per 100km in his driving routine, versus about 10 litres per 100km for the same driving routine in his 2008 Mitsubishi Triton. 
The Supreme Court heard Mr Begovic purchased the 2016 Mitsubishi Triton in part due to its reduction in the claimed fuel consumption versus his 2008 Mitsubishi Triton, and he had relied on the data published on the fuel rating labels when deciding to buy the vehicle.
Last week, the Supreme Court of Victoria upheld the earlier finding of the Victorian consumer tribunal that the figures on the 2016 Mitsubishi Triton’s fuel rating label – affixed to the vehicle at the point of sale – were “misleading and deceptive”.Further, the 2016 Mitsubishi Triton’s real-world fuel use of 12.44 litres per 100km compared to the 6.8L/100km claim for open road driving, the 9.0L/100km claim for urban driving, and the 7.6L/100km claim for a combination of urban and highway driving.

Edited by gruntguru, 01 August 2023 - 21:01.


Advertisement

#2 Greg Locock

Greg Locock
  • Member

  • 6,346 posts
  • Joined: March 03

Posted 01 August 2023 - 22:32

Aye, governments aren't very good at defining procedures that can't be gamed. I must admit 12.44 vs 7.6 is a worry. The increase between models is not like for like, the 2016 might be heavier, have DPF, off road tires, etc (I haven't checked). By the way they aren't lab tests as you might imagine (ie chassis dyno, like the old emissions tests), these days they can be calculations based on the engine dyno and parameters derived from simulations. That's why my fully loaded 4WD Everest has the same sticker (to within 1.4%) as a base model 2wd with the same engine. 



#3 404KF2

404KF2
  • Member

  • 18,742 posts
  • Joined: October 99

Posted 02 August 2023 - 04:25

Interesting.

 

When the smart fortwo cdi was introduced in Canada in late 2004, the EC estimates for the car in L/100 km were 3.9 urban, 3.1 highway and 3.4 overall. This was when they had 135 section front tires.

 

Canadian models were mechanically identical but had 145 section front tires and were rated at 4.2 L/100 km overall.

 

My first one, with 247,274 km on it when it was rear ended and written off, was a stock convertible - often driven with the top down - and rolled on 175 section front tires and 195 section rear tires. Its measured at the pump lifetime consumption was 3.95 L/100 km.

https://www.spritmon...ail/104881.html

 

My second and current one on 175/195 tires and remapped to over 60 BHP, also a convertible, has a 175K km lifetime average measured at the pump as 4.19 L/100 km.

https://www.spritmon...ail/380023.html

 

My Mercedes B 200 5 speed was rated at 8 L/100 km and mine has a lifetime average over 335,000 km of 7.39 L/100 km.

https://www.spritmon...ail/206221.html

 

So I can't claim to be disappointed there.

 

On the other hand, the EC numbers were something of an exaggeration, though the FE could get down to 3.4 and below lifetime if you were hypermiling one. Not my style.

 

I'll be getting a new Peugeot 308 SW plug-in Hybrid in October for use in Europe and its official EC rating is under 2L/100 km. I an expecting something in the 6.x range in actual use.



#4 Lee Nicolle

Lee Nicolle
  • Member

  • 11,029 posts
  • Joined: July 08

Posted 04 August 2023 - 07:30

All the way to the highest court in Australia. https://www.drive.co...sumption-claim/

Not one of those test stickers mean anything. They never have. Worldwide. They are deceptive as are not made in the real world. With wind, weather, hills taken into account.

Much the same as emmisions that manufacturers are training cars to beat the tests. VW No 1.



#5 GreenMachine

GreenMachine
  • Member

  • 2,607 posts
  • Joined: March 04

Posted 04 August 2023 - 09:49

Nothing in the article to say it is going to the High Court.  In fact, the judgement was handed down in August 2022, it has been appealed to the High Court, the HC has elected to hear the appeal, and it looks like the submissions and hearings are complete and the matter is pending judgement.

 

The nub of the appeal appears to revolve around whether a consumer should expect to substantially replicate the numbers given on the label, if they operate the vehicle in the manner of the test.  Read on:

 

Documents (on the High Court site)*

17/02/2023 Hearing (SLA, Canberra by video connection)

01/03/2023 Notice of appeal

11/04/2023 Written submissions (Appellants)

11/04/2023 Chronology (Appellants)

09/05/2023 Written submissions (Respondent)

30/05/2023 Reply

02/08/2023 Hearing (Full Court, Canberra)

01/08/2023 Outline of oral argument (Appellants)

02/08/2023 Outline of oral argument (Respondent)



#6 Greg Locock

Greg Locock
  • Member

  • 6,346 posts
  • Joined: March 03

Posted 04 August 2023 - 19:08

There's no good way to measure fuel consumption. Most car companies use a convoy test to get a 'real world' figure compared with their competitors, but that can be gamed and is hellish expensive. Using a chassis dyno seems like it would work, but the coastdown tests to calibrate it have rather large error bars. Simulations based on BSFC maps are flaky - I wrote one in 1978 during my apprenticeship and it was still in use 6 years later. So there's a standard method, and of course we develop to that method, so of course the real world fuel consumption suffers.



#7 MikeTekRacing

MikeTekRacing
  • Member

  • 11,818 posts
  • Joined: October 04

Posted 05 August 2023 - 00:07

definition of real world is also very hard - what's real world for me is not the same as real world for other people. 
The mix of short/long driving, city/highway, hills/plain, wind, temperature and driving style - we can't zone in to one number to tell a good story.



#8 desmo

desmo
  • Tech Forum Host

  • 29,165 posts
  • Joined: January 00

Posted 05 August 2023 - 05:14

Calculating realistic fuel economy numbers would be pretty darn simple if the people involved were working in good faith. Alas!



#9 Greg Locock

Greg Locock
  • Member

  • 6,346 posts
  • Joined: March 03

Posted 05 August 2023 - 21:58

Um, speaking as the person who wrote the first fuel consumption estimation program at a large car company, no, it isn't. Consider the differential. It has friction and damping. Those depend on the temperature of the oil, the state of wear of the diff, the speed, and the torque, and how the oil has aged. 



#10 MikeTekRacing

MikeTekRacing
  • Member

  • 11,818 posts
  • Joined: October 04

Posted 05 August 2023 - 22:45

Calculating realistic fuel economy numbers would be pretty darn simple if the people involved were working in good faith. Alas!

i asked above what is a realistic model? there are too many variables not to have a BIG % of error for our individual use cases.

maybe some online calculator where you put in some of your specifics and get a custom estimate?

other than that, one number for everyone is going to be significantly off



#11 scolbourne

scolbourne
  • Member

  • 554 posts
  • Joined: January 09

Posted 06 August 2023 - 04:28

Hybrid cars really exploit the tests. One of the figures provided to the public should be long distance mpg, driving after all battery storage has been depleted.

Currently as long as you make your battery big enough to complete the test , you can claim zero fuel used, resulting in infinite mpg.



#12 cbo

cbo
  • Member

  • 820 posts
  • Joined: September 21

Posted 06 August 2023 - 12:08

My current car have a real world consumption between 8 km/liter to ~ km/liter (plug-in hybrid). The old one(diesel) between 8 km/l and 25 km/l.

Laboratory tests are the only meaningfull way to go, as the values are comparable and repeatable.

"Real world" just has too many variables, including the driver.

The consumer should really just try to get a decent testdrive before buying, testing the vehicle in relevant conditions and try to monitor energy consumption.

#13 gruntguru

gruntguru
  • Member

  • 7,628 posts
  • Joined: January 09

Posted 06 August 2023 - 20:47

The consumer should really just try to get a decent testdrive before buying, testing the vehicle in relevant conditions and try to monitor energy consumption.

Which is exactly what this particular consumer did.



#14 gruntguru

gruntguru
  • Member

  • 7,628 posts
  • Joined: January 09

Posted 06 August 2023 - 20:57

Um, speaking as the person who wrote the first fuel consumption estimation program at a large car company, no, it isn't. Consider the differential. It has friction and damping. Those depend on the temperature of the oil, the state of wear of the diff, the speed, and the torque, and how the oil has aged. 

 

A few percent - maybe ten at most, adding all similar variables?

Mr Begovic's 2016 Triton uses 24% more than his 2008 model in spite of the label claiming a 13% improvement (highway). That adds up to a 42% increase in consumption vs expectations.



#15 Greg Locock

Greg Locock
  • Member

  • 6,346 posts
  • Joined: March 03

Posted 06 August 2023 - 21:17

Yes, the test has become the target. As i said we also ran convoy tests to check that real world figures compared with competitors weren't too bad, not all companies do that, or don't care.



#16 desmo

desmo
  • Tech Forum Host

  • 29,165 posts
  • Joined: January 00

Posted 07 August 2023 - 20:04

Cars today are massively online, it doesn't seem unrealistic that real-world data from them could be made into a database of empirical mileage data. Maybe use the simulated/test data to write a provisional rating, then quickly reset the figure based on the data collected as it becomes sufficiently available. 



#17 GreenMachine

GreenMachine
  • Member

  • 2,607 posts
  • Joined: March 04

Posted 07 August 2023 - 22:01

I think this is over-thinking the issue.  I wrote a massive rant earlier in this thread, but never posted it.  The gist was 'my real world is not your real world'  The tests provide a benchmark, and more usefully as a comparison between makes/models - it maybe there is a need to revisit the test methodology in the light of the gaming and cheating.

 

Demo's suggestion has some merit, except I think his premise is unsound - and if they are on-line, they are on-line to manufacturers and we are back to the source of the problem.

 

The only people who have reliable data are the vehicle leasing companies, where fuel and mileage is recorded at purchase, and I am not convinced that there aren't some problems there too.



#18 MikeTekRacing

MikeTekRacing
  • Member

  • 11,818 posts
  • Joined: October 04

Posted 07 August 2023 - 22:18

I think this is over-thinking the issue.  I wrote a massive rant earlier in this thread, but never posted it.  The gist was 'my real world is not your real world'  The tests provide a benchmark, and more usefully as a comparison between makes/models - it maybe there is a need to revisit the test methodology in the light of the gaming and cheating.

 

Demo's suggestion has some merit, except I think his premise is unsound - and if they are on-line, they are on-line to manufacturers and we are back to the source of the problem.

 

The only people who have reliable data are the vehicle leasing companies, where fuel and mileage is recorded at purchase, and I am not convinced that there aren't some problems there too.

that was my point too. I bet if we all get the same car on this forum our numbers will vary A LOT.

just because we all run very very different scenarios. 

 

For electric cars, where the energy baseline is very low (a lot less loss via heat), sometimes every 1mph difference produces a significant impact...or a few degrees of temperature, or just a mild hill climb/head/tailwind. 

If your car needs 100 KWh of energy per 100miles, adding AC or removing it or any other factor might change 5-10KWh.

If an electric car needs 30KWh of energy per 100miles, 5-10KWh becomes super sensible and noticeable 



#19 GreenMachine

GreenMachine
  • Member

  • 2,607 posts
  • Joined: March 04

Posted 07 August 2023 - 23:00

Sorry Mike, I meant to reference your post, but got distracted in my haste. :up:



Advertisement

#20 404KF2

404KF2
  • Member

  • 18,742 posts
  • Joined: October 99

Posted 08 August 2023 - 05:00

Spritmonitor has excellent empirical data posted by hundreds of thousands of owners like me. The main problem is the newest models take a while to get posted as the cars put on significant distances. Fuelly is another. Sure, some people there are hypermilers, but they're outliers in the mega data on those sites.

 

The OBCs of the two Mercedes I've owned lie like hell. The actual "at the pump" fuel consumption is ~0.45 L/100 km higher. My B200 would have me believe that it averages 6.9 L/100 km. It's still very good at under 7.4, but come on now.



#21 GreenMachine

GreenMachine
  • Member

  • 2,607 posts
  • Joined: March 04

Posted 08 August 2023 - 05:56

My point is 'so what?'  Just what do other people's (claimed) fuel consumption have to do with mine?   I maintain that as an objective standardised test, the placarded results are more meaningful to me - because I know where they are coming from, and I know my usage.

 

OBC results are irrelevant to this discussion -they are just a variant of 'my real world' results and of little use for predictive/comparative purposes.  If you are interested enough to look at them, you are already doing the work yourself.



#22 Greg Locock

Greg Locock
  • Member

  • 6,346 posts
  • Joined: March 03

Posted 08 August 2023 - 06:04

404KF2, which drive cycles are your placard numbers based on? 



#23 MikeTekRacing

MikeTekRacing
  • Member

  • 11,818 posts
  • Joined: October 04

Posted 08 August 2023 - 17:05

Spritmonitor has excellent empirical data posted by hundreds of thousands of owners like me. The main problem is the newest models take a while to get posted as the cars put on significant distances. Fuelly is another. Sure, some people there are hypermilers, but they're outliers in the mega data on those sites.

 

The OBCs of the two Mercedes I've owned lie like hell. The actual "at the pump" fuel consumption is ~0.45 L/100 km higher. My B200 would have me believe that it averages 6.9 L/100 km. It's still very good at under 7.4, but come on now.

Getting numbers is not hard.

What do. you do with the excellent empirical data. Do you average it? If so, what will that number say?

 

If the numbers in the field are wildly different, average doesn't really say anything 



#24 desmo

desmo
  • Tech Forum Host

  • 29,165 posts
  • Joined: January 00

Posted 09 August 2023 - 05:15

It says as much as any averaged quantifiable value will. You gotta ignore the leadfoot and hypermiler crazies. 



#25 GreenMachine

GreenMachine
  • Member

  • 2,607 posts
  • Joined: March 04

Posted 09 August 2023 - 06:52

... and just how do you use it?  Not the abstract you, the you-who-want-real-world-numbers you?



#26 desmo

desmo
  • Tech Forum Host

  • 29,165 posts
  • Joined: January 00

Posted 09 August 2023 - 13:22

The EPA mpg ratings are probably mostly about fleet averages and penalties which explains why the automakers always seem to want to act in bad faith to fudge/falsify them, but real-world data would also help comparison shoppers vs. artificially spun test/simulation numbers. Those artificial test numbers are junk data compared to copious real-world user data. 



#27 gruntguru

gruntguru
  • Member

  • 7,628 posts
  • Joined: January 09

Posted 09 August 2023 - 21:20

Fuel economy is not as variable as you might think. The two biggest variables for a given vehicle are drive cycle (city vs hwy) and driving style. Automatic transmission significantly reduces the variability due to driving style.

 

Many years ago I collected data from the trip computer of my 6 cylinder Ford EF Fairmont sedan. At each fill-up I recorded average economy and average speed and plotted one against the other. Over something like 40 tanks of fuel the data fitted a trendline with amazing repeatability ie, if I knew the average speed for a period of driving, I could predict the fuel consumption for that period within a couple of percent.

 

The "Average Speed" number was actually just an indicator of the percentage of city vs highway operation.



#28 404KF2

404KF2
  • Member

  • 18,742 posts
  • Joined: October 99

Posted 09 August 2023 - 23:19

404KF2, which drive cycles are your placard numbers based on?

Urban/highway combined, Transport Canada test cycle.

#29 Greg Locock

Greg Locock
  • Member

  • 6,346 posts
  • Joined: March 03

Posted 10 August 2023 - 02:05

I think that Canadia uses HWFET which is the same as EPA https://dieselnet.co...ycles/hwfet.php

 

100 seconds to get to 50 mph. Hmm. Also there's a tolerance in your speed at a particular time, this can be gamed.



#30 Greg Locock

Greg Locock
  • Member

  • 6,346 posts
  • Joined: March 03

Posted 10 August 2023 - 02:07

This is fun

 

Accounting for the Variation of Driver Aggression in the Simulation of Conventional and Advanced Vehicles

 

https://www.nrel.gov...3osti/58609.pdf

\

Figure 5 in particular


Edited by Greg Locock, 10 August 2023 - 04:00.


#31 cbo

cbo
  • Member

  • 820 posts
  • Joined: September 21

Posted 10 August 2023 - 06:32

Which is exactly what this particular consumer did.


Did he?

As I read the text, he bought a new car assuming that the lab data would correspond to real world data collected by himself.

Then learned that this was not the fact.

If he had mapped out a route, tested in his old car, and driven the same route in the new car under similar conditions, he would likely have noticed the difference between lab data and real world data BEFORE buying.

What I am saying, is that the owner could have dodged the bullet.

The lab-tests are not worth much, consumers have known this at least since VW cooked their tests.

As I understand the ruling, the car was not considered defective, Mitsubishi had conducted the labtests properly and the company was not ordered to take the vehicle back or compensate the owner.

What the ruling says, basically, is that the system is useless and has to be replaced or abandoned.

I just have some difficulty seeing the alternative.....

#32 GreenMachine

GreenMachine
  • Member

  • 2,607 posts
  • Joined: March 04

Posted 10 August 2023 - 08:10

What the ruling says, basically, is that the system is useless and has to be replaced or abandoned.
 

 

Assuming the High Court finds for the claimant, there are the two possibilities you identify, plus

 

1. Modification of the label to more clearly differentiate the test results from what a buyer might get; and/or

 

2. Modify the test to make it more representative.

 

The court may find for the defendants though, in which case there will be less immediate pressure for change.  However judges often make 'obiter' remarks, where they make all sorts of observations stemming from the evidence/circumstances as the whim takes them, about for example the usefulness of the test results.  While not binding like reasons for a particular judgement, they can be persuasive in subsequent similar cases.  Importantly, the powers that be read these, and sometimes they think, 'yeah, we really need to do something about those tests to improve them'.

 

So win lose or draw, there is a real possibility that change may come from the judgement, whatever it says about who wins the case.



#33 cbo

cbo
  • Member

  • 820 posts
  • Joined: September 21

Posted 10 August 2023 - 12:24

@GreenMachine: Good points.

I remain sceptical about the idea that you can make tests that will reflect the actual milage customers get. The guy driving on a dirt road in northern Sweden in winter and the guy driving a motorway in Spain in the summer will get different milage regardless.

Unless you make a large number of measurements using mobile labs and combine them into values for road types and climate conditions for each market.

#34 gruntguru

gruntguru
  • Member

  • 7,628 posts
  • Joined: January 09

Posted 10 August 2023 - 23:00

The consumer should really just try to get a decent testdrive before buying, testing the vehicle in relevant conditions and try to monitor energy consumption.

 

 

Did he?

As I read the text, he bought a new car assuming that the lab data would correspond to real world data collected by himself.

Then learned that this was not the fact.

If he had mapped out a route, tested in his old car, and driven the same route in the new car under similar conditions, he would likely have noticed the difference between lab data and real world data BEFORE buying.

What I am saying, is that the owner could have dodged the bullet.

 

Sorry, I missed where you said he should have arranged for an extended test drive. Pretty difficult to do - especially if the vehicle is not fitted with a trip computer - would require a "very" extended test drive and accurate refilling method.



#35 gruntguru

gruntguru
  • Member

  • 7,628 posts
  • Joined: January 09

Posted 10 August 2023 - 23:02

This is fun

 

Accounting for the Variation of Driver Aggression in the Simulation of Conventional and Advanced Vehicles

 

https://www.nrel.gov...3osti/58609.pdf

\

Figure 5 in particular

Heh heh. Up to 2.5 times the hwfet number!



#36 cbo

cbo
  • Member

  • 820 posts
  • Joined: September 21

Posted 10 August 2023 - 23:16

Sorry, I missed where you said he should have arranged for an extended test drive. Pretty difficult to do - especially if the vehicle is not fitted with a trip computer - would require a "very" extended test drive and accurate refilling method.


You can actually buy a car today that does not have a trip computer?

#37 404KF2

404KF2
  • Member

  • 18,742 posts
  • Joined: October 99

Posted 11 August 2023 - 03:10

They lie like hell so I'd not pay attention to them...



#38 GreenMachine

GreenMachine
  • Member

  • 2,607 posts
  • Joined: March 04

Posted 20 October 2023 - 00:03

Salvation is at hand: https://realworld.org.au/

 

Remains to be seen what they say about my real world, which is the only one that matters. :up:  :wave:

 

I am not holding my breath, there seems to be nothing on that site that actually spells out what is 'real world', and how it is going to be measured.



#39 cbo

cbo
  • Member

  • 820 posts
  • Joined: September 21

Posted 20 October 2023 - 14:35

Salvation is at hand: https://realworld.org.au/

Remains to be seen what they say about my real world, which is the only one that matters. :up: :wave:

I am not holding my breath, there seems to be nothing on that site that actually spells out what is 'real world', and how it is going to be measured.


"The cars, utes and vans being tested are being driven along a 93km route in and around Geelong that includes urban and rural roads, as well as a stretch of motorway.

Test drives follow prescribed testing protocols based on the European Commission’s Real Driving Emissions test procedure, adapted for Australian driving conditions and speed limits."


https://realworld.or...ke-to-the-road/

Advertisement

#40 Greg Locock

Greg Locock
  • Member

  • 6,346 posts
  • Joined: March 03

Posted 20 October 2023 - 21:57

Breathless graduate runs into the performance office "They're going to start real world mpg testing of cars! We're doomed" Cigar smoking senior engineer raises eyebrow, "OK what's the test protocol?" "Dunno" "Find out and we'll include it with all the other test cycles"

 

Both Austin-Rover and Ford used to run convoy mpg tests with the new car, the old car, and a couple of competitors.



#41 scolbourne

scolbourne
  • Member

  • 554 posts
  • Joined: January 09

Posted 29 October 2023 - 02:49

The result they need to include is for a long trip at about the max speed limit for the country. Ideally this should start with an empty battery, but otherwise should be long enough that the battery capacity is not a major factor,

This result would be as well as the standard tests, which might be more meaningful to normal daily use, but probably account for a small percentage of miles actually driven.

 

The test I propose would show the benefits of an aerodynamic shape, and give a real world mpg for hybrids, after the battery is depleted.



#42 Tenmantaylor

Tenmantaylor
  • Member

  • 17,958 posts
  • Joined: July 01

Posted 30 October 2023 - 10:05

Breathless graduate runs into the performance office "They're going to start real world mpg testing of cars! We're doomed" Cigar smoking senior engineer raises eyebrow, "OK what's the test protocol?" "Dunno" "Find out and we'll include it with all the other test cycles"

 

Both Austin-Rover and Ford used to run convoy mpg tests with the new car, the old car, and a couple of competitors.

 

This is why I tongue in cheek suggested the real world testing should be done on the same route 20 times by the same idiot on normal roads. It would be way more accurate and cheaper and impossible to cheat.


Edited by Tenmantaylor, 30 October 2023 - 10:05.


#43 desmo

desmo
  • Tech Forum Host

  • 29,165 posts
  • Joined: January 00

Posted 31 October 2023 - 14:16

Right. The more the test is nailed down in every possible parameter to make it repeatable, the more easy it will be to defeat by cheating. The testing parameters ideally should be kept completely secret so the manufacturers cannot attempt to optimize their vehicles for a specific test cycle. Real world data from actual cars driven by ordinary drivers would probably be best, but Tenman's "normal idiot" would still be better than any artificial lab-dyno test.



#44 Tenmantaylor

Tenmantaylor
  • Member

  • 17,958 posts
  • Joined: July 01

Posted 31 October 2023 - 15:30

I often try to find those out before buying a new car, parkers seems fairly reliable based on user uploaded data



#45 GreenMachine

GreenMachine
  • Member

  • 2,607 posts
  • Joined: March 04

Posted 01 November 2023 - 23:17

Right. The more the test is nailed down in every possible parameter to make it repeatable, the more easy it will be to defeat by cheating. The testing parameters ideally should be kept completely secret so the manufacturers cannot attempt to optimize their vehicles for a specific test cycle. Real world data from actual cars driven by ordinary drivers would probably be best, but Tenman's "normal idiot" would still be better than any artificial lab-dyno test.

 

No, the lab tests have a valid purpose - comparability, and for that repeatability is an absolute prerequisite.  That is their only value, as by definition no test will tell me what my consumption will be, and only a fool would look for it.  If I look at those lab results I can compare different models, and based on my actual consumption, make a reasonable prediction about what model will give me the best result as regards consumption.  That is making the heroic assumption that the gaming is under control, and the cheating eliminated, and I am realistic enough to know the pitfalls there.

 

Maybe this 'real world' testing will provide repeatability and therefore comparability, but being conducted in the 'real world' the 'degree of difficulty' is high - it will be interesting to watch how it goes, and to learn more about the methodology of the testing, and the processing of the data.



#46 scolbourne

scolbourne
  • Member

  • 554 posts
  • Joined: January 09

Posted 06 November 2023 - 02:30

If they perform the correct tests and the testing information is public then the customer can calculate what there mpg will be according for their driving routine.



#47 Lee Nicolle

Lee Nicolle
  • Member

  • 11,029 posts
  • Joined: July 08

Posted 08 November 2023 - 08:39

You can actually buy a car today that does not have a trip computer?

Yes!



#48 Lee Nicolle

Lee Nicolle
  • Member

  • 11,029 posts
  • Joined: July 08

Posted 08 November 2023 - 08:56

They lie like hell so I'd not pay attention to them...

20 years ago when doing a LOT of motorsport trips I had a few different Fords, all used cars as towcars. Instant MPG was very impressive on a 5 litre Fairlane but in real terms used more than the read out... Surprise! I also used a 4. 0 Fairlane which was a little better and the long stroke 6 was torquier than the 302. 

Though the best was a 3.9 Multipoint wagon. Best on fuel and best performance.

Worse MPG, an 89 Pajero 2'6 petrol, I could drive 300k with the Ford than the 100k or so to Mallala. My 307 HK Premier did the same trip on less than half the fuel and was a good deal faster as well.

In more recent time my 4.5 4 speed auto Landcruiser was in reality a better towcar than the 4 .7 V8 petrol. And had better final drive as well. Both use similar amounts of fuel. Though

as a car alone the V* is nicer to drive and more economical. Big long stoke 6 v short stroke V8 again. And I do not like nunga nungas, have used them as well and slow and not very economical. Even turbo early models. A friend has a 14 V8 Diesel, goes ok towing a house behind it. These Tortoise people!



#49 GreenMachine

GreenMachine
  • Member

  • 2,607 posts
  • Joined: March 04

Posted 30 November 2023 - 06:18

Salvation is at hand: https://realworld.org.au/

 

... and salvation is here!  More precisely https://www.drive.co..._medium=partner

 

The massive differences between the lab testing and this 'real world' testing should disturb us all, and lead to a massive government crackdown on these irresponsible and untrustworthy manufacturers, big fines and ... oh wait ... maybe there is not a lot wrong with the lab tests that a bit of tweaking can't fix. 

 

I suppose I should find whatever the testing people have reported, as opposed to the journo's interpretation, but if 13% is the worst we can all go back to square wun. :cool:



#50 404KF2

404KF2
  • Member

  • 18,742 posts
  • Joined: October 99

Posted 30 November 2023 - 06:39

Fuel consumption testing was once a collateral bit of info from EPA emissions testing.

 

I preferred the old "constant 90 km/h" and "constant 120 km/h" figures from Europe. I usually found them pretty accurate during highway driving. Whatever the city cycle was seemed to align with my experience too.

 

I constantly beat Transport Canada FE figures now.