Welcome to the North American Subaru Impreza Owners Club Tuesday September 23, 2014
Home Forums WikiNASIOC Products Store Modifications Upgrade Garage
NASIOC
Here you can view your subscribed threads, work with private messages and edit your profile and preferences Home Registration is free! Visit the NASIOC Store NASIOC Rules Search Find other members Frequently Asked Questions Calendar Archive NASIOC Upgrade Garage Logout
Go Back   NASIOC > NASIOC Subaru Models > Impreza Forum

Welcome to NASIOC - The world's largest online community for Subaru enthusiasts!
Welcome to the NASIOC.com Subaru forum.

You are currently viewing our forum as a guest, which gives you limited access to view most discussions and access our other features. By joining our community, free of charge, you will have access to post topics, communicate privately with other members (PM), respond to polls, upload content and access many other special features. Registration is free, fast and simple, so please join our community today!

If you have any problems with the registration process or your account login, please contact us.
* Registered users of the site do not see these ads.
Reply
 
Thread Tools Display Modes
Old 04-07-2013, 05:08 PM   #176
G-Omaha
Scooby Specialist
 
Member#: 209172
Join Date: Apr 2009
Location: Omaha, NE
Vehicle:
2012 Impreza CLL
BL2

Default

My 2012 is not a PZEV. I pre-ordered it and it was one of the early builds. I swapped out the OEM tires and pay a 1 or 2 MPG penalty, something that I am willing to do for the decreased noise and improved traction.

Something to be said about "Stats, Damned Stats, and Lies" though. Statistically, I get between 6 to 8 MPG better gas mileage than my son. He gets in the neighborhood of 24 or 25 MPG in urban driving. Now, how does that skew the "deltas"? He is also about 100 lbs heavier than I am - maybe that, along with his heavy right foot has something to do with the MPG.
* Registered users of the site do not see these ads.
G-Omaha is offline   Reply With Quote
Old 04-07-2013, 08:14 PM   #177
flyboy1100
Scooby Specialist
 
Member#: 314216
Join Date: Mar 2012
Location: ND
Vehicle:
2012 2.0i Sport 5MT
DGM

Default

Quote:
Originally Posted by G-Omaha View Post
My 2012 is not a PZEV. I pre-ordered it and it was one of the early builds. I swapped out the OEM tires and pay a 1 or 2 MPG penalty, something that I am willing to do for the decreased noise and improved traction.

Something to be said about "Stats, Damned Stats, and Lies" though. Statistically, I get between 6 to 8 MPG better gas mileage than my son. He gets in the neighborhood of 24 or 25 MPG in urban driving. Now, how does that skew the "deltas"? He is also about 100 lbs heavier than I am - maybe that, along with his heavy right foot has something to do with the MPG.
Nope, driver has no effect on mpg...
flyboy1100 is offline   Reply With Quote
Old 04-07-2013, 09:14 PM   #178
G-Omaha
Scooby Specialist
 
Member#: 209172
Join Date: Apr 2009
Location: Omaha, NE
Vehicle:
2012 Impreza CLL
BL2

Default

That's what I thought. Statistically, that wouldn't matter - even though in the "real world" it could. But then what do I know, I'm not a math major - just a computer programmer (crap in equates to crap out in all cases)...
G-Omaha is offline   Reply With Quote
Old 04-08-2013, 03:42 AM   #179
stevehnm
Scooby Specialist
 
Member#: 329526
Join Date: Aug 2012
Chapter/Region: SWIC
Location: Ground Control
Vehicle:
2013 Impreza Spt cvt
Silver

Default

Quote:
Originally Posted by G-Omaha View Post
That's what I thought. Statistically, that wouldn't matter - even though in the "real world" it could. But then what do I know, I'm not a math major - just a computer programmer (crap in equates to crap out in all cases)...
You're correct, different drivers will get different results, obviously. The thing is with the drivers and the different test procedures and all the other variables between the two tests worlds apart in methodology, the random error as standard deviation comes in at 1.18 mpg. That means that whether you compare ten or a hundred different vehicles, about 65% will differ by less than 1.19 mpg.

So basically what we are saying is that the two different tests come within about one mpg of each other most of the time.

Except when you throw in the Subarus. Then the Impreza CVT comes back six mpg low on a real road compared to its EPA highway estimate. But hey, if you think that's normal...
stevehnm is offline   Reply With Quote
Old 04-08-2013, 08:07 AM   #180
flyboy1100
Scooby Specialist
 
Member#: 314216
Join Date: Mar 2012
Location: ND
Vehicle:
2012 2.0i Sport 5MT
DGM

Default

Yes, normal.

http://autos.yahoo.com/news/here-s-w...175746134.html

Either slow down or sell the car, but everyone is tired of your bitching about it, especially when you routinely speed and have a very large hill to climb daily.
flyboy1100 is offline   Reply With Quote
Old 04-08-2013, 08:55 AM   #181
stevehnm
Scooby Specialist
 
Member#: 329526
Join Date: Aug 2012
Chapter/Region: SWIC
Location: Ground Control
Vehicle:
2013 Impreza Spt cvt
Silver

Default

Quote:
Originally Posted by flyboy1100 View Post
Yes, normal.

http://autos.yahoo.com/news/here-s-w...175746134.html

Either slow down or sell the car, but everyone is tired of your bitching about it, especially when you routinely speed and have a very large hill to climb daily.
Thank you. This is yet another perfect illustration of why you just don't or can't understand. I am beginning to think it's can't. This has nothing to do with my driving. It's not about me.

The article you linked to is about individual drivers and their correlation with the EPA test numbers. This is about systematic error.

There are two types of error. Random Error and Bias Error, aka Systematic Error.

I am talking here about how closely Consumer Reports' tests correlate with EPA tests. They won't be perfect either, especially considering the different methods. The differing methods means there likely will be a bias offset (aka systematic error) in the results, and indeed there is. However, in this case when that systematic error is removed, the two tests are usually within one mpg of each other in this case (unless Subaru is involved that is). However the Subaru Impreza CVT is 6 mpg away from the others, advertising said Impreza CVT as 6 mpg better than in reality. Reality as in the real world. That is not a random error, that is a systematic error in the measurement of the Impreza CVT mpg. That error is so big that in terms of the bell curve (which I guess you have not heard of) there is a one in three million chance it's normal. That is your position - that it's normal. Sorry, but it's not.

That's okay, so you are not good in math. Except you are apparently not good in reading or logic either. This is about the Impreza CVT. You have a manual transmission. That is different. In fact according to Fuelly.com the manual gets better mpg than the CVT.

The bottom line (what a coincidence, it actually is) is that if you are tired of looking at the posts in this thread, then don't look at this thread because THIS THREAD HAS NOTHING TO DO WITH YOU.

Last edited by stevehnm; 04-08-2013 at 09:04 AM.
stevehnm is offline   Reply With Quote
Old 04-08-2013, 09:08 AM   #182
stevehnm
Scooby Specialist
 
Member#: 329526
Join Date: Aug 2012
Chapter/Region: SWIC
Location: Ground Control
Vehicle:
2013 Impreza Spt cvt
Silver

Default

Quote:
Originally Posted by Zeeper View Post
He is an "outliar", when he started his mpg complaints Blah blah blah
Yet another manual transmission owner who knows nothing about this commenting in a CVT thread.

P.S. When I drive my other car, the same way, I do in fact get well over the EPA estimate.
stevehnm is offline   Reply With Quote
Old 04-08-2013, 09:26 AM   #183
flyboy1100
Scooby Specialist
 
Member#: 314216
Join Date: Mar 2012
Location: ND
Vehicle:
2012 2.0i Sport 5MT
DGM

Default

since reading comprehension is difficult for you
Quote:
Be Honest About Your Driving

The point here is that the common advertising disclaimer of "your mileage will vary" is something every car shopper must take seriously. Honestly assessing your specific situation is the only way to adjust the EPA ratings down — or up — to more accurately reflect the fuel economy you can expect to see once the car or truck is yours.

One way to do a reality check on a car's EPA rating is to deduct 10 percent from it if you consider yourself an aggressive driver. This means someone who routinely exceeds speed limits and hurries away from stop signs and red lights.
You can stick with the EPA rating if you consider yourself an efficient driver. You'd need to reevaluate that self-description if you find that your car or truck is consistently failing to achieve the rated combined average, however.

Real-world fuel economy can also be diminished if you use the air-conditioning a lot, haul heavy loads of cargo or passengers, or even live at the top of a hill.

The Ratings Are Here To Stay

If the EPA ratings aren't a good reflection of actual fuel-efficiency performance, can we expect to see them changed any time soon? The quick answer is no. There's not a lot of will on Capitol Hill to change the way the EPA figures fuel efficiency for passenger vehicles.

In government and policy circles, it's well known that the EPA's fuel economy figures are too optimistic, but changing the system to more closely reflect real-world averages would mean lowering the bar on all the measurements of the current state of fuel efficiency on the American highway, and no one is eager to admit that we really haven't come as far in fuel-efficiency gains as it seems.

The 54.5-mpg average for fuel efficiency among new vehicles that the Obama administration is proposing for 2025 will really be somewhere down around 36-38 mpg in real-world terms.

The MPG Testing Gap

The problem is that the federal government uses two different procedures to compute the fuel-efficiency numbers it provides to consumers. First, there's the original procedure for the Corporate Average Fuel Economy (CAFE) values that is used for regulatory purposes. Introduced in 1975, it is based on standardized highway and city driving cycles that are replicated in a computer and measured on a dynamometer.

But because people complained of a big gap between real-world fuel economy and the test values with this original CAFE test, Congress asked the EPA to make some modifications to the test, which it did in 1985. Unfortunately, the changes in this second-generation test apply only to the window sticker that consumers see for individual vehicles. Also, Congress didn't tell the EPA to modify the test for CAFE purposes.

As a result, the new CAFE measurement — the 54.5-mpg figure that's been in the headlines — is based on the original, congressional-endorsed test that was mandated in 1975. Meanwhile, the revised, EPA-endorsed fuel economy test is the one reflected in the ratings posted on the window sticker in all new cars and trucks. The EPA rating is also the one used in automotive advertising, most car-buying guides and online car shopping and information sites, including Edmunds.com.

And even though the EPA's test system was modified in 1984 and 2008 to try to make it more relevant, it still doesn't precisely reflect real-world driving. The EPA still uses stationary laboratory tests run on a dynamometer, which ensures repeatability but not necessarily the driving experience of consumers.

...


In the meantime, take any claims of fuel economy with more than a few grains of salt.
ps, my car doesn't exceed rated hwy estimates either if i drive over 68mph, so maybe i too should complain to the EPA. Have you submitted your complaint yet?
flyboy1100 is offline   Reply With Quote
Old 04-08-2013, 10:14 AM   #184
stevehnm
Scooby Specialist
 
Member#: 329526
Join Date: Aug 2012
Chapter/Region: SWIC
Location: Ground Control
Vehicle:
2013 Impreza Spt cvt
Silver

Default

Quote:
Originally Posted by flyboy1100 View Post
since reading comprehension is difficult for you


ps, my car doesn't exceed rated hwy estimates either if i drive over 68mph, so maybe i too should complain to the EPA. Have you submitted your complaint yet?
You still don't get it do you?

P.S. but your car does get a lot closer to the EPA estimate than the CVT,
which you don't have and is what this thread is about.
stevehnm is offline   Reply With Quote
Old 04-08-2013, 10:47 AM   #185
Zeeper
Scooby Specialist
 
Member#: 299286
Join Date: Oct 2011
Chapter/Region: NESIC
Location: Albany NY
Vehicle:
2012 Sport 5MT
Green/Silver

Default

Quote:
Originally Posted by flyboy1100 View Post
since reading comprehension is difficult for you


ps, my car doesn't exceed rated hwy estimates either if i drive over 68mph, so maybe i too should complain to the EPA. Have you submitted your complaint yet?
He does not know this, remember his theory is that the 5 speed does better for mpg's than the CVT, no matter how it is driven.

If you drive your 5 speed at the same speeds, on the road he is driving it on, don't you know it would give you even better mpg's?

lol
Zeeper is offline   Reply With Quote
Old 04-08-2013, 02:31 PM   #186
jd_24
Scooby Newbie
 
Member#: 318301
Join Date: Apr 2012
Location: Minnesota
Vehicle:
2012 CVT Sport Prem
Ice Silver Metalic

Default

#1 I have a 2012 CVT so maybe I'm qualified to post.
#2 I hate statistics. They can be manipulated to show whatever you want.
#3 These numbers are from Consumer reports printed edittion that I have.

first number is Consumer reports MPG. The last 3 numbers are city-average-highway for EPA.

Subaru Impreza Prem _____27_____27-30-36
Mazda 3 Skyactive (auto) _____32_____28-33-40
Mazda 3 Grand Tour(2.0) _____25_____25-28-33
Ford Focus SFE (auto) _____31_____28-33-40
Chevy Cruze Eco _____27_____26-31-39
Chevy Cruze _____26_____26-30-38
Ford Focus SE(auto) _____28_____27-31-38
Honda Civic _____29_____28-32-39
Toyota Corolla(MT) _____32_____27-30-34
Jetta TDI _____34_____30-34-42
Dodge Dart _____27_____27-31-37
Jetta (2.5) _____25_____24-26-31

Yes the Subaru number is lower than some. But matches the City EPA number. Most others are within 1 or 2 of the City Number.
Big deal?!?! I think not. Its very close to what most cars return for MPG.

Last edited by jd_24; 04-08-2013 at 02:40 PM.
jd_24 is offline   Reply With Quote
Old 04-08-2013, 02:39 PM   #187
jd_24
Scooby Newbie
 
Member#: 318301
Join Date: Apr 2012
Location: Minnesota
Vehicle:
2012 CVT Sport Prem
Ice Silver Metalic

Default

The Mazda 3 was still 8mpg less than its highway number, but better than its City number. The Corolla did well. The Jeta TDI Did well to, but still far from its highway numbers.

All the rest basically returned the CITY EPA during the Consumer Reports tests.
jd_24 is offline   Reply With Quote
Old 04-08-2013, 02:45 PM   #188
jd_24
Scooby Newbie
 
Member#: 318301
Join Date: Apr 2012
Location: Minnesota
Vehicle:
2012 CVT Sport Prem
Ice Silver Metalic

Default

With the warmer temps (30s and 40s) here in Minneapolis my MPGs have risen. Most of the winter MPG was around 30 to 32 for my commute. Current tank is displaying 36.6mpg which will equate to about 34+ MPG with slightly over 1/2 a tank used.
jd_24 is offline   Reply With Quote
Old 04-08-2013, 02:58 PM   #189
aeoporta
Scooby Specialist
 
Member#: 334126
Join Date: Oct 2012
Chapter/Region: Tri-State
Location: New York
Vehicle:
2013 5mt Sp Premium
Venetian Red & Dark Gray

Default

Can't wait to see the results of the fuzzy math based off these numbers. Could you post the month and year of this consumer reports magazine. Looks like I should get a Honda civic
aeoporta is offline   Reply With Quote
Old 04-08-2013, 03:02 PM   #190
jd_24
Scooby Newbie
 
Member#: 318301
Join Date: Apr 2012
Location: Minnesota
Vehicle:
2012 CVT Sport Prem
Ice Silver Metalic

Default

Quote:
Originally Posted by aeoporta View Post
Can't wait to see the results of the fuzzy math based off these numbers. Could you post the month and year of this consumer reports magazine. Looks like I should get a Honda civic

We just signed up for Consumer Reports this past winter. 2013 auto issue.

Why the Civic? Only +1 over the City EPA. Mazda 3 with Skyactive seems to return the best. However statistically its probably the Corolla.
jd_24 is offline   Reply With Quote
Old 04-08-2013, 04:18 PM   #191
nubsub
Scooby Specialist
 
Member#: 313575
Join Date: Mar 2012
Location: SF Bay Area, CA
Vehicle:
2012 Impreza 2.0i HB

Default

Quote:
Originally Posted by jd_24 View Post
We just signed up for Consumer Reports this past winter. 2013 auto issue.

Why the Civic? Only +1 over the City EPA. Mazda 3 with Skyactive seems to return the best. However statistically its probably the Corolla.
again, there are so many sources for variability...

i am actually quite surprised by the Honda Civic numbers being low. i haven't admittedly driven the new generation myself, but i know someone who bought one last year and i also looked at the fuelly numbers and they are significantly higher than the Impreza real world.

but i agree with the point made here. if we want to do a somewhat reasonable statistics of the CR mileage numbers, we shouldn't just cherry pick cars getting 30mpg in the test. but rather, use the entire database available ranging from high to low mileage cars.

the numbers in the above survey actually make a reasonable study, nevertheless, since it focuses on the small cars class which is presumably what a typical customer might be considering.
nubsub is offline   Reply With Quote
Old 04-08-2013, 06:19 PM   #192
stevehnm
Scooby Specialist
 
Member#: 329526
Join Date: Aug 2012
Chapter/Region: SWIC
Location: Ground Control
Vehicle:
2013 Impreza Spt cvt
Silver

Default

I took all the cars that got 35 mpg highway in Consumer Reports. I didn't cherry pick the data.
stevehnm is offline   Reply With Quote
Old 04-08-2013, 06:27 PM   #193
RJSoobs251
Scooby Newbie
 
Member#: 345206
Join Date: Jan 2013
Chapter/Region: MWSOC
Location: Minneapolis, MN
Vehicle:
2012 Impreza 2.0 Lim
DGM

Default

Quote:
Originally Posted by jd_24 View Post
With the warmer temps (30s and 40s) here in Minneapolis my MPGs have risen. Most of the winter MPG was around 30 to 32 for my commute. Current tank is displaying 36.6mpg which will equate to about 34+ MPG with slightly over 1/2 a tank used.
Same here
RJSoobs251 is offline   Reply With Quote
Old 04-08-2013, 06:39 PM   #194
nubsub
Scooby Specialist
 
Member#: 313575
Join Date: Mar 2012
Location: SF Bay Area, CA
Vehicle:
2012 Impreza 2.0i HB

Default

Quote:
Originally Posted by stevehnm View Post
I took all the cars that got 35 mpg highway in Consumer Reports. I didn't cherry pick the data.
sorry if i came across as critiquing your method. i probably should have chosen different words as i don't mean to say you were trying to bias the result in any way what the cars you chose. you chose the 35 mpg highway vehicles and you got your result which show that the Impreza was a serious outlier. JD more recently posted about using the city numbers for comparison, and they got a different result where the Impreza looks more normal.

this discrepancy speaks to the likelihood that there is no simple way to get sound statistics for this comparison. i was suggesting that for a more grounded approach, both you and JD could do the same analysis for all cars and see what kind of standard deviations you get. even then, i would be surprised if you got a different conclusion. (highway vs. city comparisons)

i do agree that CR is a good starting point, for this tricky comparison, because at least the driver variable is nominally removed. (although i'm sure they have a whole staff of drivers they use for these tests)

i should also note that I do get what you are saying. i can meet EPA numbers on the Impreza, but it requires far more effort than most of the previous vehicles i've driven/owned. if i were to rate the Impreza based on my prior cars EPA estimates, i would have thought that the Impreza were rated as 25/32, rather than 27/36 for the CVT.

(although i should note that most of the cars i have driven were MT, so maybe that makes the personal comparison somewhat unfair, too)
nubsub is offline   Reply With Quote
Old 04-08-2013, 11:07 PM   #195
stevehnm
Scooby Specialist
 
Member#: 329526
Join Date: Aug 2012
Chapter/Region: SWIC
Location: Ground Control
Vehicle:
2013 Impreza Spt cvt
Silver

Default

I went with highway mpg for two reasons. One was because it would be extremely repeatable - one the one hand we have the "hallowed" EPA test results, and on the other a two way constant speed run at 65 mph. The other reason, frankly, is because that is what I am concerned with.

Overall then, I don't do enough "city" or "combined" driving to worry about it, and making it repeatable with continuous repetitive acceleration and braking seemed difficult to replicate - and with the wide spread in results that seems to be the case. On the other hand, the highway mpg correlation was astonishing - usually within one mpg. Except the Impreza CVT... One in 3 million off the mark...
stevehnm is offline   Reply With Quote
Old 04-08-2013, 11:41 PM   #196
G-Omaha
Scooby Specialist
 
Member#: 209172
Join Date: Apr 2009
Location: Omaha, NE
Vehicle:
2012 Impreza CLL
BL2

Default

If you are using highway MPG and a constant run speed of 65 mph - that is NOT the way that the EPA measures it. Static skew? Data used for statistical analysis should be "pure" - I don't think your's is. Just MHO. Skew the data = skewed results.
G-Omaha is offline   Reply With Quote
Old 04-09-2013, 12:18 AM   #197
richde
Scooby Specialist
 
Member#: 79498
Join Date: Jan 2005
Chapter/Region: SWIC
Location: SCIC's trashy neighbor
Vehicle:
2013 XV Silver Slug
'84 Porsche 911-sorta red

Default

Quote:
Originally Posted by stevehnm View Post
Other than the fuel economy, this is a great car. However, there has been much discussion in other threads about the fact the CVT does not get the advertised MPG, backed up by this observation:

The fact that Consumer Reports has tested several vehicles that get the same highway mpg, yet are rated consistently lower. Very consistently. The sedan (at CR's 35 mpg) is particularly woeful, as the other cars are rated at 30 +/- 2 MPG EPA highway, yet still get the same mpg in real world testing, with a tight spread. The actual numbers are a mean of 29.9 MPG with a standard deviation of 1.311, meaning the 36 mpg EPA highway rating by Subaru is so far out in left field that it is 4.63 standard deviations away from the crowd. This means that if there were a half million cars in this same spread, the Subaru Impreza would have a better than 50/50 chance of being the worst.

So, the question is, how can Subaru improve the fuel economy of the Impreza? In the 200,000 miles I plan(ned?) to keep the car it will cost me an extra $5,000 extra dollars for fuel, since the real world EPA highway MPG should be 30 MPG.

I got the information here:

http://www.consumerreports.org/cro/c...erformance.htm

and the actual data is this:

Subaru Impreza sedan Premium 4-cyl CVT___ 36

Acura TL Base V6 ______________________29
Acura TSX 4-cyl ________________________30
Audi A4 sedan Premium 4-cyl______________29
Ford Mustang coupe Premium V6 MT________29
Dodge Avenger Mainstreet 4-cyl___________31
Lexus ES 350 V6_______________________31
Mazda MX-5 Miata Grand Touring 4-cyl MT___28
Nissan Altima 3.5 SL V6 CVT______________31
Subaru Legacy 2.5i Premium 4-cyl CVT______32
Volkswagen CC Sport 4-cyl_______________31
Volvo C30 T5 1.0 5-cyl MT ________________28
Volvo S60 T5 5-cyl______________________30
Quote:
Originally Posted by jd_24 View Post
#1 I have a 2012 CVT so maybe I'm qualified to post.
#2 I hate statistics. They can be manipulated to show whatever you want.
#3 These numbers are from Consumer reports printed edittion that I have.

first number is Consumer reports MPG. The last 3 numbers are city-average-highway for EPA.

Subaru Impreza Prem _____27_____27-30-36
Mazda 3 Skyactive (auto) _____32_____28-33-40
Mazda 3 Grand Tour(2.0) _____25_____25-28-33
Ford Focus SFE (auto) _____31_____28-33-40
Chevy Cruze Eco _____27_____26-31-39
Chevy Cruze _____26_____26-30-38
Ford Focus SE(auto) _____28_____27-31-38
Honda Civic _____29_____28-32-39
Toyota Corolla(MT) _____32_____27-30-34
Jetta TDI _____34_____30-34-42
Dodge Dart _____27_____27-31-37
Jetta (2.5) _____25_____24-26-31

Yes the Subaru number is lower than some. But matches the City EPA number. Most others are within 1 or 2 of the City Number.
Big deal?!?! I think not. Its very close to what most cars return for MPG.
Well, I guess that puts what he's talking about into perspective.

Like most of the cars, the Impreza got right around it's City MPG figures, and only the TDI was able to hit the combined numbers. There must be a problem with the Impreza CVT because it performs like most of the other cars tested.
richde is online now   Reply With Quote
Old 04-09-2013, 12:38 AM   #198
stevehnm
Scooby Specialist
 
Member#: 329526
Join Date: Aug 2012
Chapter/Region: SWIC
Location: Ground Control
Vehicle:
2013 Impreza Spt cvt
Silver

Default

Quote:
Originally Posted by G-Omaha View Post
If you are using highway MPG and a constant run speed of 65 mph - that is NOT the way that the EPA measures it. Static skew? Data used for statistical analysis should be "pure" - I don't think your's is. Just MHO. Skew the data = skewed results.
Yours is an insightful point. To break it down further though, there are only two types of error - that is, all error can be divided into two categories. Random Error (aka sloppiness) and Bias Error, or what you are thinking of as "skew". The two types are independent and as somebody once said - "never the twain shall meet" or something like that. Anyway, with the use of averages and standard deviations the two can be separated. Averages will give bias error, and standard deviations will give random error. So what I did was establish that the Impreza CVT is so far out of the initial field of data it can be classified as an outlier (several references to that, it fits all definitions that I've found), and is therefore rightfully dismissed. Since we have then established Subaru's measurement procedure to be ... unreliable (to be PC) ... we should, in the name of data robustness, eliminate the other result of that origin, that of the Legacy. Coincidentally the Legacy result was the closest to the Impreza CVT anyway.

Now we have the two types of error - random and bias, established as a standard deviation of 1.19 mpg and an average of 6.3 mpg systematic, respectively. It becomes clear once one recognizes and separates the two types of error.

So what we do, then is take your "skew" and remove it - it could be a constant offset or a percentage offset - it really doesn't matter in this case. Then we see that the two measurements - CR and EPA - usually differ by less than one mpg here (except for the Impreza CVT, of course, which is not just in left field, but in a different county so to speak).

I don't think we can define the different errors as well with city or combined due to different throttle responses etc, verified by the increased range of the results.

P.S. If you're still here... Standard Deviation is quite a cosmic thing, actually. It represents the Normal distribution, or Bell Curve. The bell starts from -infinity (it actually never goes to zero), then ramps up as it approaches the average - aka "concave up". Then when it gets close to the average it changes from concave up to concave down (inflects). Then it goes up to the average, then continues concave down to a point where it again inflects and becomes concave up and runs on out to +infinity. Anyway it is universal in nature and mathematics (mathematics is just a crystallization of nature when you get down to it). Those points where it "inflects" or changes from concave up to concave down or vice versa are at what are referred to as "one standard deviation" because between them are clumped most (about 65%) of the data.

Even further if you have held on this long is that highway speed limits are set (or should be, to maximize safety) at the "85th percentile of the unrestricted speed" - the 85th percentile is right about one standard deviation above the average unrestricted speed, where most of the traffic is considered encapsulated in the average, and that above is kind of out in lala land where the data starts to just spread out.

"Lala land". That's a technical term...

Last edited by stevehnm; 04-09-2013 at 01:11 AM.
stevehnm is offline   Reply With Quote
Old 04-09-2013, 01:26 AM   #199
hemophilic
Scooby Newbie
 
Member#: 301213
Join Date: Nov 2011
Chapter/Region: South East
Location: Raleigh, NC
Vehicle:
2012 Imp. Sp. Ltd.
Blue/Silver

Default

Quote:
Originally Posted by jd_24 View Post
Subaru Impreza Prem _____27_____27-30-36
Mazda 3 Skyactive (auto) _____32_____28-33-40
Mazda 3 Grand Tour(2.0) _____25_____25-28-33
Ford Focus SFE (auto) _____31_____28-33-40
Chevy Cruze Eco _____27_____26-31-39
Chevy Cruze _____26_____26-30-38
Ford Focus SE(auto) _____28_____27-31-38
Honda Civic _____29_____28-32-39
Toyota Corolla(MT) _____32_____27-30-34
Jetta TDI _____34_____30-34-42
Dodge Dart _____27_____27-31-37
Jetta (2.5) _____25_____24-26-31
Wow, Corolla and Jetta TDI are the only cars to meet or beat the combined number.
Cruze ECO gets same MPG as Impreza even though rated higher and plain Cruze gets lower MPG in spite of a higher rating.
Looks like the Impreza performs comparably.

I don't have a subscription to CR but I'd love to see the test data on the actual highway numbers for the above compacts and steves list of V6s and premium drinkers at 55, 65, and 75mph. That would be quite enlightening wouldn't it? Does that impartial data exist? My bet is that the CVT Impreza would whip the premium drinkers and V6s at 55, be a bit better at 65, and match them at 75. Any takers?
hemophilic is offline   Reply With Quote
Old 04-09-2013, 01:30 AM   #200
hemophilic
Scooby Newbie
 
Member#: 301213
Join Date: Nov 2011
Chapter/Region: South East
Location: Raleigh, NC
Vehicle:
2012 Imp. Sp. Ltd.
Blue/Silver

Default

Quote:
Originally Posted by stevehnm View Post
Yours is an insightful point. To break it down further though, blahblahblahblahblahblahblahwhere most of the traffic is considered encapsulated in the average, and that above is kind of out in lala land where the data starts to just spread out.

"Lala land". That's a technical term...
Are you high right now or something?
hemophilic is offline   Reply With Quote
Reply

Thread Tools
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Forum Jump


All times are GMT -4. The time now is 07:09 PM.


Powered by vBulletin® Version 3.7.0
Copyright ©2000 - 2014, Jelsoft Enterprises Ltd.
Powered by Searchlight © 2014 Axivo Inc.
Copyright ©1999 - 2014, North American Subaru Impreza Owners Club, Inc.