Moderator, should this be a sticky??

Great info here

I've been out of the loop for a while....... swamped with my job lately. I see that we've gotten quite a few additional respondents since I first did the regression analysis, and will provide an update either later this week or early next week.

I will show these results of this analysis on the other similar thread, titled Results of Statistical Analysis Approach.

Boy, I've really been out of the loop. I didn't realize there was a separate results thread. Read all about it here in case you also missed it.

I look forward to updated data plots and reading about the statistics underlying this whole thing.

bump

This looks like an enlightening statistical analysis of the MPG issue. See also the link given by "8_wannabe" above.

Another factor that may affect MPG is cold-short drives. A good discussion about the relationship between MPG and cold-short drives is here:
https://www.rx8club.com/showthread.p...e&pagenumber=2
( RX-8 Forum > General Topics > RX-8 Discussion > My current MPG, the saga continues......)

Starting with "pp13bnos" at 11-28-2003 08:01 PM

One interpretation is that for cold-short drives the engine isn't warm, is driving with the "choke" on, and getting lower MPG than it would if it were warm.

Everything here makes perfect sense. I know this is premature and unscientific, but my guess is there isn't a fuel economy problem necessarilly. The rotary engine (Even the Renesis) is not very efficient outside certain ranges. It's amazing how many people are getting close to or above 20mpg though (combined no less). These figures really resemble the posted numbers from Mazda. If you drove it like you stole it, you're outside the range: please don't be amongst the disappointed.
Average ambient temps and trip duration are the missing vital stats as pointed out by Trx8. I have concluded; again prematurely and unscientifically, that 8s driven in low enough temps for relatively short distances will get similar figures to the pedal to the metal crowd: end of mystery.
Can they improve it? Probably.
If you hadn't noticed, I'm rather opinionated. I apollagise.
Great thread norton.

for me short trips kill the mpg. and i think more so now in winter. ive noticed that my mileage is reduced by 1 to 3 mpg when i take short trips out for lunch. i mean like just a mile or 2 so the engine barely has time to warm up. just 2 lunches out per week seems to be enough. if i eat in or walk and drive the same route to and from work 1 get 17 to 18. just 2 trips for lunch will drop it to 15-16.

In the MPG statistical analysis I conducted several months ago, several variables were explored:
- Percent Highway
- Driving Style
- VIN
- Octane
- Region (State)
- Model
- Auto vs 6-Speed
- Windows (% time open)
- AC (% time on)

The only ones significantly impacting MPG were Percent Hwy and Driving style. Based on the data and analysis, I came up with the following estimate of MPG using Driving Style and % Hwy as determinants:

After conducting the analysis, it seems that one factor in particular continues to get mentioned as likely impacting MPG. This is average length of trip (i.e. impact of many cold starts vs few cold starts - since the RX-8 evidently runs quite rich when cold). Unfortunately, I never included this factor originally when I requested inputs from owners.

If there is much interest, I would be happy to redo the entire analysis to now include "# of cold starts" &/or "average length of trip". We could then assess the true impact of this variable on MPG. Unfortunately, even if it does impact MPG a lot, there isn't much each owner can do to change this factor.

Please let me know if you're interested. Then I'll set up a thread for inputting the data.

I dispute this. On highway driving the most significant variable is speed.

At 45mph I get 22mpg
At 60mph I get 18.5mpg
At 85mph I get 15.9mpg
At 110mph I get 12.9mpg.

This is based on actual highway trips.

Yes, there IS a road not far from here you can drive about 150 miles at 110mph. Goof thing there is a gas station at the halfway point.

The simple fact is that if one lived in Europe, and used this car on the Autobahn or similar high speed national highways it might not be affordable to drive.

Here in Alberta, where the accepted speed on major highways is around 135 to 150kmh ( 83 to 93mph) I get some pretty awful milage, averaging around 15.8mpg

It's nice to drive, though!

Norton, I think your analysis is right on. It fits with everything I've seen and read on this forum. Thank you for all your hard work on this.

Canzoomer's mileages seem a little low for the respective mph (not counting the 45 mph one) during highway cruising.

And what's up with the high mileage for 45 mph? Are you in 5th or 6th gear?

"I dispute this. On highway driving the most significant variable is speed.

At 45mph I get 22mpg
At 60mph I get 18.5mpg
At 85mph I get 15.9mpg
At 110mph I get 12.9mpg.

This is based on actual highway trips."

Canzoomer: Others have made this same observation as you - that greater highway speeds have a decreasing impact on MPG. I agree with you. When I was saying "The only ones [factors]significantly impacting MPG were Percent Hwy and Driving style", I also said, "Based on the data and analysis". What I meant was those two factors were the ONLY ones impacting MPG of those variables I collected. If I had collected "Hwy Speed" this too would have likely had a significant impact on MPG. Unfortunately, those variable (as well as other potentially impacting variables) were not included.

We need to be realistic about interpreting the results of norton's regression analysis. The usefulness of the results are bounded by the inputs and the underlying model. Its common to find cases that are out-of-bounds, and some of them have been listed above. In those cases, interpretation of the results need to be tempered by the conditions that norton has been careful to qualify.

In my opinion, norton did an exceptional job of:
- Creating the idea of using a regression analysis
- Setting up the model
- Soliciting input
- Inputting and tabulating raw data from the forum
- Cranking the numbers
- Interpreting, posting, and explaining results

His results are good and usable, they are given as:
- Tabulated raw data
- Regression output
- Simplified algebraic model
- Tabular form based on the two dominant factors
- Running discussion
- Likely things I missed

I know some of my driving is out of bounds (cold-short drives). From what he has reported, I have derived a better basic understanding of my MPG situation, and what to do about it.

norton -- thanks for the very useful results

You're welcome Trx8.
Before I started the analysis, I was really hoping to find one or two things that people could actually control to significantly improve their mileage. Unfortunately, from the data provided I only found "% Hwy" and "Driving Style" as impacting MPG. Things like % Hwy are pretty much a given for each person; that is a way of life for them determined by the area they live in, their commute, etc. They don't really have the ability to change % Hwy. People can impact their MPG by easing up on their driving style, but that kind of goes contrary to what the car was intended for.
Basically, as others have stated, it appears to me that the RX-8 overall reacts similar to many other cars to things that impact MPG (e.g. short trips, many cold starts, hwy speed, hard accelaration, etc). It just seems that the RX-8 is just FAR MORE sensitive to these factors.

This thread has given me more peace of mind about my '8 than any other. Before your analysis, I woulda sworn my car was defective and needed some kind of ECU fix. But based on my highway percent + driving style, I see I am right smack on your distribution curve. I could get better mileage if I tried, but I don't feel like trying. Thanks, Norton.

You're welcome, 8_wannabe.
I am a prospective RX-8 owner, considering for the future. If I had a shorter commute to work, I myself would not be so concerned with the mileage issue. However with a 42 mile commute EACH way, MPG becomes a consideration at some point -- both from a cost perspective and also having to refill every couple days. Luckily, the bulk of my commute (95%) is Hwy. Unfortunately, this Hwy driving is N.J. Hwy, which is a lot of stop and go, and creeping and crawling, combined with 80 mph sprints.

Therein lies the weakness in your poll: Highway does not always equal highway. Stop 'n go don't count.

I know. You are right. That's why a while back I had proposed also incorporating "Average Speed". This would have gotten around the weakness you've noted above, however it was not easy to calculate.

Re: Statistical Analysis Approach to Understanding MPG Issue
 

Originally posted by norton ...
I know this MPG topic has been beaten to death, and I apologize for bringing it up again, however I would like to propose a new approach to understanding this MPG issue.

THANKS FOR TAKING THE TIME TO RUN SOME FIGURES!!!
hope this helps ...

1. RECENT Combined City/Highway Average MPG = last tank:
12.8 mpg (and pissing me off PLENTY!!!)

2. % Highway Driving = 25%

3. Driving Style = B

4. VIN Number (last 6 digits) = 109611

5. Model (eg. Base, GT, etc) = touring (not GT though)

6. Automatic / 6 speed = 6 (of course!!! ;)

7. Octane (Number) of Fuel Used = 91

8. DSC/TCS = Yes / No(Don’t Have or Fully Disabled) almost always on

9. Miles to Date = 1600

10. Approximate % Time Driving with Windows Open = 20%

11. Approximate % Time Driving with AC on = <5%

Wow man, those actual mpg you guys are getting is pretty bad, that's about as bad as driving around in an automatic GTO - LOL. I guess you guys who are waiting for the '05 model for whatever reason are getting nervous about being hit with a gas guzzlers tax if the feds re-evaluate the RX-8 again.

- Mike

1. 15.1
2. 20%
3. B
4. 111524
5. touring
6. 6 speed
7. 93
8. yes
9. 1577
10. 0%
11. 0%
12. MD

I also have a long commute (50+ miles each way) in NJ. Virtually all of my driving is along Rt78 into Jersey City. The creeping and crawling combined with the 80 mph sprints equal out to around 19 MPG consistantly. The best I've done is a bit over 20, the worst a bit under 18. Manual transmission, just under 4000 miles. It's subpar mileage, but it's manageable. I like the 8 enough that the extra 3-5 MPG I'd get in a comparable car are a worthwhile sacrifice.

Hope that helps a bit.

more input
 

1. 18.32
2. 60
3. B
4.
5. GT
6. 6 speed
7. 93
8. Yes
9. 1196
10. 0
11. 10
12. FL

RECENT Combined City/Highway Average MPG = 14.85
(Calculated Correctly and One Number please, not a range)

2. % Highway Driving = 20%

3. Driving Style = B
A. Easy going (Don’t normally use high rpm range, only occasionally)
B. Average,
C. Above Avg (Use Upper RPM range quite often and accelerate hard)

4. VIN Number (last 6 digits) = 119291
(NOTE: If you already entered your VIN on the thread “Let's compare VIN numbers and fuel economy”, I can get it from that thread).

5. Model (eg. Base, GT, etc) = GT Loaded no Nav

6. Automatic / 6 speed = Auto
7. Octane (Number) of Fuel Used = 93

8. DSC/TCS = Yes

9. Miles to Date = 883

10. Approximate % Time Driving with Windows Open = 0%

11. Approximate % Time Driving with AC on = 0%

12. State = NJ

(I’ll later convert to Geographic Region)