Honda CR-V Owners Club Forums banner

1 - 20 of 192 Posts

·
Registered
Joined
·
866 Posts
Discussion Starter #1
*statistically it is improbable with some limitations. Read on!

I conducted a statistical analysis of 54 fillups. There's much more analysis of other variables (not shown) but 54 tanks isn't enough to go deeper than this. If everything pans out I'll be able to tell if you if summer vs. winter gas matters and if gas brand matters, among other things.

First off, I know this goes against what others have said. I haven't seen anything close to this robust, nor controlled for weather and driving style, which this analysis is.

First off, statistics is what my degree is in and I practice it daily.

Over the 54 fillups, since June 29th, 2017 to today, I have varied the octane between 87 and 93 somewhat randomly (I always tried to do 2-3 tanks in a row before switching, sometimes more). I ended up with 33 fillups @ 87 octane and 21 @ 93 octane.

I also collected, from the iMid, average speed on Trip A and then reset it after each fillup. I'm using this as a very rough analog of driving style. This is a commuter car and most of my driving is consistent on the interstate.

MPG was manually calculated, I didn't rely on iMid for that.

Finally, I downloaded NOAA average daily temperature data from NOAA's temperature station at a nearby airport, and averaged those across the dates between fillups. So I have average outside temperature per tank measured.

Even though you might be thinking 50+ fillups is a lot, it isn't. I have collected more than this but I can't make very many conclusions that I feel comfortable won't change as I collect more data. Further limitations of these results are that it is one car and one driver, and I only looked at a standard linear model. By using a linear model, it allows me to account for temperature variation and driving style (somewhat), and find out if octane has any impact on MPG. This is a lot more accurate than driving on two tanks doing a comparison.

  • For every MPH in average speed, MPG improved by 0.09
  • For every degree in ave outside temp, MPG increases by 0.07
  • As stated, Octane doesn't matter!

So far the only things that are definitively impact MPG are Avg Speed and Avg Outside Temperature. This shouldn't be a shocker (driving style and temp). It doesn't mean there aren't effects for other things, but if they exist, they're so small that they're getting obscured.

Other fun stuff: My grand total average MPG is 26.2MPG. It was highest last August at 29.1mpg and lowest in January at 24.4mpg.

The average for 87 Octane was 26.3MPG and the average for 93 Octane was 26.1MPG.

Please let me know if you agree, or disagree, but please do so with information on how I might improve the analysis!

There *may* be a confounding factor, and that is the quality of gas at different stations. I'm collecting gas station data, but there isn't enough to control for it yet so you'll just have to wait. My early analysis is trending that "gas brand matters" but nothing definitive yet.

Here is the model output for those that are interested... it basically says that octane isn't even close to being significant.

glm(formula = MPG ~ avegastanktemp + MPG.Gasoline + MPG.Ave.Speed)

Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 19.152502 0.801383 23.899 < 2e-16 ***
avegastanktemp 0.071103 0.009751 7.292 9.94e-09 ***
MPG.Gasoline 93 Octane -0.056209 0.286170 -0.196 0.845329
MPG.Ave.Speed 0.088711 0.021119 4.201 0.000155 ***
 

·
Super Moderator - How may I help?
Joined
·
2,433 Posts
...

  • For every MPH in average speed, MPG improved by 0.09
  • For every degree in ave outside temp, MPG increases by 0.07
  • As stated, Octane doesn't matter!
Up or down on average speed and up or down on average temp?

I guess the speed is obvious, but temp? Here in AZ when average temp is up, mpg goes down as A/C is used. I get better mpg in "winter" than in summer.

You don't say where you are located.

But sounds like a very detailed evaluation. Thank you for doing this and posting.
 

·
Registered
Joined
·
866 Posts
Discussion Starter #3
Up or down on average speed and up or down on average temp?

I guess the speed is obvious, but temp? Here in AZ when average temp is up, mpg goes down as A/C is used. I get better mpg in "winter" than in summer.

You don't say where you are located.

But sounds like a very detailed evaluation. Thank you for doing this and posting.
Increase in both. Northeast. I use the air conditioner heavily in summer.

Could be loss of engine efficiency in cold weather is larger than air conditioner effect.

Start collecting data!

Sent from my Pixel XL using Tapatalk
 

·
Super Moderator - How may I help?
Joined
·
2,433 Posts
Increase in both. Northeast. I use the air conditioner heavily in summer.

Could be loss of engine efficiency in cold weather is larger than air conditioner effect.

So you are saying if you go faster you get better MPG? :Huh:
 

·
Registered
Joined
·
866 Posts
Discussion Starter #5
So you are saying if you go faster you get better MPG? :Huh:
I drive mostly interstate... My average tank speed can be as high as 45mph. That's mostly an average of 75 and whatever city driving I do.

Hwy driving is more efficient than city, yes.




Sent from my Pixel XL using Tapatalk
 

·
Registered
2017 EX-L AWD
Joined
·
122 Posts
You've obviously given this a lot of thought and study. I can't comment on the octane relation, but a couple of things struck me.

Your mpg history is much lower than my experience. Over 16000+ miles, 58 fill-ups, my calculated mpgs range from 25.7 to 37.6. Life of the car average is 32.3 (calculated). All using 87 octane gas, no ECO mode, EX-L AWD, home base is Vermont. The lowest mpgs were this past winter with very cold temps and no long trips. I'm not a hyper-miler. I drive the car as I want without obsessing over fuel economy.

This is the most fuel efficient car I've ever owned from way back starting with a 40 hp VW. Ironically, this is the 2nd smallest engine car (VW was 1.2l), yet performs very well, automatic transmission, air conditioned, can actually climb a hill, go faster than 72 mph, etc. Technology has come a long way.

Wife and I are retired and have taken a few long road trips, but in normal duty, the car is used for grocery hauling, shopping, and typical suburban low mileage excursions. Miles per day driven between fillups range from 8 to 420 with a median of 48.

Your comment on speed vs mpg seems counter intuitive or needs a qualification on what speed range you're referencing. I've found that speeds approaching 70 mph or more significantly drop mpg.
 

·
Registered
Joined
·
3,182 Posts
Great thread/post/info. Tons of other articles on the net also say that premium will not improve mpg over regular presuming the vehicle is designed to run on regular and so stated in the owners manual. Where I live, 91 octane cost 27+% more than 87 octane. No thanks.
 

·
Registered
Joined
·
1,598 Posts
Very interesting report and analysis. My undocumented observations are that head or tail winds along with brand quality are significant variables that are difficult to track.
 

·
Registered
Joined
·
866 Posts
Discussion Starter #9
Very interesting report and analysis. My undocumented observations are that head or tail winds along with brand quality are significant variables that are difficult to track.
Good thoughts... I am tracking brand, but since I split it among 3 (mostly) I need more time to get enough data. There is a trend here.

I'm also including winter vs summer gas. This will be hard to account for since daily temp is correlated... But since I have almost a full year in theory this effect is not creating a confounding effect on anything else. There's pretty good info out there showing that this does matter.

Re: wind, at least in the North East our wind comes from all directions, so I don't expect a long term effect (you win some you lose some). But at the individual drive level I'm sure it does. I have all of the hourly weather data, but not sure how to include wind since I am not modeling individual drives.

Weight / number of passengers also would reduce mpg. Another variable I'm not tracking and therefore part of error.



Sent from my Pixel XL using Tapatalk
 

·
Registered
Joined
·
866 Posts
Discussion Starter #10
You've obviously given this a lot of thought and study. I can't comment on the octane relation, but a couple of things struck me.

Your mpg history is much lower than my experience. Over 16000+ miles, 58 fill-ups, my calculated mpgs range from 25.7 to 37.6. Life of the car average is 32.3 (calculated). All using 87 octane gas, no ECO mode, EX-L AWD, home base is Vermont. The lowest mpgs were this past winter with very cold temps and no long trips. I'm not a hyper-miler. I drive the car as I want without obsessing over fuel economy.

This is the most fuel efficient car I've ever owned from way back starting with a 40 hp VW. Ironically, this is the 2nd smallest engine car (VW was 1.2l), yet performs very well, automatic transmission, air conditioned, can actually climb a hill, go faster than 72 mph, etc. Technology has come a long way.

Wife and I are retired and have taken a few long road trips, but in normal duty, the car is used for grocery hauling, shopping, and typical suburban low mileage excursions. Miles per day driven between fillups range from 8 to 420 with a median of 48.

Your comment on speed vs mpg seems counter intuitive or needs a qualification on what speed range you're referencing. I've found that speeds approaching 70 mph or more significantly drop mpg.
Thanks for sharing! Interesting stuff.

I think my mpg is lower than others because I do my hwy driving at 75. I don't have an easy way to find the optimum speed without recording each drive, which I don't want to do.... Tracking at each fill up is enough. I think the key here is that I'm pretty consistent in how I drive each tank so the long term trends holds and is probably why I can see a significant effect at all.

I heard somewhere that window sticker reported fuel ratings are at 60 or 65mph, I bet that anything above is less efficient.



Sent from my Pixel XL using Tapatalk
 

·
Registered
Joined
·
51 Posts
Good stuff and interesting

Is the car garage kept?
I noticed a big different in the winter from garage kept and not (for obvious reasons).
10 degrees outside and 45 in our garage.
 

·
Super Moderator - How may I help?
Joined
·
2,433 Posts
.

I heard somewhere that window sticker reported fuel ratings are at 60 or 65mph, I bet that anything above is less efficient.
And that is why I questioned your statement that as speed goes up your mpg also goes up. If you did your interstate driving at 65 or even 70, I am pretty sure your mileage would go up.

BUT!...this is not the subject of you analysis and thread. Sorry
 

·
Registered
Joined
·
2,187 Posts
Thanks for sharing that unbelievably detailed information and I really appreciate the honesty in stating the variables that you could not account for. Of course the conclusion fly's in the face of what many people have heard before, but who knows your testing this specific engine so I have no doubts that even if your slightly off in the end result numbers the conclusion is most likely correct that their is really no real benefit in using higher octane gas in the CR-V.


Rob
 

·
Registered
Joined
·
313 Posts
I've collected per fill miles and gallons on 11 different vehicles ('86 Isuszu Trooper to '16 Honda Civic, a 2k CR-V, a 2007 CR-V, a 2009 135i, ...) we've owned dating back to the early 1990s. I have 1,585 fills, 20,900 gal, and 511,000 miles of data in my dataset. So I have some experience with the problem at hand.

There are way too many variables that you can't control for, in the kind of analysis you are trying to do, to support the fine-grained conclusions you are trying to draw. The biggest single one that affects the per-tank calculation is how full you do or don't get the tank--how accurately you replace the exact quantity used for the miles you just recorded. Your two to three fill per flavor method is not enough, in my experience, to average out the fill-to-fill variance caused by different pumps, different pump flow rates, patience, jiggling the nozzle, etc. Also there are effects like driving in light vs dark, using defroster, are the tires checked for pressure every tank, snow vs dry, etc., ad nauseum. You are trying to normalize to NOAA temperatures but doing it daily not when the vehicle was running. And assuming that a NOAA temperature somewhere close reflects what you are seeing whereever you are. I could go on...

I'm not knocking what you are trying to do nor disagreeing with your basic premise about "excess" octane as an insignificant contributor in an engine tuned by Honda to run on regular. But to get meaningful data to prove your point, you'd need to control the variables far better than you have a prayer when doing "real world" testing because, well, the real world has too many variables. You'd have to do this kind of fine grained test using OBD data and doing it on a dynamometer and with calibrated fuel supply and equal time engines for both fuel types and so on.

p.s., there's a reason the EPA numbers are obtained on a dyno under highly controlled conditions "driving" tightly specified test cycle conditions.
 

·
Registered
Joined
·
182 Posts
I would like the statistics on the difference in mpg using straight, unadulterated regular gasoline versus the 10% ethanol regular we have been mandated to use by politicians.
 

·
Registered
Joined
·
313 Posts

·
Registered
Joined
·
2,922 Posts

·
Registered
Joined
·
3 Posts
Lower speeds will consume more fuel as you'll be in the lower gears, so the engine is running faster and using more fuel. This particularly applies to urban (city, etc.) driving. As speeds increase and you stay in higher gears, less fuel is used. However, once you reach the higher speeds (I seem to recall, from a long time ago [ so it may have changed] 40mph was the threshold) the "drag factor" caused by the shape of the vehicle, which largely equates to the vacuum / low pressure caused at the back of the vehicle has an increasing effect. When the drag = the thrust of the engine (inc friction losses, etc.) you reach the vehicle's top speed. At the higher speeds fuel consumption follows a law of diminishing returns. So the graph would probably be similar to a "Normal" or "Gaussian" curve, possibly skewed.
 

·
Registered
2016 CRV Touring AWD
Joined
·
3,580 Posts
I suspect you need a staff to help you account for or discount all the disparate variables affecting mpg.
This is fascinating study. Hope you keep it up.
 
  • Like
Reactions: lmaxim

·
Registered
Joined
·
866 Posts
Discussion Starter #20
p.s., there's a reason the EPA numbers are obtained on a dyno under highly controlled conditions "driving" tightly specified test cycle conditions.
These are all fair criticisms of limitations of a real world test vs a controlled lab test. Nobody drives in a highly controlled way. If the difference between 87 and 93 were large enough to practically matter to how I drive over 15,000 miles, I'd see it in this analysis. I'm very confident of that.

Sent from my Pixel XL using Tapatalk
 
1 - 20 of 192 Posts
Top