It's not particularly interesting but I thought I'd post the mileage results they get. They tend to drive cars normally so their results are good for comparison purposes.

With the 2006 sedan MT, they got 31 mpg mixed, and 37 mpg on a highway trip.
With the 2006 sedan AT, they got 28 mpg mixed, and 33 mpg on a highway trip.
With the 2006 hybrid, they got 37 mpg mixed, and 45 mpg on a highway trip.

With the 2005 hybrid, they got 36 mpg mixed, and 44 mpg on a highway trip. So 1 mpg worse.

With the 2005 Prius, they got 44 mpg mixed, and 48 mpg on a highway trip.

With the 2000 Insight, they got 51 mpg mixed, and 61 mpg on a highway trip.

Thanks for the info!

--

Wayne, if CR altered their driving technique then the results would be totally invalid. If CR were to drive the hybrid gently then they would have to drive the other cars gently too, in which case it will be 40 mpg EX vs 50 mpg Hybrid so the difference would still be 10 mpg. Not to mention the fact that that's completely redundant since the EPA already drives them very gently and we all know the EPA scores. CR reflects more real-world values.

City was again 26 mpg. That's not unrealistic unless one is very careful with coasting but then CR would have to drive the regular cars the same way since they too would benefit (though obviously not as much). Plus it's meant to be a worst-case scenario value. The regular automatic got 18 mpg city. Anyway, for highway they got 47 with the hybrid.

I want to repeat that the results are excellent for comparison. They're not maximum values one can achieve; obviously that would be EPA or higher, anyone knows that.

"Wayne, if CR altered their driving technique then the results would be totally invalid. If CR were to drive the hybrid gently then they would have to drive the other cars gently too, in which case it will be 40 mpg EX vs 50 mpg Hybrid so the difference would still be 10 mpg. Not to mention the fact that that's completely redundant since the EPA already drives them very gently and we all know the EPA scores. CR reflects more real-world values."

Herein lies the problem, definition of "real-world". "real-world" is a term that's not real in the general sense. It is real in the specific individual sense. My drive to work is real. Your drive to work is not, from my perspective. "Real-world" only has value when interpreted from the individual's definition of "world".

"Real-world" in the CR context only has value as an artificial label assigned to a specific set of environmental conditions and actions taken to perform a test. Assumption being that these conditions and actions can be applied in a repeatable manner to achieve the same results to enable validity of making relative comparisons. CR for any of their tests to have consistent value, must be performed exactly the same every time. To be a strict purist, this is impossible. However, that does not mean that meaningful data cannot be derived with a suitable acceptance of error margin.

As you point to, the real value of this CR report is to determine relative performance, not absolute given an acceptance that error margins due to environmental conditions and repeatability of actions can be significant, especially when performed outside the lab environment and where humans are directly involved.

Jesse
ps. My definition of "real-world" is colored by a view of "can I get to work and back safely, without impeding traffic.

'06 HCH II w/Navi
'01 S2000

True but CR clearly must be testing in a consistent manner if you look at their results with the 2005 and 2006 HCH.

City/Highway/150-mile/Average:

2005 HCH: 26/45/44/36
2006 HCH: 26/47/45/37

Extremely close with the '06 HCH being just a tiny bit better, and that's what you'd expect given its improvements. In fact right now on GreenHybrid the '05 CVT is 45 mpg and the '06 CVT is 46. It's hard to argue that CR's testing is inconsistent in light of that.

Can't validate consistency based on the data meeting expected results, though it is a sanity check factor. All of my experiments through college physics and chemistry classes were notably consistent with my expectations.

Consistency of data can be a fluke at best or engineered at worst. Though what I believe is at most meaningful here, is that the relative differences between the different test scenarios per vehicle can be compared against each other.

A reasonable claim can be made that "The new technology in the HCH II did not significantly impact the ratio of city vs. highway mileage compared to the HCH I." An unreasonable claim would be to tell someone "Real-world highway mileage from this test has been shown to be 47 mpg." One could say that "From results, a driver of an HCH II may expect to see marginal highway mpg improvements vs. an HCH I"

'06 HCH II w/Navi
'01 S2000

But that's much different because every single FE result achieved in the car is 100% "correct". But it can yield vastly different FEs depending on how it's driven. The issue here was whether or not CR drives their cars in a consistent manner, and it seems like they do since the FEs of two cars, which should get similar mileage, actually did. It's not an issue of being consistent with expectations, it's an issue of being consistent. You're using a different definition of the word 'consistent'.

Uh, no. At best it is actual genuine consistency. At worst it is a fluke or a lie. I do not think CR lies, do you? And I think it's odd to attribute realistic results to mere coincidence ("a fluke"), in this situation.

What? That makes no sense! The ratio of city/highway driving is entirely up to the driver, not the car.

The first statement is inaccurate but not incorrect. CR did get 47 mpg driving probably 60-65 mph on the highway, as do many people here. But of course that doesn't mean everyone will.

The second statement is, of course, perfectly accurate.

I checked CR's results and was surprised that the 2006 Civic 5 speeed auto didn't do as well as the older 4 speed Civic 4 speed auto.The 2006 got 18 mpg in CR's city loop,the old one got 20-or 22 mpg(my memory is a bit vaque,but I think it was 22 mpg city and 40 hy) in the city loop.The corolla auto got 20 mpg in the city loop.Heck the Accord 4 cyl auto got 16 mpg in CR'S city loop(38 hy).
The 2006 was probably tested in cold months-the last 2 maybe-I'm not sure if CR attempts to correct for temp or different fuel blends.Thanks.Charlie

I'm sure they do or their results would likely be much worse. The larger engine (140 hp vs 115) and heavier weight could be responsible for the slightly reduced city mileage.