In article <1113597048.495091.18170@f14g2000cwb.googlegroups. com>, "Bucky" <uw_badgers@email.com> wrote:
>No wonder the EPA estimates are so far off.

So many say. But I've found them to come reasonably close to my
mpg. Reasonably. People have very different driving styles, in
very different driving environments. So there will be no single
(or dual) test that will do a good job for near everyone. But I
agree they can be better than the current city/highway tests, and
the EPA knows that too.

And FYI, they adjust the numbers realizing that these tests are
too optimistic. The city number is multiplied by 0.9, the highway
by 0.78(!). One might say that is a clear indictment and screams
for a re-write of the tests. I might agree. But it takes a lot
to change such tests. And in the long run, most will still bitch.

>Also, the EPA tests on a "treadmill".
>http://www.fueleconomy.gov/feg/how_tested.shtml

What's wrong with that? They could do it on a dedicated test
track. But things like air T variations, rain, snow, wind
changes will make this very problematic.

>Supposedly, they account for drag: "The energy required to move the
>rollers can be adjusted to account for aerodynamic forces and the
>vehicle's weight." But do they actually vary the simulated drag
>differently for each vehicle's drag coefficient?

Yes, and Yes.
Each car's coastdown curve is included and loaded into the
profile. So it accounts for air drag vs speed. There are some
things that are less than perfect. But as I said before, no test
is.

These are supposed to be *estimates*. They are best at giving a
relative rating. People who expect them to be perfect, or even
within +/- 1-2 mph for 90%+ of the population are just expecting
too much. NO test will do so.

The one case where it seems pretty far off is hybrids. Hybrids do
allow for a potential strong case of "cycle beating" with their
power managment strategies. And these high mpg, especially if
lightweight, vehicles can be more significantly affected by things
like defrosters and a/c on. Which the EPA tests do not do. Here,
I think we'd benefit from a comprehensive, impartial study to be
done and publicized. So people will know what to expect from
hybrids.