Hacker News Reader

I disagree.

I think that most people think about gas in the following way (I know that I do). I commute distance x a day to work, how much am I going to pay in gas?

The 100kms/liter figure makes this straightforward. I can figure out how many km I'll drive in a typical year, multiply by that figure, and I have how many liters. Multiply by the cost of gas, and I have how much I have out of pocket.

With mpg I have to do a division step in there instead of multiplication. Division is a more complex operation. Laugh if you will, but people are more likely to screw it up.

MPG comparisons are pretty intuitive, but where MPG comparisons fall apart is when you're comparing comparisons.

Let's say you have a 10-MPG truck and a 50-MPG hybrid, and you drive both equally far each year, and you've got a $40k budget to replace one or the other.

Should you spend your money on a new truck that gets 11 mpg or a whiz-bang new EREV that gets 100 mpg (equivalent)?

Most people would think that improving the miles-per-gallon by 200% saves much more fuel than improving it by about 10%, but in this case both replacements save the same amount of fuel per mile driven.

If you invert both numbers it's immediately clear that this is the case (10 versus 9 gal/100mi, 2 versus 1 gal/100mi).

Don't forget miles driven per vehicle either.

It has been found empirically that people are worse at reasoning about MPG than L/100km. That's the argument this post should have made. Source: http://wheels.blogs.nytimes.com/2008/06/20/the-illusion-of-m...

That's the argument it tantalizingly came close to making at the end, but it stopped dead in its tracks before reaching it.

First of all, the complaint has nothing whatsoever to do with the metric vs non-metric issue---we could as easily report gallons per hundred miles if it mattered that much.

Exactly.

The measurement of gallons per 100 miles shows a much clearer picture of gas usage:

  MPG = gallons per 100 miles (100/MPG)
  14 = 7.1
  17 = 5.9
  33 = 3.0
  50 = 2.0

  MPG = Cost of gasoline to drive 15,000 miles at $4/gal (15000/MPG*$4)
  14 = $4285.71
  17 = $3529.41
  33 = $1818.18
  50 = $1200.00

Yeah, most of the time when someone talks to me about gas mileage, they're interested in better mileage in order to save money or conserve gas, not to go further between fill-ups. So phrasing it as a 33% vs 18% savings hides that you actually save more gas, and so more money, going from 14 to 17 than 33 to 50.

That's one question ,but another one is:

How far can I drive per $100?

When looking at the car in part for discretionary uses, that's a valid question.....

The difficulty with this is the widely varied price of gasoline. If we assume $4/gal, $100 would buy us 25 gallons of fuel, so we can just multiply those 25 gallons by the MPG:

  MPG = Miles per $100 (MPG*25/$_per_gallon)
  14 = 350
  17 = 425
  33 = 825
  50 = 1250

> That's one question ,but another one is:

That is a question I've never seen anybody ask in my whole life. There are precious few people who drive for the sake of driving, for the vast majority driving is utility to go from point A to point B.

Therefore, why optimize for the significantly less common (almost irrelevant, even) case of "I've got money burning my pocket, how far can I drive?"

"how far can I drive for this amount of $" is a question I've only heard a few times. But I do often hear "how far can I drive on a tank" or "how far can I drive on my remaining amount of fuel".

Perhaps the disconnect we see in this comment thread is between people who do almost entirely city driving and people who drive long distances. If you only ever drive a few miles at a time, you don't really care what sort of distance you can manage on a tank, you just care how much it's going to cost you to go to work every day (you'll probably convert the fuel/distance ratio into a money/time ratio.) If you spend a lot of time driving across central Kansas, how often you have to stop for gas becomes a more significant consideration (you'll probably convert distance/fuel into stops/time.)

> Perhaps the disconnect we see in this comment thread is between people who do almost entirely city driving and people who drive long distances.

That's an interesting thought, and it probably has more than a grain of truth.

Maybe. I tend to live in rural areas so maybe that's a factor.

But I was more thinking of this:

"We have a vacation budget of X for a weekend trip. We don't want to spend more than Y for gas. What options do we have?"

How often you stop for gas is surely better covered by a "range" measurement for the car? Maybe it's different in the states, but all the cars I've ever been in show your remaining fuel as a fraction of a tank. In that case, volume of fuel is completely irrelevant, so MPG vs liters/100km has not place. e.g. My tank takes me 400 miles => I have half a tank left => I can go another 200 miles ...is exactly equivalent to... My tank takes me 600 km => I have half a tank left => I can go another 300 km

(For the purposes of this post, 1 mile = 1500m. If the size of a gallon changes with locale, why not a mile? :-p )

Linear with respect to your budget. The post is correct, that putting consumption on top gives a better "feel" for how much money you'll save.

Correct? When did comparisons of interpersonal utility schedules become a factual matter?

I agree that the article is poorly reasoned. I think part of it is the writing style. The first two sentences make no sense to me. I still cannot figure out what the "related measures" are related to and I have never understood why people type the contraction "we're." It makes me wish I could place a thousand upvotes on the recent "please learn to communicate" article.

When psychologists were invented. People understand linear measures better than inverse measures.

But linear vs inverse depends entirely on what questions you are asking.

If I want to know how much discretionary driving I can afford, MPG is a better measure. If I want to know how much the car will cost to operate over fixed driving requirements (commuting, etc) GP100M is a better measurement. Maybe we can just have both ;-)

…and most people aren't asking that question. Gasoline is an incredibly inelastic commodity.

Don't tell the economists...

In practical terms it's not that clear cut. Let's say I have an old car that gets 14 mpg and I'm considering two new cars, one that gets 33 and one that gets 50 mpg.

If I drive 500 miles a week, either vehicle is going to save me at least 20 gallons a week in fuel. But going from the 33 to the 50 is only going to save me 5 more gallons. However to get 50 mpg I probably need to buy an expensive new hybrid, whereas it's not too hard to find a good used car that will get 33 for a fraction of the new hybrid's purchase price. Then you have to consider cost of financing and insurance (both more for the new car). If I go for the 50 MPG car I'll probably never make up the difference in fuel savings unless I drive a LOT of miles.

You do not need a hybrid to get 50 mpg. Many diesel cars easily do more than 50 mpg