Another advantage of biking to work

Posted in personal at 4:49 pm by danvk

I filled up on gas today for the first time in almost exactly a month.

It’s always struck me that miles per gallon is a terrible metric. I can increase my average miles per gallon by going on a long road trip with lots of highway driving. But I certainly haven’t saved any gas or money by doing so. The better metric is gallons per week, or maybe even dollars per week. By those standards, I’m using 2.6 gallons/week = $7.70/week at $3/gal. I suspect those are both quite low, whereas I only get just over 20 miles per gallon, a pretty piddling number. What are your numbers?


  1. Jack Hardcastle said,

    August 12, 2007 at 10:38 pm

    I had several hours to think about this yesterday, while driving back from a week-long vacation. Miles per gallon is a car-specific metric. It’s not terrible. It’s just not the correct metric for the situation you’re talking about. It’s perfect if you want to compare one car to another car in a vacuum (like if you were trying to decide which to buy). Your dollars per week metric is individual-car combined. It pairs your driving habits, with your car, with your car’s mileage per gallon.

    Additionally, “dollars per week” is a poor metric, because it assumes that the price of gas remains constant, or it requires that one specify a fixed price of gas as you have done here. A better metric (in this case) would be gallons per week. Yes, this is a more abstract notion and difficult for the layperson to understand outright, but it’s a necessary intermediary calculation if you want to compute dollars per week, and it’s more flexible. The size of a gallon doesn’t change (but for once every 150 years).

    For about 18 months, I used to use [TownName]GasPrices.com (e.g. SanFranGasPrices.com) to track every gas purchase I made. By entering the total number of gallons, the price-per-gallon, and the odometer reading (which I write on the back of my receipt) you’re able to get some very precise statistics. Sadly, I have about 24 months worth of back receipts in my truck.

    Lately, I use, depending upon the week, somewhere between 12 and 25 gallons per week.

    Thanks heavens I’m not the only person thinking about this crap. Although you are much more well-spoken on the subject. Also, I apologize for the comment being longer than the original post. As I said, I had lots of time to think. :D

  2. Ryan Moulton said,

    August 13, 2007 at 8:30 am

    When I calculated my gas mileage recently I also computed a “days per gallon” metric. You can see the graphs online here: http://spreadsheets.google.com/pub?key=pYP8Gq6hLnkguFZ3UTQ9vUw
    I average ~2 days per gallon. (Dan and I have similar commute situations, but I’m lazier about biking.)

  3. danvk said,

    August 13, 2007 at 11:56 am

    Jack — the metric you use depends on what you’re trying to optimize. As you say, for comparing two cars, miles per gallon is good. For comparing lifestyles and saving the planet, gallons per week is good. For purely economic decisions, like deciding whether you get enough better mileage from using premium to offset the cost, dollars per week is good. (I’ve heard you get better mileage from better gas, but don’t know if it’s true). In our case, you’re right, gallons/week is the way to go.

    I mentioned dollars/week mainly because I’d just been reading about corn farming. Farmers always talk about “bushels/acre” as their measure of efficiency, but this is misleading for similar reasons. You can spend tons of money on genetically modified corn, fertilizer and pesticides to increase your output per acre. But have you actually increased your profits? Probably not. For them, you really need to add money to the equation and measure profit per acre.

    By the way, Jack, you’re my inspiration for tracking gas mileage, and the reason I have a pen handy in my car! =)

  4. Greg Grothaus said,

    August 13, 2007 at 10:49 pm

    Actually, Octane and MPG are orthogonal, one does not affect the other. However, cars are generally tuned for a specific octane (most of them 87) and they run most efficiently at that octane fuel, so only indirectly can octane affect MPG.