# Gas mileage

Hypothetical: If GM makes 1000 supposedly identical Chevy Cobalts and if each of those cars is driven under exactly the same conditions in exactly the same way over exactly the same 10,000 miles, what would be the expected variation in gas mileage, the normal distribution. (What about the same question for a Honda Civic?)

Interesting question. Without actual data, I can only guess. And that guess would be + or - 2 mpg varition from the norm. There are so many variables, tire pressure, wind speeds, etc. Assuming you equalize everything there will still be a variation from car to car. They can’t assembled exactly the same, so my guess + or - 2 mpg. This variation is so small as to be unnoticed in real life since all the variables simply can’t be controlled.

Just to show the subtle variations, I bet the white and silver ones would do a little better in the south.

Sometimes, there are variations that defy explanation but there isn’t usually a variation that’s large enough to matter. Nothing is ever equal so we’ll never really know.

I think with identical cars in truly identical conditions, the variation would be well below 1 mpg. I think it’s amazing how consistent city mileages actually are with so many variables in play.

The variation wouldn’t be enough to matter. That goes for a Cobalt or a Civic.

The differences would be due to the drivers more than the cars.

When the EPA measures mileage they do it indoors under controlled conditions.