## Thursday, November 6, 2008

### Fun with multiple regression

So I've added datapoints for 10% and 15% ethanol for my gas mileage and average speed. I threw out Alcohol's data, so now I've got data purely for my car. I've redone my regressions for mpg as a function of ethanol % and average speed.

It turns out that ethanol content is a statistically significant factor in mpg.

The next step was to take the regression equation and calculate MPG for various ethanol contents.

Then I looked at what price spread you need to make it worth it to use ethanol.

I found that it was anywhere from 57 to 72 cents. It's a spread that's quite a bit more than you get at Meijer!

Subscribe to:
Post Comments (Atom)

## 2 comments:

What is the breakeven % price difference Buzz? Cents per gallon differentials only have meaning at a set price.

Alcohol

Sorry. It's a 25% to 38% spread. Assuming E85 is $2, you'd need gas to be $2.50 to make it work.

In reality, today it was $2.21.

Post a Comment