[schoen]

When I saw this, I was guessing at first that the author had confused physics calories with nutritional calories; they differ by a factor of 1,000 and cause many back-of-the envelope calculations like this to come out the wrong way. For example, there was someone who had calculated that the caloric content of a scotch and soda was negative and so consuming them should cause weight loss. But the hidden flaw was that the energy required to melt the ice and warm up the liquid was actually in physical calories (which are tiny), while the energy supplied by metabolizing the alcohol was in nutritional calories (which are huge).

http://www.futilitycloset.com/2008/11/16/the-mensa-diet/

In this case, the author doesn't give his measured power output but I took a look at

https://en.wikipedia.org/wiki/Bicycle_performance#Energy_inp...

and concluded that the cited number of (nutritional) calories is the right order of magnitude, so the author's estimate is physically plausible and didn't involve a mistake about the two meanings of "calorie". The factor that Wikipedia notes that I might have neglected is about muscular efficiency: you can only actually use a fraction of the food energy you metabolize in order to move yourself around.

On the other hand, you can get this much food energy by eating about 120 g olive oil or about 200 g peanut butter. The peanut butter (choosing a random bottle from the front page of Amazon.com) offers 1213 kcal/dollar, a big improvement over the author's 205 kcal/dollar average.

I would also point out that if you're substituting at the margin, you don't have to think about "how terrible would it be if I had to eat only peanut butter all the time to save money?"; you just have to think about "how terrible would it be if I had to eat a couple of spoonfuls of peanut butter per day to power my commute?", which doesn't sound that terrible to me.