> This doesn't mean that weather forecasting, economics, agile software methodologies -- to pick just a few famous examples of subjects dealing with complex sys

> This doesn't mean that weather forecasting, economics, agile software methodologies -- to pick just a few famous examples of subjects dealing with complex systems -- are totally broken tools that must henceforth be discarded. Not that I want to ridicule you here, but it is interesting that two of your three examples fucked up so bad at various times we WOULD be better off if we discarded them (or even better: Send them back to research until they provide reliable results) - economics and weather forecasting. For the third one the bets are still open. For the general case/dieting: The observation that a net calorie deficit always leads to weight loss seems correct. The observation that you can always induce such a deficit is also correct. Using this knowledge to state "eat less, you will lose weight" leads to what I call the "dumb and proven way" - It will work(1), but your body will(2) fight you tooth and nail at every step, i.e. your body will do everything to reduce the net deficit. The open question: Is there a way to get a comparable/bigger net deficit without fighting your body? Keto, paleo and whatever try to provide an answer. For me, they fall into the agile software methodologies category: There may be benefits. Or not. (1) It will work if you can stay with it, which is unrealistic for many cases, but people still like the "idea". (2) Some bodies fight hard, some not. That's life.