> If you don't think telling people to eat less "works", then telling people to smoke less didn't "work" either. I think my metric for whether advice "works" i

> If you don't think telling people to eat less "works", then telling people to smoke less didn't "work" either. I think my metric for whether advice "works" is whether people attempting to seriously follow the advice have at least, say, a 1-in-20 chance of success. By that metric "smoke less" does work and "eat less" does not. When people follow the advice "eat less" it seems to have about a 1% chance of success 5 years later, whereas "smoke less" has at least a 10% chance of success. If only one of those numbers is meaningfully positive, only one of those sets of advice "works".