This paper looks like something put out by the No Fun Allowed Brigade; it's labelling things as 'choices' when it would be more appropriate to label them as 'lifestyles'. The factors of drinking alcohol, doing drugs, and not excercise are going to be more or less dangerous depending on your environment - living in the slums is going to make these behaviorus inherently more risky than they are in suburbia.
I suspect the real purpose of the paper is to say "Look at these dumb lower classes - they keep making the WRONG decisions for themselves. Time for us to step in, and do their thinking for them!"
'Decisions', in my mind, ought refer to the specific events leading up to the situation - not broad categories of decisions.
That's a pretty negative assessment. I see it more good news and useful information: "Pay attention to these things and you (probably) get to live longer."
I believe waist (circumference)/height predicts health problems better than BMI.
I personally choose toward the lower end of the body fat vs. lifespan plateau because it means I'm not bothered very much by an old lower back injury (it's definitely worse when I'm 15 pounds heavier). I know there are many athletic, mobile people with more than the fashion model/bodybuilder amount of body fat, but since it's easy for me to stay relatively skinny, I do so in hopes of reducing wear and tear on my joints.
Another thing to consider: even though it's fine to have quite a lot more fat than most people worry about for aesthetic reasons, eventually you form new, permanent fat storage capacity. Fat is not inert storage; it also produces hormones which will (among other things) make you hungrier, even if you lose the weight later (that is, fat clusters don't immediately die off; they just deflate, but when full, new ones form). For this reason, I will try to avoid setting a new body fat high-water mark.
What's the estimate of longevity-maximizing BMI (given other studies)? Articles linked from the post above seem to suggests about 25 (if we ignore causation/correlation question).
Whose estimate?
I'll give you my estimate, but beware of motivated cognition.
I don't think BMI tells you much about longevity unless it's extreme. The middle of the curve is pretty flat.
Many of the methods of changing BMI aren't safe, and I'm including the risk of getting an eating disorder-- a severe eating disorder will have major longterm health effects.
Afaik, the jury is out on the health effects of repeated weight loss.
I think there are only two rational ways to bet-- either eat and exercise in such a way as to maximize energy and enjoyment of life[1], let your BMI fall where it will, and hope that your qualia feedback system is well correlated with health, or go with theory and do CR.
CR's been around long enough that there should some information about whether it's doing people much good, but I haven't been following that. Have you seen anything?
[1] There are a lot of different theories about that one. Some experimentation will be required.
If I don't take care of my diet, I slowly drift towards obese. I can sustain any necessary weight, also without much limitation on diet, just by cutting some calories. Thus, if I don't control my diet, I get an undesirable outcome, and if I do control it, then it doesn't matter (in some range) what target weight I choose.
Hence, the question. Whatever uncertainty there is, I need to make a decision. The current hypothesis for best target BMI range is 23-24 (I'm currently at 24.5).
I've had this paper (pdf) in my "for LW" pile for a while, I didn't want to dump it in the open thread to have it promptly drowned out but neither could I think of much more to say to make it worth a top-level post on the main LW.
Conclusion: the impact that thinking better could have on people's lives is way underestimated.
Discuss. :)