Exercise to stay healthy seems a pretty interesting meme of the last 50 years.

You make an assumption that more is always better with doesn't seem to be the case. Recent studies suggest that exercising for 5 hours per week is actually worse than doing 3 hours.

We don't really have a good theory of how exercise provides health benefits.

Ryan Holiday recently wrote:

A winner doesn’t “exercise,” they train in something (martial arts, running, swimming, biking, cross fit, boxing, weights, whatever).

Personally the more I think about the human body the more I get convinced that while the current frame is obviously better than engaging in no physical activity at all, it's still the dark ages of knowledge.

I think it will be very interesting to see how people in 100 years will speak about how people today go jogging on concrete or go to a gym and lift weights in fixed positions and most of the people haven't shown how to do the movement in an ergonomic way.

(personal prediction that 30minutes of exercise will increase my life expectancy by 4 years)

If you assume that you have, oh, 3 hours a day free time, living an extra 4 years means that you have an extra 1/2 year of free time.

As exercise is taken from your free time, this is only a good deal if 30 minutes a day for the rest of your life totals to less than 1/2 year. 30 minutes is 1/48 of a day, so 30 minutes over a year is 1/48 of a year. This means that if you expect to live any longer than (1/2) / (1/48) = 24 years, you should not exercise.

Of course, other factors come into play (if your life expectancy increases, your extra years may be spent retired, on the other hand, you're more likely to be ill in those extra years, etc.)

And the less free time you have, the less exercise helps, because you gain less (each year of additional lifespan provides you with less extra free time) but you lose just as much (since your exercise comes out of your free time).

There are opportunities to exercise at no cost in time. I usually bike to work and to go into town, which takes no longer than driving, and the harder I ride, the less time it takes. I walk up four stories to my office, which takes no longer than the lifts, and I'm usually up and down those stairs 2 or 3 times in a day. I play taiko, which is great fun, but is also pretty strenuous. So that's 7 miles of biking and around 100 vertical feet of stair-climbing every working day, and a cardiac workout of a few hours a week, for free.

There are other things I do, but these are the ones that effectively take neither time nor money to do.

There is something that makes me skeptical about "exercise" considered as an abstract quantity. There are many different ways of doing "exercising", and they have many different effects on the body. How much exercise you do (but then, counting in hours ? spent calories ?) do matter, but in which conditions also matters a lot : how regularly, what kind of exercise, ...

Exercising usually is good for urban sedentary people who "naturally" don't do much of it, but it can also have negative side-effects, in accelerated aging or articulation damage. Professional sport players have a very short average lifespan, and it's not only because some of them use drugs.

There is surprisingly few literature on the topic (or I couldn't find it), but from what I remember finding, something important is the regularity (doing 3 hours of exercise twice a week is not nearly as good as doing 1/2 hour each day, even if 1/2 hour each day is just 3.5 hours in total) and the kind of activity : constant medium/low intensity, long duration activity (walking, long-distance running, bicycling or swimming at a moderate speed for a long while, ...) is much better than variable intensity (like ball games, where you run and stop and run again and kick and ...) or high-intensity short-duration (like sprinting).

As for opportunity cost, something to consider is doing exercise that fits a purpose, instead of just "wasting" hours doing it, for example, running/walking/bicycling instead of using a car means you do some exercise but also save on money and on accident risks, especially if your environment (public transport network, proximity of things, ...) allows to not own a car at all (making the money gain very significant), or finding an activity that you actually enjoy so the time isn't fully wasted.

I have not done the research, but I suspect that this open-loop calculation is not really worthwhile. People are very different, and if you want to achieve certain benefits you ought to close the loop and measure these benefits directly, then adjust the amount and type of exercise accordingly.

And finally; the disadvantage: opportunity cost

there are 168 hours in a week. With most people spending 1/3 of that asleep (56hrs, 8hrs/night), 20 hours on lesswrong per week, 40hours in an average work week, before we take two hours out of each day to spend exercising (14hours); what are we taking those hours away from? Can you do what you were doing before without the time spent exercising here?

Note that exercising affects your need to sleep, so more minutes spent exercising don't necessarily translate to fewer minutes available for other stuff on a one-to-one basis.

But there are two contrasting effects here -- by exercising, you'll need to sleep more, but you'll also get more hours of actual sleep per hour spent in bed trying to sleep -- and it's not obvious which one will dominate. It probably depends on lots of things.