I'm happy to report that scientists have been seriously testing if stuff actually works for over a century (Slonaker, 1912).
They typically try an intervention, like diet or exercise, on lab animals, and then see how long they live.
For what it's worth, I maintain the world's biggest spread sheet of these life span experiments.
It summarizes over 14,000.
One column is the intervention (diet, exercise, etc..) another is the change in lifespan (+10%, -2%, etc...) another is the species (human, mice, etc...) and so on.
I suppose it may suggest increasing code complexity and effort to maintain.
For what it's worth, I also tried to quantify how robust Debian is my measuring the ages of packages installed on my main computer.
My results are at
http://morse.kiwi.nz/kingsley/lib/exe/detail.php?id=technolo...