This is a static site. Nothing is sent anywhere.
Each input maps to a hazard ratio (HR) from a published study, e.g. meeting exercise guidelines corresponds to HR 0.69, a 31% lower mortality hazard. HRs are combined multiplicatively, assuming independence. Sometimes the numbers differ from the study, as we try to get our reference points (HR = 1) to be the average value for a given input. For example, 20 BMI = 1.0 HR, and 30 BMI = 1.07, but we instead divide and get 20 BMI = 0.93 and 30 BMI = 1.0 as 30 BMI is closer to the average. We cap the total at a 55% hazard reduction (HR 0.45) and a 4× increase (HR 4.0)
HR → years of life expectancy uses a Gompertz approximation (adult mortality risk doubles roughly every 7 years).
In an attempt to be more accurate, some math is applied. Exactly how can be seen in the source code.
A lot of mistakes are likely being made in the calculation, so don't take the final result as accurate. There's NO scientific basis for the math we do in this site. Take nothing on this site as truth uncritically, and please read the attached sources.
Correlation does not equal causation. Some of these sources just check for a variable and output a number. Exercisers generally sleep better, eat better and move more daily, so how can the study be sure, when measuring exercise, whether it's the exercise, sleep, diet or daily movement that's increasing your life expectancy? As such, multiplying the -20% mortality from both the exercise, sleep, diet and daily movement studies as if they were independent likely conflates and overstates benefit.
Furthermore, some of these studies have a lot more caveats than we can fit into a slider. For example, creatine's cognitive effects are most prevalent in vegetarians and older adults. The real number that that applies to *you* specifically is very hard to find. Even more so when some studies do account for confounding variables, like one of our sun exposure studies accounting for the fact that exercisers spend more time in the sun. If you'd like to help, feel free to reach out on the GitHub
Similarly, cognition and happiness are deliberately very fuzzy, as there's no real measure of "brain score".
The reference attempts to be the "average citizen", meaning better than the reference is only better than average, not necessarily "good".
The goal of the website is only to give a collected overview of insights into what kind of lifestyle factors affect your health and by how much, whilst maybe sharing some new information. Is a daily multivitamin 'healthy' enough to offset never exercising? (no).