Longevity protocols: why a ten-part stack cannot tell you anything

The arithmetic problem

Suppose you adopt ten changes on the same Monday and feel better a month later. What have you learned? That the bundle, applied by you, in that month, coincided with feeling better. You cannot attribute the improvement to any component, you cannot rule out that one item is actively harming you while the others compensate, and you cannot separate any of it from the season, your workload, the novelty, or the simple fact that people who begin protocols usually also start sleeping and eating more deliberately. Ten changes produce one bit of information. One change, run properly, produces one bit about something specific — which is strictly more useful.

The part everyone skips: what you were before

A protocol adopted without a baseline cannot be evaluated even in principle, because the comparison does not exist. This is the least glamorous requirement and the one that decides whether the whole exercise means anything: several weeks of ordinary data, gathered before you change anything, in a period you would describe as typical. Not your worst month, not a training camp, not December. If you are about to start something, the useful first move is to wait three weeks and record what normal looks like.

The unglamorous list has the evidence

It is worth saying plainly what the strongest evidence actually supports, because it is not what the protocols emphasise: not smoking, keeping blood pressure controlled, regular aerobic and resistance training, adequate sleep, moderate or no alcohol, a diet you can sustain, and maintained social connection. The effect sizes there dwarf anything demonstrated for any supplement on any stack, and they are boring precisely because they are settled. A protocol that adds eleven items on top of an unaddressed sleep problem or an untreated blood pressure has optimised the wrong end of the list.

Testing one thing, properly

Pick a single change and a single outcome, and write down before you start what result would count as a success — because deciding afterwards is how everything becomes a success. Run it for at least four weeks, because most things that work take that long to show and most noise resolves over that window. Hold the obvious confounders roughly steady: training, alcohol, caffeine, travel. Compare rolling averages rather than days. And expect a small effect, because almost everything real is small — which is also why a protocol of ten small effects is indistinguishable from a protocol of ten nothings.

What your own data can and cannot answer

Wearable data is genuinely good at within-person comparison: it can tell you whether your sleep, resting heart rate, HRV baseline or temperature behaved differently across two four-week blocks, which is exactly the question a self-test asks. It cannot tell you anything about your lifespan, and the “biological age” numbers on consumer devices are constructed from a handful of ordinary metrics rather than validated against any outcome — a lower one is not evidence that a protocol is working. Vitra is deliberately built for the first job and not the second: it compares you against your own history, on your own machine, and does not score you against a population. This is general information, not medical advice — anything involving medication or an existing condition belongs with your doctor.

Frequently asked questions