Why we tell you what we cannot prove
Your model is a map of what your own data has actually verified - not a pile of confident-sounding claims. It shows associations, never proven causes; it starts every link assumed-zero and makes it earn its place; and when it cannot prove something, it says so. That restraint is the product, not a limitation of it.
- Your model shows associations in your own data - links, not proven causes. We removed the word "confirmed" for exactly that reason.
- Every link starts assumed-zero and has to earn its way up, so one lucky week cannot fake a pattern.
- At realistic amounts of data, most personal effects can never be confirmed - and that honest "not yet" is the correct answer.
- The only honest way to prove something helps you is a small randomized experiment on yourself.
A map of what your data actually verified
Most apps in this space are built to sound sure. They will tell you that blue light wrecked your sleep, that your 3pm slump is your cortisol, that a cold shower fixed your focus. Confident, clean, and usually unearned.
Sarenica is built the other way. Your model is a living map assembled only from analyses run on your own data - and each link is labelled honestly as an association, not a cause. We deliberately removed the word "confirmed" from the product, because a confirmed correlation and a proven cause are not the same thing, and pretending otherwise is how these apps quietly mislead people.
“The restraint is the product, not a limitation of it.”
Every link starts at zero
Under the hood, each relationship in your model is a belief, not a headline. It starts from a skeptical position - assumed to be zero, no effect - and it only moves off zero as your own evidence accumulates. A single good-looking week is not enough to shift it.
When two of your own findings disagree, the model does not quietly average them into a tidy answer; it marks the tension and asks to test it. When a finding gets old and is not re-confirmed, its weight fades - the model is allowed to forget what it can no longer stand behind.
Most patterns can never be confirmed - and that is correct
Here is the uncomfortable truth almost nobody in consumer health will say out loud: with a normal amount of personal data, you simply cannot confirm most effects, no matter how good the app is. To call a personal link real, the effect has to be both big enough and backed by enough days. Small effects and short histories cannot clear that bar - not because the model failed, but because the evidence genuinely is not there.
The chart below is the honest ceiling. It shows the smallest effect your data could even detect, as your history grows. At two months, only a strong personal effect can ever be confirmed; most real-but-modest patterns stay "suggestive" indefinitely. So instead of hiding that, we are moving toward showing it to you up front - before you reorganise your life around a number that was never provable.
The only honest route to "this helps you"
Watching your data can show that two things move together. It can never, on its own, tell you which one caused the other - or whether a third thing drove both. A shorter gap before your first phone pickup might track lower recovery, but the arrow could point either way, and a bad night could be causing both.
The honest way to settle that for you specifically is a small randomized experiment on yourself: some days the model nudges a change, some days it does not, decided by a coin flip, and it measures what follows. That is the difference between "these go together in your data" and "this actually helps you" - and it is the direction the product is built to grow into.
What you get instead of false certainty
You get a map you can trust: the handful of links your own data genuinely supports, shown boldly; the many that are only suggestive, shown faintly; and the honest gaps, shown as open questions with a one-click way to test them.
It is a quieter product than the ones promising to decode you in a week. But every line in it is one your data earned - and when it does not know, it tells you. For something making claims about your body, that honesty is not a nice-to-have. It is the whole point.
