Search "life score" and you'll find half a dozen apps using the exact phrase, and no two of them mean the same thing by it. One version is a weekly nutrition score out of 150. Another combines study time and workouts into a number you compete with friends over. A third builds it from location data and a wellbeing questionnaire. The name is common. The actual math behind it almost never gets explained. That's worth fixing, because the design choices buried in that math determine whether the number tells you something true or just something simple.
The basic idea, stripped of branding
At its core, a composite score like this takes several separate inputs (commonly: sleep, physical activity, and some measure of mental/stress state) and reduces them to one number, usually on a 0-100 or similar scale, meant to represent "how are you doing overall" at a glance. The appeal is obvious — checking one number is faster than checking four dashboards.
The math to get from "several inputs" to "one number" isn't as simple as it looks, and the specific choice made has real consequences for what the number actually means.
Choice one: averaging vs. gating
The most common approach is a weighted average — each input gets a percentage weight, and the final score is roughly (sleep × weight) + (training × weight) + (other × weight). This is easy to compute and easy to explain, and it has a specific, often-overlooked property: every input is treated as substitutable for every other. A terrible sleep score can be mathematically offset by a great training score, because addition doesn't care where the points come from.
The alternative is a gating or "floor" model, where one input (often sleep, given the research on how thoroughly it affects the others — see the mechanism section below) caps how high the total can go, regardless of how strong the other inputs are. This is mathematically more complex — it's not just addition — but it more closely matches how the body actually seems to work: research on sleep and training recovery (covered in more detail elsewhere) consistently finds that poor sleep reduces how much benefit a workout actually delivers, meaning a "great training day" on top of "terrible sleep" plausibly didn't deliver a great training day's worth of adaptation. An averaging model can't represent that interaction at all; a gating model is built specifically to.
Choice two: self-report vs. sensor-derived
Some composite scores (Whoop's Recovery Score, Oura's Readiness Score) are built primarily from biometric sensor data — heart rate variability, resting heart rate, skin temperature. Others are built from self-reported behavior — did you sleep, did you train, did you meditate, logged by hand. These aren't measuring identical things: sensor data captures autonomic nervous system state with reasonable objectivity but requires hardware; self-report captures consistency of behavior with reasonable honesty-dependency but requires no hardware at all. Research comparing the two has found they correlate — modestly, and more strongly for subjective states like mood than physical ones like soreness — which suggests they're each catching a real but different slice of the picture, not that one is simply a lesser version of the other.
Why the label "Life Score" tells you almost nothing on its own
Given how differently the apps using this exact phrase build their number — nutrition-weighted average, sensor-based recovery estimate, self-reported behavioral composite, wellbeing questionnaire — the label itself carries almost no information about what you're actually looking at. The only way to evaluate any specific "life score" is to ask the two questions above: is it averaging or gating, and is it built from sensors or self-report? Those two choices determine what the number can and can't tell you far more than the name on the label does.
Why this matters for how you should read any composite score
If you're looking at a single wellness number from any app — this one included — the useful question isn't "is my score good or bad today." It's "what specific math produced this number, and does that math match how the underlying physiology actually behaves." A gating model that treats sleep as a limiting factor is making a specific, falsifiable claim about physiology (that sleep quality bounds what training and other inputs can deliver) — a claim with real research behind it, discussed at length elsewhere, rather than an arbitrary design choice.
Key takeaways
- "Life score" is used by multiple unrelated apps with genuinely different underlying calculations — the name alone tells you nothing about the method.
- Averaging models let strong inputs offset weak ones; gating models let one input (typically sleep) cap the total, which better matches documented sleep-training interaction research.
- Sensor-derived scores (HRV-based) and self-report-derived scores measure related but distinct things, correlating modestly rather than perfectly.
- Judge any composite health score by its method, not its label.
Ready to start your streak?
High performance. Zero equipment. No excuses.