BEHIND THE SCENES

How a session is scored

The number out of a hundred is a weighted average of several separate measurements. How much weight each carries, why, and what the score doesn't tell you.

7 min read

The number on screen when a session ends doesn't measure one thing. There are several separate measurements, each with its own weight, and on top of that a multiplier that affects them all.

The measurements

Rhythm match. The gap between the moment the circle starts shrinking and the moment your exhale starts. Measured separately for each cycle, then averaged. This is the heaviest component, because it is the app's real job: settling your breath into a pace.

If your exhales start late or early by roughly the same amount every cycle, someone who is consistently late doesn't lose as many points as someone who doesn't keep the rhythm. The shift itself still costs points: next to nothing when it's small, more and more as it grows. When the shift costs points, the report says so under Measurements and in the suggestions.

Exhales heard. How many of your target exhales were heard, and how many of those were genuinely near the target. Both are measured, because looking at the first alone would mislead: someone breathing out at random intervals would "catch" every target. So an average that rewards both at once is taken: if one is low, the result comes out low.

Exhale length. How close the exhale is to the target length. Against a three-second target, breathing out for two seconds counts as the same amount of deviation as five.

Inhale and inhale length. The inhale is far quieter than the exhale, so it is measured separately with lower thresholds. Not hearing the inhale at all doesn't sink the score by itself: some people's nasal inhale really is silent.

Breath volume. How far the exhale rises above the noise floor. Both too quiet and too loud are penalised. Too loud isn't a good sign: it means forcing the breath, pushing instead of letting go.

Exhale smoothness. How much the sound level varies through the exhale. A steady "ffff" gives a small deviation; a broken or rasping exhale gives a large one.

Cycle consistency. How much the cycles resemble one another. One clean cycle among scattered ones brings this down.

Holds and hold length. In techniques with a hold phase, whether a sound was heard in that window — if it was, air escaped — and whether the hold lasted as long as it should.

Completion. How much of the session you did.

Which measurements count depends on the technique: a technique without holds has no hold components at all, and if the inhale could not be measured the inhale components drop out. In Kapalabhati the exhales are too fast to separate one by one, so rhythm match and exhales heard don't count towards the score. The weights of whatever remains are shared out again among themselves.

In the three techniques where sound is part of the technique, the technique's own sound is a separate component: the hum in Humming Bee Breath, the rustle of the exhale in Ocean Breath, the hiss of the inhale in Cooling Breath. In these three techniques a good part of the score comes from that sound. Even with no sound at all, rhythm, exhales and inhale earn their own share, but only a session that gets the sound too reaches full marks.

The weights and the rhythm multiplier

The weights are fixed: rhythm match takes the largest share, followed by exhales heard and exhale length, with the inhale-related components carrying the smallest share.

On top of the weighted average there is one more multiplier. If most of the exhales fell outside the target, the whole score is lowered proportionally: a session that never caught the rhythm can get at most half marks even if its other measurements are good. The reason is simple — a well-formed breath outside the circle is a well-formed breath, but not that session's breath.

Part scores in long sessions

In sessions longer than two minutes, the report also shows part scores. Each part is scored with the same measurements, the same weights and the same multiplier, looking only at that part's exhales. The score at the top isn't their average: it is worked out from the whole session, as described above.

A near-flawless session gets a hundred

A session that is very nearly flawless shows a hundred on the screen. The reason is presentation: writing ninety-six makes a session that was, in fact, close to flawless look smaller than it is.

There is one exception: if one of the components is clearly low, no rounding happens and the real score is written. A report that says "your exhales are short" doesn't show a hundred at the top.

What the score doesn't tell you

This number is not a measure of health. It doesn't measure whether your breathing is "correct", whether the exercise did you good, or whether you relaxed. The only thing it measures is how closely you followed the pace on screen.

It is possible to get a low score and feel very good. It is possible to get a high score and feel nothing. Both are normal — the number is a feedback tool, not a grade.

There is also the microphone's limit: the app measures the sound of your breath, not the breath itself. Someone who breathes out very quietly, sits far from the microphone or is in a noisy room can score lower than what they actually did. The technical measurements in the report exist so you can see this: the noise floor and the exhale peak level are written there.

← Behind the scenes