"You're carrying a sleep debt" is one of the most repeated pieces of wellness advice, and one of the least precisely defined. Most people who hear it have no idea what the actual number is, how it's calculated, or how it's supposed to change based on what they do tonight. It becomes a vague feeling of guilt rather than a number you can act on.

The problem starts with the reference point. Sleep-debt discussion almost always assumes everyone needs 8 hours. Some people genuinely need 6.5; others need 9. Subtracting last night's sleep from a population average produces a number that's wrong for almost everyone individually — either exaggerating a non-issue or missing a real one.

A useful sleep-debt number has to solve two problems at once: what do you actually need, and how has your recent history — not just last night — tracked against it. Apple Health has enough nightly sleep data to answer both, once you build the right ledger on top of it.

"Sleep debt isn't 8 hours minus last night. It's your own need, measured against your last seven nights — and a big Saturday lie-in doesn't erase a week of short ones."

Step One: Your Need Isn't 8 Hours, It's a Personal Number

The single biggest error in most sleep-debt tools is using a fixed target for everyone. Sleep need varies by genuinely large margins between individuals — the range across the healthy adult population commonly spans 6 to 9+ hours, and it's substantially heritable, not a matter of discipline.

Wrong approach
Fixed 8-hour target
Treats everyone as needing the same amount. Manufactures debt for naturally shorter sleepers and understates it for people who genuinely need more than 8.
Example: someone who's felt consistently sharp on 7 hours for years gets flagged as "in debt" every single night against an 8-hour bar that was never theirs.
Right approach
Personal need from your own history
Your need is derived from your own recent sleep durations — specifically the 75th percentile of your main-night sleep over the last 4 weeks, which approximates what you get when nothing is cutting a night short.
Example: someone whose nights over the last month cluster mostly around 7h with occasional 7h40m nights gets a personal need of ~7h40m — their own ceiling, not a population average.

Why the 75th percentile rather than the average or the maximum? The average is dragged down by the short nights that are themselves evidence of debt — using it as the target would be circular. The single longest night in a month is often an outlier (illness, a lazy Sunday). The 75th percentile sits in between: it's close to what a well-rested night for you actually looks like, without being distorted by one extreme data point.

This calculation needs a minimum amount of history to be trustworthy — at least a week of recorded main-night sleep within the last 28 nights. Below that, there isn't enough signal to distinguish your real need from noise, and a personal need shouldn't be guessed from too little data.

Step Two: The 7-Night Ledger, Not Last Night's Number

The second common mistake is treating sleep debt as a single-night comparison. One short night against your need tells you almost nothing — it might be a one-off late flight, a birthday dinner, or a normal fluctuation. Debt is a cumulative pattern, and the most useful window is the last seven nights: long enough to smooth out noise, short enough to stay relevant to how you feel right now.

📆
7-night rolling window
Each of the last 7 calendar nights is compared individually against your personal need — not averaged first, then compared once.
Per-night shortfall only
Each night contributes max(0, need − actual). A night that met or exceeded need contributes zero — never a negative number.
🚫
Missing nights are skipped
If a night has no recorded main sleep — watch not worn, data gap — it contributes nothing. Missing data is never assumed to mean zero sleep.
🧢
Capped at 8 hours
Total debt display caps at 480 minutes. Beyond that point the number stops being informative and starts just being alarming.

Why Oversleeping Doesn't Repay Debt

This is the detail that surprises people most, and it's the one that matters most for how you should actually think about "catching up" on sleep: the ledger only ever sums the shortfall. A night where you slept beyond your personal need contributes exactly zero to the total — never a negative number that offsets a previous bad night.

Night Slept Need Contribution to debt
Mon 5h 30m 7h 30m +2h 00m
Tue 6h 00m 7h 30m +1h 30m
Wed–Fri 7h 30m each 7h 30m +0m each
Sat 9h 30m (the "catch-up" night) 7h 30m +0m — not −2h
Total 3h 30m still outstanding
Why this asymmetry is correct, not a bug: Sleep researchers have found that a single long recovery night restores alertness and mood, but it doesn't fully reverse the accumulated physiological cost of a run of short nights — some of the deficit in areas like cognitive performance persists even after a big catch-up sleep. A ledger that let a 9.5-hour Saturday erase Monday and Tuesday's shortfall entirely would be more comforting and less accurate. The debt total instead keeps declining gradually as you string together nights that meet or exceed your need, which better reflects how recovery actually works.

Reading Your Number

Once you have a personal need and a 7-night ledger, the total debt figure becomes genuinely informative rather than a vague sense of tiredness:

1
Under ~1 hour
Normal week-to-week variance. Not worth restructuring anything around — most weeks will show some small debt from ordinary life.
2
1–3 hours
A noticeable pattern of short nights. Worth being deliberate about the next few bedtimes rather than letting it compound further.
3
3–6 hours
A meaningful deficit that likely correlates with reduced HRV, elevated resting heart rate, and lower recovery scores over the same window.
4
Near the 8-hour cap
A sustained shortfall across most of the week. The single highest-leverage change available is protecting the next several bedtimes, not one big weekend sleep-in.

Because debt only clears through nights that meet or beat your need, the practical response to a high number is unglamorous but effective: several ordinary, adequate nights in a row, rather than one heroic one. The ledger rewards consistency, which happens to be exactly what sleep physiology rewards too.

Where This Feeds Into the Rest of the Picture

Sleep debt doesn't exist in isolation — it's one of the inputs that helps explain movement in your other Apple Health-derived metrics. A rising debt total is often the quiet explanation behind a lower Recovery score, a faster-draining Reserve, or a suppressed HRV reading that otherwise looks unexplained.

It also becomes forward-looking. Metrya's Tonight Simulator projects what tomorrow's debt will be if you go to bed at a given time tonight, letting the ledger respond to a decision you haven't made yet rather than only reporting on decisions already made.

Ask your AI advisor: With sleep debt in context, you can ask "is my current fatigue actually sleep debt, or something else?" and get an answer that cross-references the ledger against your training load, HRV trend, and recent illness signals — rather than a generic "get more sleep."

What Sleep Debt Can't Tell You

A Number Worth Trusting

The reason a sleep-debt figure is worth paying attention to — rather than dismissing as another wellness score — comes down to these two design choices: measuring against a need that's actually yours, and treating the ledger as cumulative rather than resettable by one good night. Together they turn "I feel tired" into a specific, trackable number with a specific, unglamorous fix.

Know your real number.

Metrya calculates your personal sleep need from your own history and tracks a rolling 7-night debt ledger — computed entirely on-device from your Apple Health data.

Download Metrya Free