The Sleep Scoreboard

A recovery audit needs sleep numbers it can actually act on, and that turns out to be a short list. Three of them — how long you sleep, how regular the timing is, and how quickly you fall asleep — carry nearly all the practical signal, and all three can be logged without a wearable. This page explains why those three matter, how to measure each honestly, and what a drifting column means for the rest of the audit.

Why Only Three Numbers

A good audit metric has two properties: you can measure it accurately at home, and a change in it tells you something you would otherwise miss. Most of what a sleep tracker displays fails one or both tests. Stage percentages are not measurable accurately at home, and an overnight "score" moves for reasons nobody can name. The three numbers below pass both tests, and they capture most of the measurable relationship between sleep and health.

• ⏱️ Duration — total sleep per night, the U-curve number with roughly seven hours as the anchor for most adults.
• 📐 Regularity — the spread of your wake times across a week, the strongest individual predictor in the set.
• 🚪 Latency — how long it takes you to fall asleep, the early-warning column for insomnia risk.

Each has a clean measurement procedure, files as a single number per quarter, and means something when it drifts. Everything else on your device is either redundant or too noisy to file.

Duration: The U-Curve, Not a Target

The relationship between sleep duration and mortality is a U. Risk sits lowest around seven hours and rises toward six hours or below on one side and nine or above on the other. The shape moves the question from how much sleep to demand to where on the curve you sit.

• 🕖 The anchor — seven hours is the midpoint for most adults; six and a half to seven and a half is a reasonable band. Older adults frequently settle at six to seven and cannot force more — the curve adjusts with age.
• 📉 The short edge — regular sleep below six hours carries elevated cardiovascular and metabolic risk in cohort data; the association is consistent and dose-like.
• 📈 The long edge — habitual sleep beyond nine hours also carries higher risk, though long sleep often marks underlying illness rather than causing it, so treat that edge as a flag to investigate, not a number to avoid.
• ⚖️ Quality is not quantity — an unrefreshing seven hours is a different problem from a short seven-by-alarm week; the audit records both the number and the one-line context.

Regularity: The Load-Bearing Column

Of the three, regularity has the strongest claim to being the load-bearing number. In a large UK Biobank-derived cohort, sleep regularity predicted mortality more strongly than sleep duration — irregular sleepers carried higher all-cause and cardiovascular mortality than regular sleepers regardless of how long they slept. The mechanism is plausibly circadian: a shifting sleep-wake anchor disrupts the timing of metabolism, blood pressure, and hormone release.

• 🎯 The band — the working target is wake time within ±30 minutes of your usual wake time, across seven nights. Wake time anchors the clock more than bedtime, and it is the one you control first thing in the morning.
• 📓 The measure — log your wake time for a week and compute the spread between the latest and earliest morning. All seven inside a 60-minute window passes; two nights an hour late is the quarter's finding.
• ⌚ The wearable shortcut — this is where consumer devices genuinely help, because they are simply timestamping sleep onset and offset. The stage fiction does not matter for column two.
• 🚧 Shift work is its own rule — anchor relative to each shift block rather than one fixed hour, and log the honest context line; the audit detects drift within the schedule you actually have.

Latency: The Honest Onset

Sleep latency — the time between lights-out and sleep — is the least glamorous column and the one most worth tracking. For good sleepers it is typically in the range of ten to twenty minutes, and it rises with age. Latency is also the early-warning sign for insomnia: when falling asleep takes longer and longer, night after night, the pattern of lying awake in bed quietly entrenches, which is how chronic insomnia starts.

• ⏰ The honest measure — you cannot time this objectively without a device or lab, but a self-estimate to the nearest five minutes is workable for trend purposes. The point is the trend, not the precision.
• 🎯 The band — up to about 20 minutes is unremarkable; 20 to 30 is worth watching; consistently over 30 minutes on most nights is the flag that belongs in the red column.
• 🚩 The trap — a long latency combined with a short duration is the classic insomnia signature. If you are also waking to use the bathroom, worrying in bed, or dreading bedtime, the flag is more than observational — see the When Sleep Won't Come protocol.
• 🧪 The lab alternative — a false sense of lying awake for hours when the log says otherwise is worth knowing about; if your perceived latency and your partner's observation differ wildly, that is a conversation for a clinician, not a tracker forum.

How to Log the Three Without a Lab

One week of logging per quarter is enough. You are filing a trend, not running a sleep study — three columns, one line of context, done.

Reading the Scoreboard

The quarterly read is a traffic-light summary of the three columns, not a grade. Each column gets one verdict and one planned move.

• 🟢 All three steady for two quarters — the sleep layer is holding; recheck in three months.
• 🟡 One column drifting — pick the column with the clearest lever. Regularity responds to the alarm clock and morning light; duration to the back end of the evening (caffeine, alcohol, screens); latency to stimulus control — out of bed if sleep does not come.
• 🔴 Two or more columns red — coordinated failure, not three separate ones. Fix wake time first (it drags duration and latency along), then read the Seven Habits of Great Sleepers topic before reaching for anything stronger.
• 📅 The honest calendar — one bad week of sleep is a data point; two red quarters in a row is the finding. The quarterly format exists precisely so you do not overreact to a single hard week.

When the Numbers Point Past the Scoreboard

Three situations send the sleep audit outside the scoreboard, and all three need a clinician, not a better log. The audit is a monitoring tool; it is not a diagnostic instrument, and these columns cannot rule sleep disorders in or out.

• 🌬️ Breathing pauses, gasping, or choking at night — especially with loud snoring and daytime sleepiness despite full hours, these point toward obstructive sleep apnea and deserve clinical evaluation and possibly a sleep study. The Sleep Apnea topic covers the warning signs in detail.
• 🛏️ Chronic insomnia taking shape — latency over 30 minutes most nights, for more than three months, with daytime consequences, is chronic insomnia; the first-line treatment is a structured program (cognitive behavioral therapy for insomnia), so the earlier you hand off, the less entrenched the pattern becomes.
• 💤 Unrefreshing sleep at full duration — a solid seven hours with zero recovery is not "bad sleep hygiene"; it is a symptom that deserves a medical conversation about possible restless legs, apnea, hormonal factors, or medication effects. Do not treat it with more supplements.

None of these columns should be changed with lifestyle advice alone once they show this profile. The scoreboard's real value is catching the drift early enough that the fix is still a habit change.

Questions, Answered Briefly

• ❓ Do naps count toward duration? — No. Log naps separately in the context line; the evidence is about consolidated night sleep, and a nap-heavy week usually signals short nights.
• ❓ My device says I average 6 hours but I feel fine — which wins? — The device wins on the number; your daytime function and the scoreboard shape the verdict. Feeling fine at six hours is real information, but a regular six is still worth watching.
• ❓ Can I fix latency with supplements? — Melatonin helps some people with timing and onset, but it is not a treatment for the escalation pattern here; if latency keeps climbing, the play is stimulus control and a clinician conversation, not a nightly capsule.

Related Topics