Free U.S. shipping on orders $45+ · Subscribe & save
Sleep Debt Calculator
Fourteen nights in, a debt in hours and a repayment range out. It runs in your browser, it stores nothing, and every figure under it names the study it came from.
Use the figure you know you need, not a round number. Kitamura 2016 measured an average of 8.4 hours in a lab that let people sleep as long as they wanted, and the spread between people was wide.
What that adds up to
| Nights below target | 10 of 14 |
|---|---|
| Hours short, added up | 12 h |
| Hours banked on the long nights | 1 h |
| Total time asleep | 101 h |
The net figure subtracts the banked hours from the hours short, which is the most generous reading of your fortnight. Leger 2020 (n=12,637 adults) is the reason not to trust it too far: of the people carrying a severe debt, only about one in four balanced it out through weekend catch-up or naps, so for most of them the number that described their fortnight was the raw shortfall of 12 h, not the net.
The repayment range is division, not biology. Kitamura 2016 is the only study that watched the repayment happen: about one hour of accumulated debt took roughly four days of a twelve hour sleep opportunity before sleep length settled. Expect recovery to run behind the arithmetic, and read the range as a floor.
What this calculator does
It does one piece of arithmetic and refuses to do several others. You give it fourteen nights of sleep and the number of hours you personally need. It returns the hours you fell short, the hours you banked on your longer nights, the difference between the two, and a repayment range worked out by dividing that difference by one to one and a half extra hours a night. Nothing is fitted, modelled or personalised, because nothing on a web page can be. Two weeks is the window because two weeks is the span the strongest evidence covers.
Why the feeling is not the measure
Van Dongen 2003 (n=48 healthy adults, aged 21 to 38) randomised people to 4, 6 or 8 hours in bed for fourteen consecutive nights, alongside a separate arm kept awake for three nights straight. The restricted groups got steadily worse at reaction time and working memory, with the deficits still accumulating on day fourteen rather than levelling off. Restriction to six hours or less produced deficits the paper describes as equivalent to up to two nights of total sleep deprivation.
The part that matters for a calculator is what those people said about themselves. Their subjective sleepiness ratings rose sharply when the restriction began and then barely moved, and they did not reliably separate the six hour group from the four hour group. People whose measured performance was still sliding in the second week reported feeling much as they had in the first. That gap between how impaired people were and how impaired they felt is the entire reason to count hours on a page rather than ask yourself how tired you are.
Where the repayment estimate comes from
Kitamura 2016 (n=15 healthy men, mean age 23) gave people a twelve hour sleep opportunity for nine consecutive nights and measured how long they actually slept. Sleep length started high, fell night by night, and settled around the fourth day. Habitual sleep at home averaged 7.37 hours, the settled laboratory figure averaged 8.41 hours, and the roughly 1.04 hour gap between them is what the authors called potential sleep debt. Their finding in plain terms: about one hour of debt took roughly four days of an unusually generous sleep opportunity before sleep length settled at its own level.
That is why the repayment number here is a range and why the tool calls it a floor rather than a schedule. Fifteen young men, a laboratory, twelve hours in bed every night. It gives you the direction and the rough order of magnitude. It says nothing about what ten or fifteen hours of accumulated debt takes to clear, because nobody has measured that.
Why Saturday morning does not settle it
Leger 2020 surveyed a nationally representative sample of 12,637 French adults about sleep debt, naps and weekend catch-up. Among those carrying a severe debt, defined there as more than ninety minutes, 18.2 percent balanced it out with weekend catch-up sleep and 7.4 percent with naps. The remaining three quarters did neither. That is an observation about what people actually do rather than proof that catch-up sleep cannot work, but it is the reason this tool shows your raw shortfall next to the net figure instead of quietly letting two long weekend nights cancel ten short weekday ones.
What it cannot tell you
It cannot tell you why you slept badly, and it does not try. It sees duration and not quality, and the two are different inputs: eight hours broken into fragments is not eight continuous hours, and this page cannot see the difference. It cannot tell you whether your target is the right one, since the spread between individuals in Kitamura was wide enough that a single recommended number does not fit everyone. It does not know your age, your shift pattern, your medication, your alcohol, whether you are ill, or whether there is a new baby in the house, all of which change both what you need and what you get. It diagnoses nothing. What it can do is turn a vague sense of being behind into a specific number of hours, which is a better thing to take to a conversation than a feeling.
Caffeine is not repayment
Caffeine borrows alertness against the same debt rather than paying any of it back, and Van Dongen is the reason to be careful with it: the measured deficit kept growing while the felt one stayed flat, and a stimulant flattens the felt one further. If you do use caffeine, the timing matters more than the amount. Gardiner 2023, a systematic review of 24 studies, found that a 107 mg coffee needs about 8.8 hours before bed to avoid cutting total sleep time. That cutoff is about the caffeine, not the format, so it applies to a Roon pouch exactly as it does to coffee. A pouch is 80 mg, below the 107 mg the review modelled, and the review did not publish a figure at that dose.
None of this is medical advice and none of it is a diagnosis. If your sleep has looked like this for more than two weeks, it is worth raising with a clinician rather than a calculator.