"I'm catching up on sleep"

You are, partly. It's worth being accurate here because the usual version of this argument overreaches, and once you've seen it overreach you stop trusting the rest.

The shortfall is real. Hans Van Dongen and colleagues kept forty-eight healthy adults in a laboratory for two weeks on four, six or eight hours in bed a night, and the people on six hours, which is a very ordinary weekday amount, built up deficits in attention and reaction time across the fortnight that by the end matched two nights with no sleep at all. The striking part was that they didn't feel it. Their ratings of their own sleepiness rose in the first few days and then more or less levelled off, while their performance kept falling. So if you feel roughly fine on Thursday, that's not evidence you're fine.

And extra weekend sleep does put some of it back. Christopher Depner, Kenneth Wright and colleagues put healthy young adults through five short nights, then two nights of sleeping as long as they liked, then two more short nights, against a group that was short all nine nights. Over the weekend the catch-up group slept about an hour and ten minutes more than their baseline, and their late-evening eating, which sleep loss had pushed up, came back down. So far so good, and that's the part you feel on Sunday.

Then the week started again. In the short nights after the weekend, the catch-up group's body clocks had shifted later, their evening eating and weight went back up, and their insulin sensitivity fell by somewhere between nine and twenty-seven per cent depending on the measure, which was no better and by some measures worse than the group that never got a weekend at all. The authors' conclusion was that weekend recovery sleep is not an effective way to prevent the metabolic effects of a repeatedly short week. This was a small study of healthy young people and the numbers matter less than the shape. The weekend put back an hour. It didn't put back the week, and it moved the clock.

Shingo Kitamura and colleagues found something related from a different direction. They estimated each person's optimal sleep length by letting them sleep as long as they wanted for several days in a laboratory, worked out how far short of it their normal life fell, and found that even one hour of that shortfall took about four days of extended sleep to recover from. Two mornings can't do the job of four days. So yes, you're catching up. You're catching up on some of it, by a method that makes next week's shortfall bigger.