Bryan Johnson cancelled training after bad sleep. He was probably right, and his statistic was wrong.
At the end of August, Bryan Johnson posted that a café had given him caffeinated coffee instead of decaf, that his sleep was wrecked as a result, and that he had cancelled his morning training because acute sleep disruption can raise injury risk by 70 to 130%.
Seven million views and seventeen hundred replies later, the internet had decided he was ridiculous.
I think he probably made the right call. I also think his number doesn’t say what he thinks it says, and that the reason he was probably right has nothing to do with it.
That word “probably” is doing real work, and I’ll come back to it, because it’s the whole point.
First, the number
I went looking for the 70 to 130% range, and you can find where it plausibly comes from.
The lower end matches Milewski and colleagues, who followed 112 adolescent athletes and found that those averaging under eight hours of sleep had 1.7 times the injury risk of those getting more. A 70% increase.
The upper end matches a study of US Army Special Operations Forces soldiers, where those habitually sleeping four hours or less were 2.35 times more likely to sustain a musculoskeletal injury than those sleeping eight or more. A 135% increase.
So the range is real. It just isn’t about what happened to Bryan Johnson.
Every one of those figures describes habitual sleep. Milewski measured averages across a school year. The Army study measured how much soldiers routinely slept. A meta-analysis of adolescent athletes found chronically poor sleepers had higher injury odds, and even that pooled only three studies with one athletic population between them.
The literature on acute sleep loss and injury, one bad night followed by one training session, is thin to the point of inconclusive. Nobody has established that a single disrupted night raises your injury risk by 70%, let alone 130%.
A man who has built a public identity on measurement should be more careful about which population a statistic came from. If you take one thing from this article, take that, because the same error is everywhere and it’s usually pointing at something you’re being sold.
Now, the decision
Here’s the part that gets lost: he didn’t need the number. Cancelling that session was defensible on evidence he didn’t cite.
We know a good deal about what one bad night does, and none of it is about tendons.
A single night of deprivation produces roughly 60% greater amygdala reactivity to negative images, alongside weakened connectivity with the prefrontal cortex that normally regulates it. The emotional accelerator gets more sensitive while the brake gets less responsive.
Sleep loss also degrades executive function specifically. Reviews of the decision-making literature describe impaired judgement, increased acceptance of risk, and a shift toward short-term rather than long-term consequences.
Read those together and think about what heavy training actually requires. Accurate self-assessment of what you can lift. Willingness to stop when something feels wrong. Coordination. Patience with a technical movement that isn’t working. Sleep loss degrades every one of those, and does it while making you more inclined to accept risk and less able to notice you’re doing it.
The reason to skip a hard session after a broken night isn’t that your connective tissue became fragile overnight. It’s that the person deciding how hard to push is temporarily worse at that judgement, and unusually confident about it.
Why this matters at work more than in the gym
Most people reading this won’t be doing dunk training tomorrow. The same impairment shows up somewhere far more consequential.
Christopher Barnes and colleagues found in research written up by Harvard Business Review that sleep-deprived leaders were rated as less charismatic, apparently because they were less able to regulate their own mood. The same work found sleep-deprived employees were harder to inspire in the first place. A tired leader talking to a tired team is a compounding problem.
There’s a useful old idea, often attributed to Viktor Frankl, that between stimulus and response there is a space, and in that space lies our power to choose.
That space is where judgement lives. Sleep deprivation narrows it.
The tired executive who makes the irreversible decision, the tired manager who fires back an email they’d have softened, the tired parent who snaps at a child over nothing: same mechanism, higher stakes than a missed workout, and no one posts about it.
The instrument he didn’t use
Here’s the irony, and it’s the most useful thing in this story.
Johnson reached for a population statistic to justify a decision about one specific body on one specific morning. But there’s a considerable literature on how athletes should make exactly this call, and it points somewhere else entirely.
A systematic review in the British Journal of Sports Medicine compared subjective self-reported measures of athlete wellbeing against objective ones: resting heart rate, blood markers, oxygen consumption, heart rate response to exercise. The finding is in the title. Subjective measures trumped the objective ones. They were more sensitive and more consistent at detecting how an athlete was actually responding to training load.
Ask someone how well they slept, how sore they are, how much they fancy training, and you get a better readiness signal than most of the instruments you could strap to them.
Which means the man who measures everything had the better instrument available and didn’t mention it. Not a figure from a study of soldiers. A straight question: given how I actually feel this morning, is today a day for heavy training?
That’s why I said probably. I don’t know whether he was right, and neither does anyone in his replies, because none of us know how he felt. If he woke up wired and jittery from unexpected caffeine but physically fine, the session might have been perfectly safe. If he felt wrecked, cancelling was obvious and needed no citation at all.
The statistic can’t answer that. Only he can.
I want to be careful here, because this is where people take the wrong lesson and start waving away evidence in favour of vibes.
None of this is an argument against data. I build measurement apps. It’s an argument against substituting someone else’s data for your own information.
The three inputs do different jobs. Population research tells you which questions are worth asking, and it’s why you know to consider sleep at all. Your own history and trends tell you what has changed relative to your normal. And the subjective check tells you where you actually are this morning, which is the only one of the three that knows about the caffeine, the argument last night, or the fact that your shoulder has felt odd since Tuesday.
A go or no-go decision needs all three. Johnson reached for one, and it happened to be the one that knows least about him.
That’s the bit I’d want people to take from the whole episode. Research doesn’t make the decision. It tells you what to look at. You still have to look.
When the number changes how you feel
There’s a failure mode here that I see constantly in coaching, and it’s the reason I care about the order you do things in.
You wake up feeling fine. You pick up your phone. Your sleep score is 61. And within about ninety seconds you feel tired, foggy and slightly cheated, in a way you demonstrably were not two minutes earlier.
That isn’t imagination, and there’s a neat experiment showing it.
Researchers gave 164 people fake feedback about their previous night’s sleep, using an invented technique they claimed could measure REM retrospectively. Participants were randomly told they’d spent either 28.7% or 16.2% of the night in REM, regardless of how they’d actually slept.
Their cognitive performance then followed the fake number. Told they’d slept badly, they did worse on tests. Told they’d slept well, they did better. The number was meaningless and it still moved the outcome.
The clinical version has a name. Baron and colleagues coined orthosomnia for patients arriving at sleep clinics with tracker data, whose pursuit of perfect sleep numbers had itself produced insomnia. The detail that stays with me is their observation that these patients were believing the device over their own experience of how they had slept, and in some cases over their doctor.
So data has a specific hazard that a coffee or a late night doesn’t. It can manufacture the symptom it claims to be reporting.
The practical rule I use, and coach, is simply about sequence. Check in with yourself first. How do I actually feel, before I have been told anything? Then look at the number, and hold it as information rather than a verdict.
If you feel fine and the score is poor, you are allowed to feel fine. The score is an estimate produced by a wrist-worn accelerometer making inferences about your brain. Your own experience of your body is not the less reliable instrument in that pairing.
That’s a large part of why my own apps don’t show you a score out of a hundred in the morning. A number like that is very good at telling you how to feel, and I would rather you decided that for yourself.
What I do now, and what I did before
I’m a certified sleep and stress coach, and I still sleep badly sometimes. The difference is what happens next.
After a bad night I defer anything irreversible. No big decisions, no difficult conversations, nothing I’d struggle to walk back. I take the mechanical work over the creative work, because tired me is adequate at admin and unreliable at judgement. And I modify training rather than skipping it: I’ll walk, or lift light and technical, and leave anything near a limit for a day when I’m equipped to notice a warning sign.
None of that is discipline. It’s a set of rules I follow precisely because I don’t trust my in-the-moment assessment on those days.
Before I understood any of this, I did the opposite. I made decisions I later unpicked. I was short with people I love, over nothing, and then puzzled about why. And I trained heavy while exhausted until I hurt myself doing something I had done safely hundreds of times before.
That’s the honest version of my credentials. Not that I sleep well, but that I’ve paid for the alternative.
The objection, which deserves taking seriously
A large share of the replies to Johnson made the same point: must be nice. Tell that to a shift worker. Tell that to someone with a newborn. Cancelling your morning is a luxury good.
That’s fair, and it’s the strongest thing said in that thread.
Most people cannot cancel. The nurse still has a shift. The parent still has a school run. Nobody’s boss accepts poor sleep as an agenda item.
But it argues for the opposite conclusion to the one people drew. If you can’t remove the day’s demands, the ability to adjust within them matters more, not less.
You can’t cancel the day. You can move the difficult conversation to Thursday. You can do the filing instead of the strategy. You can drop the heavy lift and walk. You can tell one person, once, that you slept badly and are working with less patience than usual, which costs nothing and prevents a surprising amount.
Johnson’s version of the principle is privileged. The principle isn’t. What his critics actually objected to was him, which is understandable, and it meant a reasonable idea got buried under the noise.
The thing worth keeping
Everyone is told to sleep well. Almost nobody is taught what to do on the mornings they didn’t, which is most mornings, for most people, for most of their lives.
The skill was never sleeping perfectly. It’s knowing how to read what state you’re actually in, and what to change because of it.
That means using what the research tells you to look for, what your own history tells you is normal, and what this particular morning tells you about today. Not a number from a study of somebody else, and not a rule you follow regardless of how you feel. Some mornings after bad sleep you’re fine and should train. Some mornings you’re not, and the honest check takes about thirty seconds.
He got the check right. He just reached for a statistic he didn’t need to justify it, and it cost him the argument.
I’m Paul Calver, a Precision Nutrition Certified Master Health Coach, certified in sleep, stress management and behaviour change, and an iOS developer building humane health apps. This article is educational and isn’t medical advice.