We Measured 35,735 Sessions: How Long You Can Take Off Before You Lose Fitness
Miss a week and lose all your fitness? We measured output per heartbeat across 35,735 sessions from 406 athletes. Five days off changed nothing. Two weeks off cost about 5%. Here is the real detraining curve.
The Detraining Panic Is Mostly Wrong
Every endurance athlete carries the same low-grade fear. You miss a few days, work gets busy or you catch a cold, and a clock starts ticking in your head on all your fitness draining away. We have felt it too. So we went and checked it against real data instead of the folklore.
The short version: up to about five days fully off, nothing measurable happens. The loss only starts showing after a week, and by two to three weeks off you are down roughly 5%. A missed week is basically free. That is good news, and it matches the science better than the panic does.
Why We Didn't Use CTL
Most apps show you a fitness number, usually CTL. The problem is that CTL falling during a break is fake in a specific way. It is a model that decays on a fixed schedule whether or not your body actually changed. Plot CTL dropping during a layoff and all you have plotted is the formula. We wanted something physiological.
So we used aerobic efficiency: output per heartbeat. For runs, speed divided by heart rate. For rides, power divided by heart rate. When you are fit, your heart does more work per beat, so efficiency is high. When you detrain, your heart rate creeps up at the same pace or power and efficiency falls. It is the flip side of the resting heart rate creeping up that athletes notice during long breaks. It is a real bodily response, it gets recorded on nearly every session, and it is exactly what should fade if time off costs you fitness.
How We Measured It
We pulled 35,735 aerobic sessions from 406 athletes, scored each one against that same athlete's own normal, and then sorted them by how many days the athlete had taken completely off beforehand.
The within-person scoring is the whole game. A category racer and a weekend rider have wildly different efficiency, so averaging them together measures who they are, not what a break did. Comparing each athlete only to themselves strips that out. A score of zero means a normal session for that person. Negative means less efficient than their usual.
Finding 1: Five Days Off Did Nothing
Through about five days off, the comeback sessions were indistinguishable from a normal session. The three-to-five-day bucket came in around -0.7% versus the athlete's own baseline, which sits well inside the noise.
So the "I missed four days and lost everything" story is just not in the data. If anything, a short break behaves like a light taper. The rest is doing its job, not undoing your work. This is the same logic behind a planned deload week: a few easy or empty days let you absorb the work rather than erase it.
Finding 2: After a Week, the Floor Starts to Tilt
Past the one-week mark, the loss becomes real and grows steadily:
- 6 to 9 days off: about 2% down
- 10 to 14 days off: about 3% down
- 15 to 21 days off: about 5% down
It is a slope, not a cliff. It becomes visible right around a week and is clearly there by two weeks. Nothing falls off a ledge.
Finding 3: It Happens to Almost Everyone
Bucket averages can hide behind who is in them, so we ran the stricter version. We found 26 athletes who had both their everyday trained-through sessions and at least one comeback after a 10+ day layoff, and compared each athlete only to themselves.
84.6% of them came back less efficient than their own normal, with a median drop of about 5%. That is the part that convinced us. This is not a handful of detraining-prone people skewing an average. It is close to universal, and it is roughly the same size for almost everyone.
This Agrees With the Research
The detraining literature has been consistent for years. In trained endurance athletes, VO2max holds for the first week or so, then meaningful decline shows up around 10 to 14 days off, on the order of 4% by two weeks and 4 to 7% by three weeks (Bosquet and colleagues' review; Sousa and colleagues). Our efficiency drop of roughly 5% by two to three weeks sits right on top of that.
The science was never the alarmist part. The alarmism is what happens when "you hold your fitness for about a week" gets retold as "you lose it the moment you stop."
The Caveats (One of Them Helps You)
The friendly one first: the first session back is usually easy on purpose, and an easy session tends to look more efficient, not less, because the heart rate is low. That means our estimate of the loss is probably conservative. The real number could be a touch larger, and it would still be this manageable.
A few others worth stating plainly. We did not weather-correct, so a hot comeback day adds noise, though the within-athlete comparison limits it. A long gap sometimes means illness or injury, which carries its own hit. And this is submaximal aerobic efficiency, not a lab VO2max test. Beyond three weeks off the sample thins to fifteen-odd athletes, and who logs a structured session after a five-week break is self-selected, so we cut the curve at three weeks rather than pretend the tail is clean.
What To Actually Do With This
Stop treating a missed week as a setback. Up to about five days off costs you essentially nothing, and the rest is often the point of a recovery week or a forced break. Past two weeks, plan to come back a few percent down. Start with easy aerobic work rather than a benchmark session you will only be annoyed by, and let the efficiency climb back over a couple of weeks. Most of it returns faster than it left. A coach that adapts the plan to your recovery data will do this for you, dialing the first week back to what your heart rate is actually telling it.
The timeline that actually matters is not 406 strangers. It is the one written into your own training history, where you can see exactly what a week off does to your heart rate at a given pace. That is the version worth knowing, and it is sitting in your own data right now.
athletedata connects your Garmin, Strava, Oura, WHOOP and the rest, scores every session and night against your own baseline, and coaches you over Telegram. If you want to see your own detraining curve instead of the average of 406 strangers, start here.
