We Measured 16,513 Comeback Sessions: How Fast Fitness Comes Back After a Break

The First Session Back Feels Awful. It Does Not Last.

A couple of days ago we measured the other side of this: how long you can take off before you actually lose fitness. The calming version turned out to be true. Five days off changes nothing, and two weeks off costs around 5% of aerobic efficiency. But a detraining curve only answers half of the worry. The half that keeps people up is the return. You have had the forced break, you are finally back out there, the first run feels like wading through wet sand, and you want to know whether that goes away and how long it takes.

So we ran the other direction. The short version: the first session back really is down about 3% from your own normal. Then it closes fast. About half of that dip is gone by your fourth session, you are within a percent by six to eight sessions, and you are back at your own baseline in roughly a month of consistent training. The dead-legs feeling is real, and it is also the fastest-improving part.

How We Measured It

We used the same signal as the detraining piece, because it is recorded on nearly every session: aerobic efficiency, which is 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 are detrained, your heart rate sits higher at the same pace or power and efficiency falls. It is the same number that dropped during the break, which makes watching it climb back a clean way to see fitness actually return, rather than a CTL line decaying on a formula.

We pulled 16,513 sessions that followed a break of seven or more days completely off, from 246 athletes, across 3,978 separate comebacks. We scored each session against that same athlete's own normal, then tagged it by how many sessions, and how many days, into the comeback it was. The within-person scoring is the whole game. Some people are simply more efficient than others, so comparing each athlete only to themselves leaves the thing we care about: how far below their own normal they are, and how fast that gap closes.

The Comeback Curve

The first session back averages about 3.1% below the athlete's own normal. That dip is not in your head. It is roughly the loss the detraining curve predicted, sitting right there in the first session.

Then it climbs, and it climbs steepest where it feels worst:

Session 2: down 2.5%
Session 3: down 2.0%
Session 4: down 1.5%, so about half the dip is already gone
Sessions 6 to 8: down 0.9%
A dozen or more sessions in: back at or slightly above your own normal

In calendar terms, efficiency sits about 2.8% down in the first three days back, around 1.8% down by the second week, and essentially level by roughly four weeks of consistent training. A two-week layoff that cost you about 5% is mostly clawed back inside two weeks of showing up, and fully home in about a month.

The Within-Athlete Check

Bucket averages can hide behind who is in them, so we ran the stricter version. We found 242 athletes who had both a fresh comeback session and a settled session later in the same return, and compared each athlete only to themselves. 70% were more efficient once settled than on their comeback, by a median of about 3.4%.

The number that reframed it for us: across 429 breaks where the comeback genuinely started below the athlete's normal, the median time to get back to baseline was two sessions. Not two weeks. Two sessions.

Does a Longer Break Take Longer to Come Back From

This is the question most people actually have, so we split the comebacks by how long the break was. The starting dip behaves exactly like the detraining curve predicts. The longer you were off, the deeper the first session back.

One week off: comeback session about 2.7% below your normal
Two weeks off: about 3.0% below
Three weeks off: about 4.9% below, with the first three days back averaging closer to 5.7% down

No surprise there. More time off means a bigger hole. The climb-out is where it gets interesting. We expected the deeper holes to take proportionally longer to fill, the way the common half-time rule implies. They mostly do not. A one-week break is back within half a percent of normal by about a month of training. A three-week break, despite starting almost twice as low, is also back within half a percent by about a month. The hole scales with the length of the break, but the time to climb out of it stays roughly fixed at a few weeks. The body fills a deeper hole about as fast, because it is rebuilding on machinery it already has.

We would not push that conclusion past three weeks off. Beyond there the sample thins fast, just 31 athletes at the one-month mark and two dozen past six weeks, and the people who log a structured aerobic session after a long layoff are a self-selecting crowd who tend to ease back gently. So we held the length comparison to the one-to-three-week range where the data is dense, and left the long tail as an open question.

Why It Comes Back Faster Than It Left

This is the muscle-memory effect, and it is well documented. When you train, you build cellular and neural machinery: extra nuclei in your muscle fibers, more capillaries, a nervous system that knows the movement pattern. A lot of that infrastructure persists through a break even as the performance fades. So when you start again you are not rebuilding from scratch, you are switching the lights back on in a house that is already wired.

The detraining and retraining research agree on this asymmetry. You lose fitness faster than you would like, and you get it back faster than you fear (Bosquet and colleagues' detraining review; retraining and VO2max recovery summaries). A common rule of thumb is that a break shorter than a few months takes about half its length to undo, which is roughly where our numbers land. The gloomier internet version, two months to regain two weeks, is not what our data shows.

The Caveats

The same honesty that applied to the detraining curve applies here. The first session back is often easy on purpose, and an easy session reads as more efficient because the heart rate is low, so if anything the early dip is understated and the real climb is a touch steeper than it looks. We did not weather-correct, so a hot comeback week adds noise, though the within-athlete comparison limits it. A long break sometimes means illness or injury, which carries its own return curve. And this is submaximal aerobic efficiency, a proxy that happens to be recorded everywhere, not a lab VO2max retest. None of that changes the shape. The dip is real, and it closes fast.

What To Actually Do With This

Come back easy and trust the first few sessions to feel worse than they are. The wading-through-sand feeling on day one is genuine, and it is also the steepest part of the climb, so it improves the fastest. Do not test yourself in the first week back, you will only be annoyed at a number that is about to fix itself. Give it a couple of weeks of consistent aerobic work and most of the loss is gone. Give it about a month and you are level. A coach that adapts the plan to your recovery data will pace that first week back to what your heart rate is actually telling it, instead of throwing you at a benchmark on day one.

The curve that actually matters is not 246 strangers. It is the one written into your own history, where you can watch your heart rate at a given pace drop session by session as you come back. 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 watch your own comeback curve instead of the average of 246 strangers, start here.