How fast do you lose fitness? Strength vs engine

Your engine fades faster than your strength. In one study of endurance-trained people, VO2max dropped 7% in the first three weeks without training and settled about 16% below where it started after eight weeks.¹ Two weeks off barely touched power athletes' bench press and squat.² Both can be held with much less training than you'd guess, as long as you keep the intensity.

If you're stopping because of an injury or illness, how and when to come back is a question for a clinician or physio. This page is about what happens to fitness, not about medical recovery.

The two clocks side by side#

What was lost after time off

	Value (% lost)
VO2max, 21 days off	7
VO2max, 56 days off	16
Bench press, 14 days off	1.7
Squat, 14 days off	0.9
Eccentric knee extension, 14 days off	12

Different studies and different people: 7 endurance-trained subjects (Coyle et al. 1984) and 12 power athletes (Hortobágyi et al. 1993). Bench and squat changes were not statistically significant. Sources: both abstracts.

These come from different people and different time spans, so read the chart as two shapes, not as a head-to-head. The shapes are what the reviews describe, though. Mujika and Padilla's two-part review of detraining says that in highly trained athletes, a few weeks without enough training brings a rapid fall in VO2max and blood volume, while strength performance declines are limited.³ Over longer breaks, VO2max keeps falling markedly and force production declines slowly.⁴

The engine side#

Coyle's group followed seven endurance-trained people after they stopped completely. VO2max fell 7% in the first 21 days and levelled off after 56 days at 16% below the trained value. Even at 84 days they were still above a group of people who had never trained, 50.8 against 43.3 ml/kg/min.¹ The early drop was linked to the heart pumping less blood per beat; the later one to the muscles extracting less oxygen.¹

Two details matter for you. First, fitness you built recently goes first. Mujika and Padilla note that over longer breaks, recently acquired VO2max gains are completely lost, while long-trained athletes stay above untrained levels.⁴ If you only added running this year, a long break can take that whole gain back, while someone with years of mileage keeps a floor above untrained. Second, the fall isn't linear. Most of the visible loss in Coyle's data happened in the first couple of months, then it flattened.

The strength side#

Hortobágyi's group took 12 power athletes off training for 14 days. Bench press moved -1.7%, parallel squat -0.9% and vertical jump +1.2%, and none of those changes was statistically significant. What did drop was eccentric knee extension force, by 12% in every subject, and the size of type II muscle fibres, by 6.4%.² So the bar numbers held, but something underneath was already starting to move.

Over longer breaks, strength does go down. Bosquet's meta-analysis pooled 103 studies of people stopping resistance training and found losses in submaximal strength, maximal force and power, with bigger losses the longer the break lasted. The losses were larger in people over 65, and larger in inactive people than in recreational athletes.⁵ Bickel's trial put 70 adults through 16 weeks of lifting three times a week, then had some stop entirely for 32 weeks. Strength was largely retained through those 32 weeks, with only a slight drop at the final test.⁶

That's a long time to keep most of a strength gain with no training at all, and it's why the strength half of a hybrid athlete tends to be the more forgiving one when life gets busy.

What a break looks like, by length#

Putting those studies together for the breaks people actually take:

None of these studies looked at people who lift and run at the same time, so the table is a guide to the shape, not a prediction for you.

For a planned break, like a taper or a deload, this is also the reason not to panic. A week or two of reduced training sits at the very start of both curves, and Mujika and Padilla's point is that cutting volume hard costs little as long as some intensity stays in.⁴

How little keeps each one#

You rarely stop completely. The more useful question is what the minimum is. Spiering and colleagues reviewed the maintenance research and found, in general populations:⁷

Their headline finding is that intensity is the variable that matters most, and they're explicit that there isn't enough data to make specific recommendations for athletes.⁷ Mujika and Padilla land in the same place: keep intensity, cut frequency only moderately, and cut volume as much as you need to.⁴ Bickel's young lifters kept their muscle size on a third and even a ninth of their training dose, but the older group didn't.⁶

For a hybrid athlete with a crushed month, that translates into a short list. One heavy lifting session a week, kept heavy. Two runs a week with some real intensity in them. Short is fine. What the research doesn't support is turning everything into easy, low-effort sessions and calling it maintenance. This is extrapolated from general-population studies, so treat it as a floor, not a guarantee.

What the hybridAF score does when you stop#

Here are the exact clocks the score runs on when you stop. They're our design choices, not physiology.

How long a result counts toward the hybridAF score

	Value (days)
Grind (sessions and tonnage)	56
Runs and power rides	90
Watch VO2max reading	120

Source: hybridAF scoring code (src/lib/hybrid-score.ts), 22 October 2026. Lifts have no time limit, so they aren't charted.

The Engine Index only counts runs and power-meter rides from the last 90 days (the run you typed in at signup ages out on the same clock), and a VO2max reading from your watch for 120 days. Your resting heart rate is shown beside those estimates but isn't counted, so it can't hold the engine up on its own. When they age out, the engine falls back to your most recent HYROX race sim if you've logged one, and otherwise to a starting estimate from your experience level, with the score marked as an estimate. The Grind spoke on the radar, which reads how often you train and how much you lift, looks back 8 weeks, so it drops first.

The Strength Index is the odd one out. It reads your best estimated one-rep max on record, with no time window, so it doesn't fall when you stop lifting. That fits the first few weeks of the research above, where strength barely moves. For long breaks it's generous: Bosquet's data says strength does fall eventually.⁵ If you've been off the bar for months, your Strength Index is describing your best, not your today.

Put together, a full stop looks like this in the app. Your strength number doesn't move. The Grind spoke empties out after 8 weeks without a session. Your runs stop counting at day 90, and any watch VO2max reading at day 120, after which the engine side rests on a race sim or the starting estimate. So the headline score changes mostly through the engine, which is roughly the order the research above would predict, just on slower and simpler clocks. Once your engine data has aged out, that half of the score is an estimate, not a measurement.

If you've turned on score-change notifications, have had no session for 21 days, and one side of your score has dropped 4 or more points from its high in the last 180 days, the app can send one notification saying so, at most once every 30 days. That's the whole "decay" feature.

Related reading#

The flip side of this page is the minimum effective dose, which builds a maintenance week around the same research. If you're stepping back on purpose rather than by accident, deload weeks for hybrid athletes covers planned reductions, and session RPE training load shows how to rebuild your weekly load without spiking it. Everything on recovery is in the recovery hub.