The off-season: how long to rest and what to keep doing

Every autumn the same question: how long should I stop? The two wrong answers are the common ones. Some riders never stop, carry a season's fatigue into the winter and are stale by February. Others put the bike away in October, do nothing until January and spend the whole of base rebuilding what they had. The detraining literature is precise enough to give a better answer: take a genuine break of two to four weeks in which riding is light, unstructured or replaced by something else; keep one short hard session in the week; and do not let the break run past about six weeks, because after that the losses stop being cheap to recover. This post explains where those numbers come from and what "keep doing" means in practice.

What you lose, and how fast

The clearest single experiment is still Coyle's from 1984. Seven endurance-trained athletes stopped training completely for twelve weeks and were measured throughout. Maximal oxygen uptake fell by about seven per cent in the first twelve days and by about sixteen per cent over the full twelve weeks. The early loss was almost entirely cardiovascular: stroke volume dropped quickly, and a follow-up study by the same group showed why. Blood volume falls by several per cent within two to four weeks of stopping, the heart has less to pump per beat, and when the researchers artificially restored blood volume the stroke volume came back. The muscle side was slower: the oxidative enzymes in muscle declined steadily over the first eight weeks and then stabilised at a level still well above that of untrained people, while capillary density did not change at all over the twelve weeks. And even after twelve weeks of nothing, the athletes' VO₂max remained well above that of untrained controls.

The reviews since have confirmed the shape. Mujika and Padilla's two-part review puts short-term (under four weeks) losses in VO₂max at roughly 4 to 14 per cent, driven by blood volume and stroke volume, with muscle adaptations lagging; longer stops lose more, and the muscle changes catch up. A 2022 meta-analysis of detraining in athletes found the same split: a small drop in VO₂max after short-term detraining and a considerably larger one after long-term detraining. In elite cyclists measured across a season, aerobic indices fell over the winter transition and recovered through the following preparation period, which is the expected pattern rather than a failure.

Three practical points fall out of this:

• The first two weeks are mostly plasma. The early, alarming-looking drop is blood volume, which returns within days once training resumes. It is not the fitness you spent a year building.
• Muscle keeps for a while. Capillaries stay; mitochondrial enzymes decline slowly. A break of a few weeks does not undo the base.
• Six weeks or more is a different story. By then the muscle changes are substantial and the rebuild is genuinely a rebuild. We go through the full timeline in detraining: how fast you lose fitness.

Why you should still stop

If a break costs fitness, why take one? Because the alternative costs more. A season of training and racing accumulates fatigue that a recovery week does not clear, small injuries that need time without load, and, more than either, a kind of mental wear that the sports-science consensus on recovery treats as seriously as the physical kind. Kellmann and colleagues' consensus statement describes recovery as having physiological, psychological and social components, and notes that inadequate recovery over long periods is on the path to overtraining. A rider who has been "on plan" for ten months needs a period where nothing is measured, and the riders who skip it are the ones who write to us in March saying they are inexplicably flat. See overtraining vs overreaching for the warning signs.

The break also lets you do the things a training plan makes difficult: fix a niggling bike-fit issue, get a dental appointment, walk in the hills, see people. That is not soft; it is the reason the next season is possible.

How long: two to four weeks, in our view

Put the physiology and the psychology together and a window appears. Under two weeks is not really a break; the fatigue does not clear and the head does not reset. Over four to six weeks, muscle losses become expensive. Between two and four weeks of light, unstructured activity, the losses are mostly the cheap-to-recover kind and the benefits are real. These are our figures, and the right length within that band depends on how hard the season was: a rider who raced twenty times needs the four; a rider whose season was one sportive may need the two, and may not need a formal break at all if the year was already gentle.

"Break" does not mean bed. The reduced-training studies are useful here. Houmard's runners cut their weekly volume by about 70 per cent for three weeks while keeping their intensity and maintained their VO₂max and their 5 km time. A large cut in volume with a little intensity kept is a recovery, not a detraining. That is the model for the off-season: less, not nothing.

What to keep doing

• One short hard session a week. This is the most evidence-backed item on the list. Rønnestad's group had well-trained cyclists do either low-intensity riding only or low intensity plus one high-intensity interval session a week through an eight-week transition. The group that kept the intensity held its power at threshold and its performance through the winter, while the low-intensity-only group declined; and the intensity group was ahead when the next season's training began. Twenty to thirty minutes of real work once a week, indoors if it is easier, is enough. It is the cheapest insurance in cycling.
• Move most days. Walking, running if your knees allow it, swimming, an easy spin. The aim is to keep blood volume and habit, not to train. Anything that gets you out of breath a little for half an hour counts.
• Start strength work. The off-season is the one time you can afford to be sore. Two sessions a week of heavy, simple lifts (squats, deadlifts, step-ups) over eight to twelve weeks is the pattern in the cycling strength studies, and it is easier to start in November than in March. See strength training for cyclists.
• Sleep and eat normally. Not less, on the grounds that you are riding less. Under-eating in the off-season is a common way to arrive at base run down.
• Fix the bike and the fit. This is when a saddle change or a cleat adjustment can bed in without disturbing key sessions.

What to stop doing

• Structured training. No plan, no intervals apart from the one session, no power targets on the easy rides.
• Tracking. Leave the head unit at home for the easy rides, or at least do not look at it. The point of the break is partly to stop measuring.
• Racing. Including the unofficial kind on the club ride.
• Long rides. Two hours is plenty. The long ride returns in base.

Coming back

The return should be gradual but not timid. The first two weeks after a break restore blood volume and reacquaint the legs; keep them easy and frequent, and resist testing yourself, because the numbers will be poor for reasons that are about plasma, not fitness. From week three, base training proper begins. If you left the break at four weeks, expect to feel like yourself again after about two weeks of riding and to see your old numbers after four to six, in our experience. Riders who took a much longer break are in different territory; see getting back on the bike after a break.

Do not re-test FTP in the first fortnight. A test done when blood volume is still recovering gives a number you will then train to, and it will be low. Give it a month. If you use Moveee, the Power Profile estimates your threshold from ride data rather than a single test, so the estimate drifts back up as you do, without a bad test anchoring it.

A note on the calendar

The off-season does not have to be in the autumn. It has to follow your last important event and precede your base. A rider whose main goal is an October hill climb takes the break in November; a rider whose goal is a spring classic may take it in May and again briefly in September. What matters is that it exists, that it is short, and that one hard session survives it. The plan wizard lets you set a start date up to a week out and blocks dates you tell it you cannot train, so the break can be a real one and the plan starts when you do. How the break fits into the rest of the year is in the cycling year.