The standard-issue advice for athletes over forty is a slow apology. Train smarter. Listen to your body. Focus on mobility. It's delivered in a tone that suggests the racing part is behind you and what's left is management.

It's also, on the evidence, mostly wrong about the mechanism.

Masters athletes adapt. Put a genuine hypertrophy stimulus on a fifty-year-old and muscle grows. Put a structured endurance block on them and aerobic capacity improves. The response curve isn't flat and it isn't a fraction of what it was. Adaptation capacity is not the age-limited resource. Recovery capacity is.

That distinction sounds academic until you notice it produces completely different training decisions. If you believe adaptation is the problem, you train harder to compensate. If you understand recovery is the problem, you change the frequency of the bill, not the size of the stimulus. One of those works.

The first change: deload cadence, not intensity

In general training populations, the standard rhythm is a four-to-six-week mesocycle followed by a deload. For masters athletes, the sensible default moves to every three to four weeks.

Note what that adjustment does not touch. It doesn't lower the loads. It doesn't cut the hard intervals. It doesn't replace your squats with a resistance band. The blocks stay the same shape; they just get shorter before you empty the tank and pay it back.

This is the single most useful reframe available to a masters athlete, because it's the one that lets you keep training like an athlete. The stimulus doesn't need softening. The accumulation window does.

The second change: your maintenance floor is higher than the textbook's

If you read the last piece on lead-and-maintain, you'll know the model depends on maintenance being cheap: hold intensity, shed volume, and a non-leading discipline costs you almost nothing.

That's true. It's just less true with every decade, and this is the clearest age-specific finding in the whole literature.

In young adults, muscle has been preserved for months on roughly one-ninth of building volume. Genuinely. One hard session a week, load maintained, and the hypertrophy stays.

Run the equivalent studies in adults aged sixty to seventy-five and it stops working. Neither a one-ninth dose nor a one-third dose reliably preserved fibre hypertrophy. The minimum effective dose for maintenance — not for growth, for merely holding on — rises with age.

That's the masters asymmetry, and it has a practical consequence. Every maintenance number you read in a textbook or a coaching course was probably established in twenty-five-year-olds, and every one of them is a floor you should treat as too low.

For the 35–55 athlete, the honest position is: nobody has studied you properly. You sit in the gap between the two cohorts that have been. The defensible default is to interpolate conservatively — call it a third of building volume, spread across at least two sessions a week, with load held at building intensity — and then calibrate against your own data rather than trusting the interpolation. That is a starting point, not a finding, and it should be treated as one.

The third change: the tissue that doesn't keep up

Here's the part that actually ends masters seasons, and it isn't the engine.

Muscle and the cardiovascular system respond to training on a timescale of weeks. Tendon, ligament and bone respond on a timescale of months. That mismatch exists at every age — it's why young athletes get shin splints in their first big mileage jump — but two things make it sharper for a masters athlete.

First, connective tissue turnover slows. Second, and more dangerously: you have a training history. Your engine is capable of writing cheques your tendons haven't been prepared to cash. A fit forty-eight-year-old coming back from six weeks off can hold a pace their Achilles has no current tolerance for, and the aerobic system will happily let them do it for ten days before anything hurts.

In hybrid racing, where impact volume and loaded carries stack in the same block, the dominant reported injury pattern is exactly this: tendinopathy — Achilles, patellar, rotator cuff. Not acute injury. Not cardiac events. Tendons, gradually, then suddenly.

So connective tissue needs to be a programming variable, budgeted in the plan before the block starts — not a note in your injury log after it fails.

Why your RPE can't see it

The natural objection is that you'd feel it. You wouldn't, and there's a specific reason.

Session-RPE — your rating of the session multiplied by its duration — is a genuinely excellent tool. It's the standard answer to the problem of making a heavy squat session and a long run commensurable, because it measures internal load: the strain the work imposed on you, normalised to you, rather than kilograms and kilometres that don't convert into each other. One number, every modality, addable across a week.

But sRPE tracks cardiometabolic strain, and it has a documented dissociation from mechanical stress under eccentric loading. Two examples make the problem obvious:

  • A bodyweight eccentric calf-raise protocol: RPE 3, near-zero heart-rate response, twelve minutes. By systemic load it barely registers. It is simultaneously one of the highest-value tendon-protective elements in the entire programme — and if you spike its volume, one of the more reliable ways to irritate an Achilles.
  • A downhill-heavy long run and a flat long run of identical duration and identical RPE. Same systemic load. Completely different mechanical cost, showing up 24 to 72 hours later as soreness the session itself never predicted.

Collapse both axes into one number and you hide precisely the thing you needed to see. Which is why load needs tracking on two axes:

  1. Systemic load — session-RPE × duration. The universal currency, comparable across every discipline on your calendar.
  2. Mechanical / eccentric-tissue load — modality-aware and tracked separately. Loaded eccentric work as sets × reps × load. Bodyweight eccentric work as total reps or time under tension, since there's no weight to multiply. Running impact as a distance-and-elevation proxy.

And a third thing that is not a load axis at all: readiness. HRV, resting heart rate trend, sleep. These measure your response to load, not the stimulus. Keeping them structurally separate matters, because "did a lot" and "recovered badly" are different situations that call for different responses, and averaging them into a single readiness score destroys the distinction.

The number that actually predicts injury is a ratio

The instinct with tissue load is to look for a ceiling — a maximum weekly mileage, a rep cap on eccentric work. There isn't a useful one, because any absolute threshold ignores what you've been doing for the last month.

The injury literature — the acute-to-chronic workload work most associated with Gabbett — consistently finds that rapid increases relative to your own trailing baseline predict injury far better than any absolute number. The signal is the spike, not the size.

Practically, that means the number worth watching is:

this week's eccentric volume ÷ trailing 4-week average

with anything meaningfully above roughly 1.3–1.5 as the trigger to look harder at the week ahead. Not a law — a flag. But it's a flag with a real evidence base behind it, and it's individualised by construction, because the denominator is your own history.

It also explains the two moments masters athletes most reliably get hurt, both of which look innocent on a calendar:

  • The return from a break. Two weeks off drops the denominator. Coming back to "what I was doing before" is, arithmetically, a spike — even though nothing about the session changed.
  • The lead change. Moving from a strength-led block into a running-led one can hold systemic load perfectly flat while tripling impact load. The week looks identical in every metric you were watching. It isn't.

What to actually do with all this

Six things, none of which involve training less seriously:

  1. Deload every three to four weeks, by default, before you feel you need it. Keep the loads and intensities; shorten the accumulation window.
  2. Raise your maintenance floors above the published young-adult numbers, then adjust them down only if your own data says you can.
  3. Protect intensity above all else. Volume is what you cut when life intervenes. The moment maintenance work gets comfortable it stops maintaining anything.
  4. Programme tendon work as work — dosed, progressed and logged like a lift, not appended as rehab after something hurts.
  5. Watch the ratio, not the total, especially across returns from breaks and lead changes.
  6. Bias low-impact modalities for endurance maintenance during strength- and hypertrophy-led blocks. The engine barely notices the difference. Your Achilles does.

None of that is "listen to your body." It's the opposite: it's a set of decisions you make in advance, on a calendar, because by the time your body has something to say about tendon load it has usually already happened.

What this doesn't know

The 35–55 maintenance interpolation is exactly that — an interpolation between two studied cohorts, chosen conservatively. Per-athlete calibration is the plan, not a fallback for when the default fails.

The 1.3–1.5 ratio band comes from team-sport populations with much denser data than a self-coached athlete produces. Treat it as a direction of travel rather than a threshold with a decimal point.

And the specific tendon-load prescriptions for hybrid racing rest on a young literature — the sport's first dedicated studies are very recent. The injury pattern is well-observed. The optimal dose response to it is not settled, and anyone telling you otherwise is selling something.