How Much Running Is Too Much? The Answer May Be a Single Run

For thirty years, the question “how much running is too much?” has had a weekly answer. Don’t add more than 10% to your mileage a week. Watch your acute-to-chronic load ratio — the number your watch now computes for you. Keep the weekly chart climbing gently. The entire mental model lives on a seven-day clock.

A 2025 study in the British Journal of Sports Medicine — the largest running-injury cohort ever assembled — suggests that clock may be set to the wrong interval. When the researchers looked at single sessions instead of weeks, a clean injury signal appeared. When they ran the popular weekly metrics on the same runners, the signal mostly evaporated.¹ The uncomfortable implication: the run that hurts a runner is often one specific run — much longer than anything they’ve done recently — not a month of creeping volume.

This is a finding worth taking seriously and worth not over-reading. Here’s what it actually says, where it’s solid, and the one place it would be easy to draw exactly the wrong conclusion.

What they did, and what they found

The Garmin-RUNSAFE study followed 5,205 adult runners for up to 18 months — roughly five times the size of any previous running-injury cohort — using GPS data straight off the runners’ own watches and weekly injury check-ins, across 588,071 running sessions.¹ After every single session, they recalculated three competing definitions of “did you spike?” and asked which one predicted the next overuse injury:

• Single-session ratio — today’s distance divided by your longest single run in the previous 30 days.
• ACWR (acute:chronic workload ratio) — this week’s volume over your rolling three-week average. The number on your watch.
• Week-to-week ratio — this week’s volume over last week’s.

All three were binned identically: a flat-or-down session (≤10% increase) as the safe reference, then a small spike (10–30%), a moderate one (30–100%), and a large one (over 100% — a doubling). Every result below is adjusted for age, BMI, sex, prior injury, and running experience.¹

Only one of the three tracked injury — and it was the new one.

Read the top and bottom rows together: a session even modestly longer than your recent best carried about a 60% higher overuse-injury rate, and one that more than doubled your recent longest more than doubled the injury rate.¹ A companion analysis from the same group adds a mechanistic thread — when they classified the injuries by onset, most overuse injuries looked sudden rather than slow-building, exactly what you’d expect if a single session is the trigger rather than gradual accumulation.²

The metric that predicted injury wasn’t weekly mileage or the load score on the watch. It was a single run measured against the runner’s longest run of the past month.

Why “longest run in 30 days” is the clever part

The exposure isn’t weekly volume and it isn’t an average — it’s the ceiling. Today’s run is judged against the highest single session you’ve reached in the last month. Run 11 km when your longest recent run was 10 km, and you’re at a 10% spike. Run 16 km off that same 10 km ceiling, and you’ve spiked 60% — squarely in the elevated-risk zone — even if your weekly total looks unremarkable.

That reframes the lever. The session most likely to be 30%, 60%, or 100%+ above a runner’s recent ceiling is almost always the long run — or a race, or an enthusiastic time trial. Those are the sessions this study points at. Two athletes can finish a week at identical mileage; the one who got there by jumping their long run from 8 to 14 km is, on this evidence, in a different risk place than the one who sprinkled easy kilometers across several days.

The result you must not misread

Here is the place the headline gets dangerous. In this same study, a higher ACWR looked protective — more weekly spiking was associated with lower injury rates.¹ It is tempting to read that as “spiking is fine, maybe even good.” Do not.

The week-to-week ratio fared no better — essentially no association with injury at any spike size.¹ And that null is the part of the whole picture with the most outside corroboration: it has now turned up in novice runners,³ and — notably for the school crowd — in a prospective study of 434 Wisconsin high-school cross-country runners, where week-to-week training changes were likewise not associated with injury.⁴ Three independent datasets pointing the same direction is far more persuasive than any single result.

So the weekly toolkit we inherited — the 10% rule, the load score, the weekly chart — has weaker prospective support than its ubiquity suggests. The weekly lens may simply be the wrong lens.

What’s genuinely unsettled

Honesty about the edges is what separates a finding from a slogan. Four caveats matter before you act on this.

Two more, briefly. The exposure was distance only — not pace, intensity, vertical, or surface.¹ A hard interval session at unchanged distance is invisible to this metric, yet most coaches would rightly call that a spike too. And the “longest run in 30 days” denominator drifts: a taper, an illness, or a two-week break lowers the ceiling, so an ordinary return-to-training run can get flagged as a dangerous “spike.” Both are reasons to apply the principle with judgment rather than as arithmetic.

A note for the multisport coaches: there were no triathletes in this cohort — it was runners.¹ Applying it to a triathlete’s run leg is reasonable coach inference, not a study finding, and two wrinkles matter. The metric only sees run distance, so a brick run off a long ride looks short while landing on already-fatigued legs — the bike that preceded it is invisible to the math. And a bike- or swim-heavy block thins the recent run history, so the first real long run afterward can flag as a “spike” relative to a ceiling that only dropped because the running paused. Judge the run ceiling against recent runs, and treat it as a flag to investigate, not a verdict.

What this means for your athletes

Not a protocol — you know your roster. What the evidence makes actionable: