Nobody can tell you where the 10% rule came from. Sports scientists have searched. No founding paper exists. The rule spread through coaching manuals in the late 1980s and early 1990s — not through research. And in 2008, a 532-person randomized controlled trial tested it directly and found it did absolutely nothing to prevent injury.
Yet it’s still the first piece of advice every new runner hears.
The 10% Rule Has No Science Behind It
The 10% rule — the idea that you should never increase your weekly mileage by more than 10% — is officially classified as “injury-prevention folklore” by researchers who have tried to trace its origin. The 2018 systematic review by Damsted et al. in the International Journal of Sports Physical Therapy put it plainly: no evidence exists for its use.
The single most damning piece of data comes from Buist et al. (2008). Researchers randomized 532 novice runners into two groups. One followed a 13-week graded program built around the 10% rule. The other followed a faster 8-week standard program. Injury incidence was 20.8% in the 10% rule group vs. 20.3% in the standard group (p = 0.90). Statistically identical. Zero benefit from the slower ramp.
That’s the only RCT ever run on this rule. It failed.
The Real Injury Data — What 23,047 Runners Actually Show
The most comprehensive review of running injury and training load comes from Fredette et al. (2022), a systematic review of 36 studies covering 23,047 runners. Their conclusion was stark: no universal training progression recommendation can be issued based on current evidence.
What they did find is a pattern. Injured runners averaged 31.6% weekly progression before getting hurt. Uninjured runners averaged 22.1%. In the week before injury, the spike was even sharper — injured runners showed 86% greater progression than uninjured ones (p = 0.026).
A threshold closer to 30% is where risk rises, not 10%.
That gap matters. Think of your body’s connective tissue like a suspension bridge cable. Each week of steady loading sends a slow, controlled signal to strengthen the cable. A sudden large load doesn’t send a stronger signal — it just snaps the wire before it has time to adapt. The 10% rule treats all cables as equally fragile. They aren’t.
What the Weekly Ramp Rule Gets Wrong
The 10% rule treats every runner as identical. The data says the opposite.
Videbæk et al. (2015) pooled 13 studies and found that novice runners sustain 17.8 injuries per 1000 hours of training. Recreational runners average just 7.7 per 1000 hours. Ultra-marathon runners: 7.2 per 1000 hours.
That’s a 2.3x difference. Same rule, two completely different risk profiles.
There’s a second problem. At 10 miles per week, a 10% increase is 1 mile. At 60 miles per week, it’s 6 miles. The same percentage produces 6x the absolute stress. The rule is mathematically incoherent across fitness levels.
Under-ramping has a cost too. Rasmussen et al. (2013) followed 662 marathon finishers and found that those training below 30 km per week before race day had twice the injury risk of the 30–60 km group (RR = 2.02, p < 0.01). Being too conservative isn’t safe — it just shifts the risk from overuse to under-preparation.
Session Spikes Beat Weekly Ramps as the Real Injury Signal
Here’s what changes everything. A 2025 British Journal of Sports Medicine study tracked 5,205 runners over 18 months, logging 588,071 individual sessions via GPS. The researchers tested both week-to-week volume changes and per-session spikes against injury outcomes.
Week-to-week load change: non-significant predictor of injury.
Single-session spikes: dominant injury signal.
A single run exceeding 110% of your longest run in the prior 30 days raised overuse injury risk by 64% (HRR = 1.64, 95% CI: 1.31–2.05). A run exceeding 200% of that threshold doubled the risk (HRR = 2.28). The week didn’t matter. The individual session did.
That’s a different model entirely. The danger isn’t slow weekly creep — it’s the one big day.
The Better Framework: ACWR and Your Fitness Score
The evidence-based replacement for the 10% rule is the acute-to-chronic workload ratio (ACWR). It compares your last 7 days of training load against your rolling 28-day average.
ACWR = 7-day load / 28-day average load
A ratio between 0.8 and 1.3 is the sweet spot. Above 1.5 is the danger zone. Above 2.0, studies show very high injury risk — English Premier League data found a relative risk of 6.7 for non-contact injuries at low chronic load combined with high acute load.
But the calculation method matters. Murray et al. (2017) compared simple rolling-average ACWR against an exponentially weighted moving average (EWMA) in 59 elite athletes over 2 years. The EWMA model was significantly more accurate at detecting injury risk (p = 0.001 in the very-high-ACWR zone). EWMA gives more weight to recent days, which better mirrors how your body actually recovers.
The difference: simple rolling averages treat Monday and Sunday of last week equally. Your body doesn’t.
Most consumer platforms — including TrainingPeaks and WHOOP — use EWMA under the hood. The 10% rule uses nothing.
A Better Benchmark Table
| Runner profile | Current weekly mileage | 10% rule cap | ACWR-safe cap (approx.) | Risk if under-ramping |
|---|---|---|---|---|
| Absolute beginner | 10 mi | +1 mi | +2–3 mi | High — connective tissue adapts slowly but needs minimum stimulus |
| Recreational runner | 25 mi | +2.5 mi | +5–7 mi | Moderate — under 30 km/wk pre-marathon = 2x injury risk |
| Marathon build | 45 mi | +4.5 mi | +7–10 mi | Low — fitness base handles larger absolute adds |
| Experienced runner | 60 mi | +6 mi | +10–15 mi | Very low — high CTL tolerates aggressive weekly gains |
The 10% column produces meaningless ramp increments at low mileage and overly conservative ones at high mileage. The ACWR column adapts to the runner.
Meet the Rule in Practice: A Case Study
James is 42, running his third marathon, currently at 38 miles per week. He decided to jump from his usual Saturday long run of 13 miles to 18 miles in a single week. That one session sat at 138% of his longest run in the prior 30 days — well past the 110% spike threshold the 2025 BJSM study identified.
His weekly mileage increase was only 13%. Under the 10% rule, he was already “over” the cap. Under the ACWR model, the red flag was the single session, not the week total.
He strained his left calf the following Tuesday. Three weeks off.
His training partner, Elena, runs the same mileage. She increased her long run from 13 to 14 miles the same week — 108% of prior 30-day max, just under the spike threshold. Her week-to-week increase was about 6%. Her ACWR hit 1.18. She ran the marathon eight weeks later in 3:41.
One session made the difference. Not the weekly percentage.
How AthleteOS Handles Your Mileage Ramp
AthleteOS doesn’t apply a blanket 10% cap. The AI coach tracks your fitness score (CTL) and calculates your ACWR using an exponentially weighted moving average — the Murray et al. approach that outperforms rolling averages in injury prediction. Your weekly mileage cap adjusts to your current fitness, not a fixed percentage.
It also flags when any single session is set to exceed 110% of your longest run in the prior 30 days. That’s the session-spike trigger the 2025 Garmin-RUNSAFE study identified as the primary overuse injury driver.
For a complete picture of how training load connects to injury risk, see how Zone 2 builds your aerobic base without accumulating injury stress and how aerobic decoupling signals when your base is ready for more volume. If you’re returning from time off, the return-to-run framework covers how to restart your ACWR safely from zero.
Set up your running training plan in AthleteOS to get your ACWR tracked automatically from day one.
The 10% rule sounds sensible. The research doesn’t support it. Your long run on Saturday is far more important than what percentage it represents of your week.
Watch the session. Not the cap.