The 10% Rule Is Wrong: What 5,200 Runners Taught Us About Injury

Every running book, every coach, every injury prevention article says the same thing: do not increase your weekly mileage by more than 10% per week. It is one of the most repeated rules in distance running. It is also, according to the largest running injury study ever conducted, wrong.

The Garmin RunSafe study - published in the British Journal of Sports Medicine in 2025 - tracked 5,205 runners over 18 months using continuous Garmin watch data. Lead researcher Rasmus Ostergaard Nielsen and his team at Aarhus University did not rely on training logs or surveys. They had GPS data, pace data, and distance data from every single run.

Their conclusion flips decades of conventional wisdom: weekly mileage increases are not predictive of injury. What is predictive is a single run that exceeds what your body has recently handled.

What the Study Actually Found

The RunSafe study was designed to test the foundational assumptions behind running injury prevention. The researchers tracked three main variables: weekly mileage changes, the acute-to-chronic workload ratio (ACWR), and individual run distances relative to each runner's recent history.

Here is what the data showed:

Weekly mileage increases did not predict injury. Runners who bumped up their weekly volume showed flat or even slightly decreased injury risk compared to those who held steady. The sacred 10% rule had no statistical support in a dataset of over 5,200 runners.

The acute-to-chronic workload ratio was not predictive either. This is significant because the ACWR is the theoretical foundation behind Garmin's Training Load metric and the Training Status labels that millions of runners check daily. The ratio between your 7-day load and 28-day load - the number your watch uses to decide whether you are Productive, Overreaching, or Detraining - did not reliably predict who would get hurt.

The real predictor was single-run distance relative to your 30-day maximum. If you ran more than 110% of your longest run in the past 30 days, your injury risk increased significantly. The researchers described it as a paradigm shift: "Overuse injuries occur much more frequently during a single training session when the runner runs too far compared to what they are used to."

In other words, it is not the weekly total that breaks you. It is the one run where you went further than your body was prepared for.

Why This Makes Intuitive Sense

One of the most upvoted comments in the massive Reddit discussion of this study - a thread with over 900 upvotes - said it simply: "Fits with what I've found anecdotally. My overuse injuries always started on long runs."

Think about your own injury history. Was it the accumulated mileage of a big training week that got you? Or was it that one long run where you added an extra couple of miles because you felt good, or that race where you pushed further than anything you had done in training?

Runners across the discussion reported the same pattern. Achilles tendon flare-ups after an unexpectedly long run. Tibial stress reactions after a single effort that was far beyond recent training. Not from a gradual buildup over weeks - from one session that crossed the line.

This aligns with how tissue adaptation actually works. Your bones, tendons, and connective tissue adapt to the specific loads you regularly impose on them. A sudden jump in single-session duration applies mechanical stress that exceeds what those structures have been conditioned to handle - even if your weekly volume looks reasonable on paper.

What This Means for Garmin's Training Load

If you use a Garmin watch, you are probably familiar with the Training Status labels - Productive, Overreaching, Detraining, and so on. These labels are built on the ACWR model: they compare your 7-day acute load to your 28-day chronic load and flag when the ratio gets too aggressive.

The RunSafe study found this ratio was not predictive of injury. That does not make Training Load useless - it still helps you understand your overall training trajectory and spot patterns of overtraining or detraining. But it does mean you should not treat it as an injury prevention tool.

Your watch might say Productive while you are actually at elevated injury risk because last Saturday's long run was 120% of anything you have done in a month. Conversely, your watch might flash Overreaching after a high-volume week of moderate runs, even though the individual sessions were all well within your recent capacity.

The gap here is that Garmin aggregates load across the week. The RunSafe data suggests the critical variable is what happens in a single session. These are fundamentally different ways of measuring risk.

The Math: What 5% Per-Run Progression Looks Like

If single-run distance matters more than weekly volume, the practical guideline shifts from "increase weekly mileage by 10%" to something like "increase your longest run by no more than 5-10% beyond your 30-day max."

The researchers suggested 5% per individual run as a more sensible progression limit. Here is what that looks like in practice:

| Current Longest Run (30 days) | Max Next Long Run (5%) | Max Next Long Run (10%) | |-------------------------------|------------------------|-------------------------| | 5 km | 5.25 km | 5.5 km | | 8 km | 8.4 km | 8.8 km | | 10 km | 10.5 km | 11 km | | 15 km | 15.75 km | 16.5 km | | 21.1 km (half marathon) | 22.15 km | 23.2 km |

At 5% progression per long run, getting from a 5 km max to marathon distance (42.2 km) would take roughly 45 runs. If you do one long run per week, that is about 11 months. This is notably slower than most marathon training plans suggest - and it might explain why so many first-time marathoners get hurt during the peak long run phase of their programs.

The practical math gets granular fast. A 5% increase from a 10 km max means you can only add 500 meters to your next long run. That feels conservative. But consider the alternative: a runner whose longest run in 30 days is 10 km decides to jump to 16 km for a weekend long run - a 60% increase. According to this data, that runner has entered the danger zone.

One Underappreciated Implication: Double Days

If the risk is concentrated in individual run distance rather than weekly volume, there is a logical follow-up: splitting the same weekly mileage across more frequent, shorter runs may be safer than running fewer, longer sessions.

Running 60 km per week as six 10 km runs is - according to this framework - a very different injury risk profile than running 60 km per week as three 20 km runs, even though the weekly load is identical. The first approach never exceeds the runner's recent maximum. The second potentially does.

This has practical implications for how you think about recovery between runs. More frequent, moderate-distance sessions with adequate recovery between them may be a better structure than the classic approach of a few hard, long sessions per week. Your Garmin Recovery Time estimates can help you space those sessions appropriately.

Study Limitations: Why You Should Not Overreact

Before you throw out your training plan and start micromanaging every run by 5%, there are important caveats.

Most injuries still happened at "safe" levels. Even using the study's own metric, 87% of injuries occurred in runners who stayed within the 10% single-run threshold. The metric identifies elevated risk, not a binary safe/unsafe boundary. Running injury is complex and multifactorial.

The study did not track intensity. A 15 km easy run and a 15 km tempo run are dramatically different in terms of musculoskeletal stress. The study only measured distance. Pace, terrain, surface, elevation gain, and footwear were all unaccounted for. A runner who does their long run on soft trails faces different injury mechanics than one grinding out miles on concrete.

Injury data was self-reported. Runners logged their own injuries, which introduces bias. Some runners push through pain that others would report as an injury. There was no clinical verification of injury type or severity.

Retention was modest. Of the 13,311 runners initially enrolled, only 5,205 - roughly 46% - completed the full 18 months. The runners who dropped out may have been systematically different from those who stayed (perhaps more injury-prone, or less consistent in training).

The study ran during COVID. Training patterns during 2020-2021 were atypical for many people. Gym closures pushed athletes outdoors. Race cancellations changed motivation and training structure. This may limit how well the findings generalize to normal training periods.

One commenter in the Reddit discussion - a rowing coach with decades of experience - offered useful perspective: "Anyone over 40, I start with the mindset they are returning from a serious injury." That kind of conservative default may be wiser than any single metric, no matter how well-validated.

And another observation that made runners laugh: only 35% of the study participants sustained an injury over 18 months. For a community that treats injury as inevitable, that number seemed remarkably low.

What This Actually Means for Your Training

The RunSafe study does not invalidate everything you know about training. What it does is shift the focus from weekly totals to session-level spikes. Here are the practical takeaways.

Track Your 30-Day Running Max

Start paying attention to the longest single run you have done in the past 30 days. Before your next long run, check: am I exceeding that distance by more than 10%? If yes, consider scaling back. This is a different mental model than tracking weekly mileage, and it requires different awareness.

Progress Long Runs Conservatively

The 5% guideline is aggressive in its conservatism, but directionally right. If you have been running 10 km long runs, do not jump to 16 km because your training plan says so. Build to 11, then 12, then 13. Your weekly volume can increase faster than your long run distance.

Distribute Volume Across Sessions

If you want to increase weekly mileage, consider adding a run rather than making existing runs longer. Five 8 km runs (40 km/week) may carry less injury risk than three 13 km runs (39 km/week), because no single session pushes far beyond your norm.

Do Not Ignore Your Watch Entirely

Training Readiness and HRV Status still provide valuable recovery signals. The ACWR-based Training Load may not predict injury specifically, but chronic under-recovery is still a risk factor. Use your watch data for recovery management even if the weekly load number is not the injury predictor everyone assumed.

Strength Train

The study measured only running metrics, but the broader sports medicine literature is clear: runners who do targeted strength work - particularly for calves, glutes, and hip stabilizers - have lower injury rates. Strong connective tissue tolerates load spikes better. This is especially true if you are over 40 and need to account for the natural decline in tendon resilience.

Be Extra Careful After Breaks

If you took two weeks off for vacation or illness, your 30-day max is still showing your pre-break fitness. But your tissues have deconditioned. Coming back and immediately matching your old long run distance is exactly the kind of spike this study warns about. Rebuild conservatively, even if it feels slow.

The Bigger Picture

For decades, running injury prevention has been framed around weekly totals. The 10% rule was simple, memorable, and felt scientific. The RunSafe study - the largest of its kind, with real watch data instead of training diaries - suggests we have been watching the wrong number.

The right question is not "how much did I run this week compared to last week?" It is "how does today's run compare to the hardest single run I have done recently?"

That shift has implications for how you build training plans, how you interpret your Garmin Training Load, and how you think about the relationship between volume and injury. It does not mean weekly mileage is irrelevant - it means session-level spikes deserve more attention than they have been getting.

If you are trying to make sense of all your Garmin data - training load, recovery time, HRV, readiness - and figure out whether today is a good day for that long run or whether you should keep it short, Should I Train synthesizes all of it into a single daily recommendation. No spreadsheets, no manual 30-day max calculations. Just a clear answer based on what your body is actually telling you.

FAQ

Is the 10% rule for running completely wrong?

The 10% weekly mileage rule is not supported by the RunSafe study data as an injury predictor. However, it is not dangerous advice either - it naturally limits how fast you ramp up, which indirectly constrains single-run spikes. The study suggests a per-run progression limit (staying within 110% of your 30-day max single run) is a more precise and evidence-based guideline.

What is the Garmin RunSafe study?

The RunSafe study is a research project led by Rasmus Ostergaard Nielsen at Aarhus University, published in the British Journal of Sports Medicine in 2025. It tracked 5,205 runners over 18 months using Garmin watch data, making it the largest running injury study ever conducted. Its key finding is that single-run distance relative to your recent maximum - not weekly mileage changes or the acute-to-chronic workload ratio - is the strongest predictor of overuse running injuries.

Does Garmin Training Load predict running injuries?

According to the RunSafe study, the acute-to-chronic workload ratio that underlies Garmin's Training Load and Training Status features was not predictive of injury in their dataset of 5,200 runners. Training Load is still useful for understanding your overall training trajectory and recovery needs, but it should not be relied on as a primary injury prevention metric.

How do I prevent running injuries based on this research?

The most actionable finding is to avoid single runs that exceed 110% of your longest run in the past 30 days. Progress your long run distance by 5-10% at a time, distribute weekly volume across more frequent shorter runs when possible, and rebuild conservatively after any break from training. Combine this with strength training and adequate recovery between sessions.

How long does it take to build from 5K to a marathon using the 5% rule?

At a strict 5% per-long-run progression starting from 5 km, you would need approximately 45 long runs to reach marathon distance (42.2 km). With one long run per week, that is about 11 months of consistent progression - slower than most 16-20 week marathon plans, but potentially with a much lower injury risk.