There are a few posts doing the rounds about a recent study on running distance and injury risk. Published in the *British Journal of Sports Medicine* last year, the study followed just over 5,200 runners for 18 months using data from their Garmin devices. Frandsen and colleagues were looking at whether sudden increases in running distance were linked to injury.
The take I keep seeing is that the old 10% rule wasn’t wrong, we were just applying it to the wrong number. The authors’ own recommendation is to avoid any session more than 10% longer than your longest run in the previous 30 days to reduce injury risk.
The study does give us useful evidence. Runners who made unusually large jumps in a single session got injured at higher rates. But I don’t think we can take that finding and turn 10% into a new universal rule.
There are several reasons for that. The first is the design of the study. This was an observational study, so the researchers can show that certain training patterns and injuries tend to turn up together, but they can’t show that the training pattern caused the injury. Runners who make bigger jumps may differ from everyone else in ways the analysis doesn’t pick up, whether that’s experience, injury history or what they’re training for.
But there are also some important questions about what the different measures in the study actually tell us, and how much we can take from them when we’re deciding how to progress training.
1. It only looks back 30 days
The researchers analysed the training data from just over 5,200 runners over 18 months. For each run, they compared the distance with the longest run that athlete had done in the previous 30 days. The problem is that 30 days is a pretty short window.
Take two runners. Runner A ran a marathon six weeks ago and has spent most of the time since recovering. Their longest run in the last 30 days is 15 km. They go out and run 20 km, a 33% increase.
Runner B has never run further than 15 km. Over the next month they build to 16.5, then 18, then 19.5 km. Each step is 10% or less, so every one of those runs sits in the study’s lowest-risk category.
On the 30-day metric, Runner A has made the big risky jump and Runner B looks like they’ve been progressing sensibly.. As a coach, I could make a pretty good argument that Runner A’s jump is the less concerning of the two. Runner A handled 42 km six weeks ago, and their 15 km runs may simply reflect what they’ve been doing during recovery. Runner B is now 30% beyond anything they’d run a month ago, covering distance they’ve never done before.
So the metric tells you how much the latest run has changed relative to the last month. It can’t tell you how far the athlete is moving beyond what they’ve handled over the longer term. That’s a pretty big limitation.
2. It doesn’t account for weekly volume
Two athletes both extend their longest run from 20 to 24 km, a 20% increase each. One averages 50 km a week and the other averages 100. For the first athlete that 24 km run is almost half the week, while for the second it’s under a quarter. The study records both as the same 20% increase, and it didn’t test whether weekly volume changes the risk attached to that jump.
But that 20% increase represents a very different change in training load for those two athletes.
3. Bigger jumps don’t mean steadily bigger risk
The estimated injury risk was lowest in the 10% or less group which is consistent with the authors’ recommendation. But after that, things get a bit strange. The injury rate was actually lower in the 30–100% group than in the 10–30% group.
So the finding isn’t as simple as saying the bigger the increase, the greater the injury risk. If that was the case, you’d expect injury rates to keep going up as the size of the increase went up. They didn’t.
And again, this is an association. Runners making the biggest jumps may have been different in ways that also affected their injury risk, so I wouldn’t use these numbers to decide how much you can safely increase a run.
4. The researchers didn’t find 10%, they started with it
The analysis didn’t reveal that injury risk changes at 10%. The researchers took the existing 10% rule, used it to define the reference group, and compared larger increases against it. That’s a fair way to test whether the traditional rule has any relationship with injury. It doesn’t show that 10% is where the body tips from safe to unsafe.
Look at who’s in that reference group, too. It includes everyone who increased their longest session by 10% or less, which means it also includes everyone who reduced it. Someone who dropped from 30 km to 20 km sits in the same reference group as someone who went from 20 to 22. That’s a broad mix of training behaviour to be treating as the safe baseline.
The authors acknowledge this themselves. In the discussion they point out that increases of up to 10% aren’t necessarily safe either. When they moved the cutoff down to 1%, runners who increased by between 1% and 10% had a 19% higher injury rate than those who reduced or increased by less than 1%, although that result wasn’t statistically significant.
So you could just as easily ask: why 10%? Why not 1%? The analysis changes depending on where you put the cutoff, and the 1–10% result doesn’t give us evidence that 10% is a meaningful line between safe and unsafe training.
They also point out that their findings don’t account for steady progression across several sessions. Their example is 10, 11, 12.1, then 13.3 km in the same week. Every one of those runs is within 10% of the one before, but without enough recovery between them it can still add up to too much. So the study doesn’t support the idea that anything under 10% is automatically safe.
5. The study also looked at weekly volume
The researchers also looked at progression across whole weeks. One method was the acute:chronic workload ratio, which compares distance in the most recent week with the average of the previous three. You’d expect a higher ratio, meaning a week well above your recent average, to go with more injuries. The study found the opposite. As the ratio went up, injury rates went down, and runners who more than doubled their recent weekly average had about a 25% lower injury rate than the reference group.
Now apply the same logic people are applying to the long run finding, and you’d conclude the new safe rule is to more than double your weekly volume. Nobody is saying that, and for good reason. An association between big weekly increases and fewer injuries doesn’t make big weekly increases safe. We need to be just as careful with the long run result.
The second method simply looked at the change in distance from one week to the next. Here, they found no relationship with injury at all.
So the two weekly measures gave different answers. Higher ACWR was associated with fewer injuries, while week-to-week change showed no relationship at all. That doesn’t tell us which measure is right. It shows how much the answer depends on how you define progression, and why one association from one measure shouldn’t be turned into a training rule on its own.
6. Distance isn’t the whole load for trail runners
The study measured the distance of each athlete’s longest running session, but didn’t distinguish between types of session. Runs with less than 15 minutes between them were combined into one, so an interval session with its warm-up and cool-down counts as a single run and could end up as the athlete’s longest.
For trail and ultra runners, distance is only part of what determines how demanding a session is. Elevation gain, technical terrain, descending, hiking, time on feet and intensity can all change the load. A 30 km road run and a 30 km mountain run are very different sessions, and the study has no way of knowing which one the athlete’s previous longest run was.
Despite all those limitations, I think the study has highlighted something that hasn’t received nearly as much attention as weekly volume or overall training load: what happens when you suddenly make your longest run much longer than anything you’ve done recently.
If an athlete’s longest recent run is 20 km and they ask me if they can go out and run 35 km with a mate, the answer is probably no. This study gives us some evidence that making a big jump in your longest run is associated with a higher injury rate.
What it doesn’t give us is enough evidence to treat 10% as a universal rule, and it tells us nothing to suggest anything under 10% is automatically safe.
So what do you do instead?
Before increasing your long run, ask yourself whether you’ve adapted to the distance and duration you’re already doing. If you feel ready to increase it, then look at the bigger picture — your previous running history, your normal weekly volume, any recent injuries, and what the new long run actually involves.
Are you adding more flat distance, more uphill hiking, more technical downhill, or more runnable downhill? Those are all different changes in the demand of the session.
The increase might be as small as five minutes or as large as 30 minutes or more, depending on the athlete’s training history. An athlete returning to long runs after a race might go from one hour to two hours to three hours over the space of three weeks. An athlete gradually building their long run might take two months to go from 10 to 20 km.
There isn’t one number that makes sense for both athletes. The important thing is whether the athlete is ready for the next increase based on what they’ve already adapted to.
The 10% figure can still be a useful warning flag, because injury rates were higher above it in this dataset. But they may have found the same thing if they’d used 5% or 15% instead. We don’t know. So I’d use 10% as a reason to stop and think about the increase, rather than as a line between safe and unsafe training.
