The sprint vaccine

Where the fear came from

For most of the time I have been around football, sprinting was treated as a hazard to be rationed. The received view was that hamstrings tore at top speed, so the way to protect hamstrings was to limit the time players spent there. Conditioning was aerobic, technical work was submaximal, and full-speed running arrived on Saturday whether the player had been near it that week or not.

That view was not irrational. Hamstring injuries do overwhelmingly occur during high-speed running, and the problem has been getting worse rather than better: the UEFA Elite Club Injury Study reported that hamstring injury rates rose across recent seasons and now account for around a quarter of all injuries in men's professional football. Twenty years of prevention work, and the number went up.

What changed the thinking was GPS. Once every sprint could be counted, the question stopped being whether sprinting was dangerous and became how much of it, for whom, and how quickly the amount had changed.

The workload ratio and its unravelling

The idea that dominated the next decade was the acute:chronic workload ratio — this week's load divided by the rolling average of recent weeks, with spikes above a certain value flagged as dangerous. It was intuitive, it fitted what coaches already believed about doing too much too soon, and it spread through professional sport faster than almost any idea in our field.

It has since been taken apart. Impellizzeri and colleagues argued that recommending coaches manipulate the ratio to change injury rates assumes a causal effect that no study had properly attempted to estimate, making the practice conjecture and an overinterpretation of the data — and that the statistical properties of a ratio make the metric itself inaccurate and hard to interpret. A follow-up paper went further and called for the ratio and its underlying theory to be dismissed altogether. Separate work identified mathematical coupling, where the acute load appears on both sides of the calculation, as a further problem.

The tool that taught a generation of coaches to think about sprint dose turned out not to survive its own statistics.

Practically, that does not mean load monitoring was a waste of time. It means the numbers are descriptive rather than prescriptive. Knowing that a player covered 40m above 90% of their top speed this week and 380m last week is useful information about what has happened. It is not a threshold that tells you what to do next.

Exposure as protection

The reversal came from studies suggesting that regular exposure to fast running was associated with lower injury risk. The most cited is Malone's season-long work with 37 elite Gaelic footballers, which found that players carrying higher chronic training loads tolerated greater distances and more exposures at maximal velocity than players carrying low chronic loads.

That paper is usually summarised as "sprinting protects you". Its actual finding is more interesting and almost always dropped: the relationship between maximal velocity exposure and injury was U-shaped. Both under-exposure and over-exposure raised risk. There is a window, it differs by player, and you can fall off either side of it. Malone's later work in soccer pointed the same way, suggesting that weekly sprint loading above 90% of maximum velocity was worth monitoring specifically for hamstring risk.

A 2025 scoping review of sprint training for hamstring injury prevention, covering twelve studies using efforts at 80–90% of maximum sprint speed or above, concluded that progressive sprint exposure combined with eccentric strengthening and technical work appeared more effective than any of those in isolation. That is a reasonable summary of where we are — but a scoping review maps evidence rather than weighing it, and twelve studies is a thin base for a claim that has become coaching orthodoxy.

The defensible recommendation is a floor and a ceiling, not a target: get players above 90% of their maximum velocity regularly, and do not let weekly exposure swing wildly in either direction.

Fit players run more

Here is the problem sitting underneath almost all of this evidence, and it is the reason I am cautious rather than evangelical.

Nearly every study in this area is observational. Players who accumulate high chronic loads and high sprint exposure are, by definition, the players who have been available to train. Players who are carrying something, who are old, who are managing a chronic issue, or who are quietly about to break down are the players who do less. So when the data shows that high-exposure players get injured less, at least part of what we are seeing is that healthy players are healthy. Epidemiologists call this the healthy-worker survivor effect, and recent modelling work in professional basketball has argued explicitly that failing to correct for it distorts the whole load-injury picture.

The direction of causation is genuinely unresolved. One 2024 study in professional football found that hamstring injuries were preceded by a short period of higher running demands. Another, from the same group, found that reduced match exposure across the previous two matches accounted for hamstring injury incidence. Both cannot be the simple story. Both can be true of different players.

The biomechanical evidence offers no rescue either. A systematic review of running mechanics and hamstring injury found mostly null or contradictory results, with only lateral trunk kinematics and horizontal propulsive forces showing any consistent signal, and flagged real concerns about study quality and sample size. We do not currently have a running-technique screen that identifies who is about to tear a hamstring, and coaches should be sceptical of anyone selling one.

The threshold problem

Even if the dose were settled, most clubs are measuring it wrong. The conventional approach applies one absolute speed threshold to an entire squad. Take a 25 km/h sprint threshold: for a player whose maximum is 35 km/h that represents roughly 71% of their capacity, while for a player whose maximum is 30 km/h the same threshold sits at about 83%. Those are not the same stimulus, and calling them both a sprint makes the resulting data close to meaningless for individual decisions.

A scoping review of 36 studies on absolute versus individualised thresholds found wide methodological variation and no settled standard, which is itself the finding. Set your thresholds as a percentage of each player's own measured maximum velocity, and re-test that maximum at least a few times a season. This is the single change that most improves the quality of the decision, and it costs one timing session.

One further caution on prescribing submaximal speeds. When field-sport athletes were asked to run at 75% and 90% of maximum, they actually produced around 78% and 89% on average — reasonable — but individual errors ran to roughly 15%, particularly on the slower prescription. Percentages of maximum velocity are a monitoring unit, not a coaching cue. If you want a player at 95%, you have to create a situation where they run fast, not ask them to.

What I do with my footballers

Test maximum velocity properly, from a rolling start over a long enough run-in that they actually reach it, and rebuild the speed zones around each player's own number. Get every player above 90% of that figure at least twice a week, in short exposures, early in the session when they are fresh. Keep weekly high-speed volume broadly stable rather than chasing a target, because both directions of large change look risky in the data. Protect the exposure in congested weeks instead of cutting it first, which is what usually happens.

And hold all of it loosely. The strongest claim the evidence supports is that regularly exposing players to near-maximal speed is better than not doing so, within a window whose edges we cannot yet locate for an individual. That is a genuine advance on where we were twenty years ago. It is not a vaccine, and the language of protection has run some distance ahead of the data.