Injury Prevention and Load

01What the measurements show

The Numbers Behind Load and Injury Prevention

Eight findings on what a training spike does, and on what the ratio built to catch it turned out to measure.

02Load is the lever

Tissue Fails When the Demand Exceeds What It Was Prepared to Tolerate

Injury prevention in sport is load management, full stop. Tissue does not fail because an athlete trained. It fails because the demand placed on a tendon, ligament or muscle exceeded what that structure was prepared to tolerate at that moment. Every preventive strategy reduces to one equation, the load applied against the load capacity built.

That equation now has a large measurement behind it. Adult runners were tracked by Garmin device across 18 months, and 1,820 of 5,205 reported an overuse injury during 588,071 recorded sessions Schuster Brandt Frandsen 2025. The exposure that raised the injury rate was a single session running more than 10 percent past the longest run of the previous 30 days.

The spike is a session, not a season

The size of the spike graded the risk. A rise of 10 to 30 percent carried a hazard rate ratio of 1.64, a rise of 30 to 100 percent carried 1.52, and a rise above 100 percent carried 2.28. An athlete who doubles their longest recent effort in one session more than doubles the rate at which injury arrives.

Note where that measurement sits. It is not a weekly total and not a monthly average. It is one session, read against what this athlete has recently done. The injuring load is a particular afternoon, and the tissue that fails is the one holding the demand when capacity runs out.

How Sports Injuries Happen carries that moment in full, as a capacity event rather than a tissue accident. Prevention operates on both sides of it: lowering the demand that arrives on any single day, and raising the capacity the athlete brings to it.

03The workload ratio

The Acute to Chronic Workload Ratio Spread Because Training Protects as Well as Injures

The acute to chronic workload ratio was sport's attempt to measure that balance directly. It compares the most recent week of training, the acute load, against the rolling four week average, the chronic load that is taken to reflect what the body has actually adapted to. A value above one says the athlete is doing more this week than they have been doing.

The reasoning behind it inverted a piece of coaching dogma, and that is why it spread. Higher training load was assumed to cause injury. The counter-argument held that training protects, that well developed physical qualities are associated with reduced injury risk across a wide range of sports, and that under-training may raise risk instead Gabbett 2016.

Where the numbers in circulation came from

Two seasons of data on 53 elite rugby league players supplied the figures every performance department went on to quote. A ratio at or above 2.11 carried a 16.7 percent injury risk in the current week and 11.8 percent in the week after Hulin 2016. High chronic workload combined with a very high two week ratio at or above 1.54 reached 28.6 percent.

A companion paper carried the idea into return to play Blanch 2016. It reported a polynomial relationship between the ratio and injury likelihood with an R squared of 0.53, and recommended that the ratio join the return to play decision. That paper now carries a published correction. Return to Play owns the clearance question and the criteria that belong in it.

The finding inside it that survived

One result from that first cohort matters more than the ratio it was reported in. Players carrying a high chronic workload were more resistant to injury across ratios of 0.85 to 1.35 than players carrying a low one, with relative risks running from 0.3 to 0.7. Acute and chronic workloads considered in isolation did not consistently predict injury.

Read plainly, that says the protection came from the base. Sharp spikes, where the acute load races ahead of the chronic base, are where hamstrings tear and tendons give way. The athlete who absorbs one of those weeks without failing is the athlete who built a wide base first.

04The critique of the ratio

The Workload Ratio Shares Its Numerator With Its Denominator

The acute to chronic workload ratio has a statistical problem at its foundation. The most recent week sits inside the four week average it is divided by, so the top of the fraction is part of the bottom. Two quantities built that way correlate with each other whatever the athlete does. The objection is named mathematical coupling, and it produces spurious correlation inside the conventional calculation Lolli 2019.

This is not a quibble about decimals. It means a relationship can appear between the ratio and injury when nothing in the athlete's training produced it. Every threshold, every colored zone on a monitoring dashboard, and every recommendation to hold a player inside a band rests on that relationship being real.

Uncoupling the ratio does not repair it

The obvious repair is to remove the acute week from the chronic average and recompute, which is called the uncoupled ratio. Elite basketball and weightlifting data were used to test whether that changes anything Coyne 2019. Differences between coupled and uncoupled values were almost all trivial, at effect sizes of 0.04 to 0.21, and coupled and uncoupled ratios correlated at 0.88 to 0.99.

Those authors read their result as reassurance, since coupling appeared to change little in practice. The deeper objection reads the same numbers the other way. A metric that gives the same answer whether or not its known defect is removed has not been vindicated. It has been shown to be insensitive to the thing it is supposed to be measuring.

A methodological review of the ratio reaches that conclusion directly Impellizzeri 2020. It lists the failure to normalize the numerator by the denominator even when uncoupled, an absence of background rationale for a causal role, and ambiguity in what the number means. It also records a relation to injury risk that is neither consistent nor unidirectional.

Random denominators produced the same result

The sharpest test replaced the chronic load with numbers that carry no information at all. Using previously published data, acute load was divided by a fixed value of 1,510 and by randomly generated chronic loads, then modeled the same way as the original Impellizzeri 2021.

The real ratio returned an odds ratio of 2.45 for injury. The fixed denominator returned 1.95. The random denominators returned odds ratios from 1.16 to 2.07, and their mean of 1.89 was inflated rather than flattened. Neither the ratio nor acute load alone beat a model containing nothing but an intercept, at c-statistics of 0.574 and 0.544 against 0.5.

A denominator that can be replaced by a random number without changing the answer is not carrying information. The ratio rescales acute load, magnifies its effect estimate, shrinks its variance, and reclassifies players who have not changed what they did.

The bins were chosen rather than found

The zones layered on top of the ratio add a second problem. Continuous training data gets cut into discrete categories at boundaries nobody measured, and players land in a risk band because of where a line was drawn. A review of the analytical issues names discretization alongside sparse data, confounding that varies over time, bias introduced by injured athletes, and the initial load problem created by exponentially weighted moving averages Wang 2020.

Those reviewers also record that the exponentially weighted method does not apply cleanly to sports where athletes taper, which is most of them. Their verdict on current recommendations is that the limitations affect their validity and should discourage the use of the ratio.

The case is not unanimous. A systematic review of 22 studies, at a median quality score of 8 on the Newcastle-Ottawa scale, supports the association between the ratio and non-contact injury Griffin 2020. It treats the ratio as one tool inside a larger monitoring system and favors the exponentially weighted model for its greater sensitivity.

The largest dataset settles the practical question. Among 5,205 runners, the session spike predicted injury and the acute to chronic ratio showed a negative dose-response, while the week to week ratio showed nothing Schuster Brandt Frandsen 2025. Load matters. The ratio is the wrong instrument for reading it.

05Screening before the season

The Injury History Is the Strongest Item on the Screen, and Its Reach Crosses Sites

Prevention begins with the preparticipation physical exam, the screen that maps an athlete's status before a single rep is loaded. A comprehensive baseline covers vitals, cardiovascular and pulmonary screening, a neurological examination, and a static and functional musculoskeletal evaluation. It surfaces pre-existing conditions before competition can expose them. The Preparticipation Exam carries what that examination yields and where it falls short.

The functional layer matters most for soft tissue risk. Movement screening exposes the asymmetries, the limited ranges and the compensatory patterns that quietly funnel load into the wrong structures. Asymmetry and the Dominant Side carries what a side-to-side difference does and does not predict, including the reviews where it predicted nothing.

What the history is actually telling you

A previous injury is the strongest item on the intake form, and its reach is wider than a recurrence at the same site. Pooling three studies, a history of anterior cruciate ligament injury raised the risk of a subsequent hamstring injury with a relative risk of 2.25 Toohey 2017. Previous lower limb muscular injury raised the risk of a lower limb muscular injury at a different site.

That transfer is selective rather than general, which is the useful part. In the same review a history of chronic groin injury did not raise subsequent hamstring risk, at a relative risk of 1.14 with a confidence interval from 0.29 to 4.51. A history of concussion and a range of joint injuries did raise later lower limb risk.

So the athlete who has been hurt before earns a sharper screen and a more deliberate ramp back into full workload. The screen is aimed at what the old injury changed rather than at the old injury itself. That is a nervous system question as much as a tissue one, since an injured joint changes what it reports upward and the system reorganizes around the report.

06The neuromuscular warm up

The FIFA 11+ Cut Football Injuries by 39 Percent and Its Predecessor Cut Nothing

A structured warm up is one of the most validated injury prevention tools in sport, and it works on the nervous system as much as the muscle. Comprehensive neuromuscular warm up programs combine running, strength, plyometric and balance drills, and the FIFA 11+ carries the most randomized evidence behind it.

The landmark trial randomized 125 Norwegian clubs and followed 1,892 female players aged 13 to 17 across one season Soligard 2008. The program was built to improve strength, awareness and neuromuscular control during static and dynamic movements.

What those trials actually found

The result deserves reporting exactly as it came out. The primary outcome, lower extremity injury, did not reach significance, at a rate ratio of 0.71 with a confidence interval of 0.49 to 1.03. Injuries overall fell to 0.68, overuse injuries to 0.47 and severe injuries to 0.55, and those three did reach significance.

Pooling the trials sharpens the picture and adds a null beside it. Four cluster randomized trials of the FIFA 11+ returned an injury incidence rate ratio of 0.61, a 39 percent reduction Thorborg 2017. The two trials of the earlier FIFA 11 program returned 0.99, with a confidence interval from 0.80 to 1.23. One program in the same family moved nothing.

The male collegiate record runs in the same direction. Sixty-five National Collegiate Athletic Association teams were randomized, and injury incidence ran 15.04 per 1,000 athlete exposures in the control group against 8.09 in the intervention group Silvers-Granelli 2015. That is a 46.1 percent reduction, at a rate ratio of 0.54. Those are trial findings under trial conditions rather than a promise to any athlete.

Adherence is the dose

A prevention program is only as good as its completion rate, and the completion rate has been measured. Teams in the Norwegian trial completed the program in 77 percent of all training and match sessions, at a mean of 1.3 sessions a week Soligard 2010. Players completed it in 79 percent of the sessions they attended.

Players with high compliance carried a 35 percent lower risk of all injuries than players with intermediate compliance, at a relative risk of 0.65. Coaches who had used injury prevention training before coached teams at a 46 percent lower risk, at an odds ratio of 0.54. The program and the belief in it arrived together.

Why the warm up reaches the nervous system

The mechanism is readiness. Most of the effects of a warm up are attributed to temperature, through decreased stiffness, an increased nerve conduction rate, an altered force-velocity relationship and increased anaerobic energy provision Bishop 2003. Mechanisms that do not depend on temperature are also proposed, along with an increase in preparedness.

A warm up raises tissue temperature, speeds nerve conduction and rehearses the proprioceptive and reflexive loops that protect a joint under sudden load. It primes the central integrative state that governs an athlete's readiness, tuning autonomic balance and motor drive so the first explosive movement meets a system already awake rather than one caught flat.

07Strength buys capacity

Strength Training Cut Sports Injuries to Under a Third and Stretching Did Not Move Them

Strength training is the most powerful injury prevention intervention available, and the pooled trials are unambiguous. Across 25 randomized trials, 26,610 participants and 3,464 injuries, strength training carried a relative risk of 0.315 Lauersen 2014. Stretching carried 0.963, with a confidence interval from 0.846 to 1.095.

Proprioception training reached 0.550 and multi-component programs 0.655. Acute injuries fell to 0.647 and overuse injuries to 0.527. Progressive strength work reduces sports injuries to a fraction of the rate seen in unstrengthened athletes, and it outperforms stretching and proprioception drills taken in isolation.

The mechanism is capacity. A stronger tendon, muscle and bone simply tolerate more load before failure, which moves the whole distribution of demands the athlete can meet. Eccentric work earns special weight, since exercises that load a muscle as it lengthens build resilience in the position where injuries happen. Lower-Extremity Rehab carries the Nordic hamstring evidence and the progressions built on it.

The goal is load tolerance, raising the chronic capacity high enough that the demands of sport land well inside what the body is prepared to absorb. None of these trials prevented injury by removing load. Every one that moved the injury count added something, whether strength work, a rehearsed warm up or a wider base. Building a body and a nervous system that meet load and ask for more is the intervention. Principles of Rehabilitation carries the loading progressions and the tissue timelines they are matched to.

The manual therapy prevention trial was retracted

One study in this area is regularly cited for chiropractic care and injury prevention, and it should not be. A 2010 randomized controlled trial added a sports chiropractic intervention to best practice management in 59 semi-elite Australian Rules footballers and followed them across a season Hoskins 2010.

It reported no significant difference in hamstring injury incidence, at an odds ratio of 0.116 and a p value of 0.051. It did report a significant reduction in primary lower limb muscle strains and in weeks missed to non-contact knee injury. Both 2010 reports from that trial were retracted in 2011.

The retraction notice for the companion report states the reason Hoskins 2011. The publisher was advised by the authors' institution, Macquarie University, that its Human Research Ethics Committee did not approve the study. The work was retracted because it had been conducted without institutional ethics committee approval.

A retracted trial is not evidence, and citing it does not make it one. The claims made here for care aimed at the nervous system are claims about identity and mechanism, stated as the model's, and each one carries the measurement that would test it.

08Recovery and readiness

Time on the Training Base Lowers the Risk of the Next Injury

Recovery is the half of load management that gets skipped, and it is where adaptation is actually banked. Tissue and the nervous system grow stronger during rest rather than during the session. An athlete who never absorbs enough recovery accumulates fatigue until capacity quietly falls, and the next ordinary load becomes the injuring one. Overtraining is load management that lost its balance.

Time spent building the base is protective in its own right. Team sport athletes who completed more than 18 weeks of training before their first injury were at reduced risk of a subsequent one Gabbett 2016. Well developed physical qualities track with lower injury risk across a wide range of sports, and under-training raises risk rather than lowering it.

Reading readiness rather than counting load

This is why monitoring readiness belongs inside prevention. Heart rate variability, the RMSSD measure of beat to beat variation, reads the autonomic recovery state and flags the athlete trending toward overreach before an injury declares itself. Heart Rate Variability owns that measurement, including the trials where training guided by it beat a fixed plan and the trials where it did not.

The readiness signal and the training log answer different questions. A workload number counts what was asked of the athlete. An autonomic reading reports what the athlete did with it. When the two disagree, the athlete is the one holding the information.

Sleep, nutrition and planned deload weeks are the inputs that let the chronic workload base keep rising safely. Adaptation and Supercompensation carries the line between productive overreaching and the state that costs a season, and the recovery time that separates them.

09What we corrected

Three Claims Removed From This Page

This page previously carried a risk window of a workload ratio above 1.5. That figure traces to one rugby league cohort, where players on a high chronic base became less resistant to injury at ratios near 1.5 Hulin 2016. It is gone as a threshold, because the ratio it belongs to does not survive its own statistics. What replaces it is the session spike measured in 5,205 runners.

The page also stated that the spike is the single most reliable predictor of soft tissue injury. That superlative could not be traced to a source and has been removed. The measured claim stands in its place Schuster Brandt Frandsen 2025. A session more than 10 percent past the longest run of the previous 30 days raised the injury rate, at hazard rate ratios of 1.64, 1.52 and 2.28 across three spike sizes.

A third line held that ratios kept inside a balanced range protect the athlete. The protection recorded in that cohort tracked the size of the chronic base rather than the position of the ratio, so the sentence has been rewritten to say what was measured. A display quotation also sat at the foot of the strength section. Its claim now appears in the body, with the trials that support it.

10The model's claim

Protection Is the Width of the Load an Athlete Can Absorb and Return From

Two layers run through this page and they should not be confused. The established science is the coupling critique, the random denominator demonstration, the session spike cohort, the pooled FIFA 11+ result with the FIFA 11 null beside it, and the strength meta-analysis. Those results belong to the investigators who ran the programs.

The Unified Model of Tone starts by relocating the question. The ratio asks how far an athlete sits from a number. The model asks how wide a band of load this athlete can absorb and return from. Health lies in the width of the range a system can move inside rather than in proximity to any particular value.

Read that way, the failure of the ratio is expected rather than surprising. A ratio reports a position. Injury follows from the width of the band and from how fast the athlete comes back inside it, and no single position on a scale carries either quantity. The Hulin cohort measured the band without naming it, which is why the chronic base predicted and the ratio band did not.

Load management is dosing

The second half of the reading is about magnitude. Every input a program applies sits somewhere on a continuous scale of size, from a light session to a full contact match. Load management is the act of choosing where on that scale the next input lands. The model treats an underdose as cheap and informative. "A small input that fails has cost the system little and has taught the clinician something."

That is why a progression built from small increments beats a progression built from a target number. The athlete who stays inside 10 percent of their longest recent session and comes back clean has bought information at almost no cost. The athlete who doubles it has bought a hazard rate ratio of 2.28 and no information at all.

The prediction this page makes

The band an athlete can absorb has never been measured directly, and every instrument needed to measure it already sits in a performance department. Record four readouts in the same athletes on the same mornings. RMSSD for variability structure. Stride-interval timing variability for coupling. Reactive strength index from a drop jump for reflex responsiveness. Time to return to baseline heart rate after a standardized submaximal load test for recovery time.

Define the absorbable band as the largest single-session increase above that athlete's recent maximum that leaves all four unchanged by the following morning. The model predicts that band width separates the athletes who go on to be injured from those who do not. It predicts that band width does so better than any acute to chronic ratio computed on the same training logs, and that the band narrows before mean session output falls.

This is a claim about how protection is organized rather than a claim about what treatment does. If RMSSD, stride-interval timing variability, reactive strength index and time to return to baseline are shown to move together within the same athletes across a season, the unification claim is confirmed.

11The tone reading

Load, Constraint and Time Course Inside a Training Week

Three signatures of tone carry load management on this page, each inside a number a performance department already keeps.

The rest of the library carries the same logic through its other foundations. Set-point is the value the ratio treats as the thing to defend, and the model relocates protection away from it. Gain decides how large an answer a given session produces, which is why the same program returned a rate ratio of 0.61 in one family of trials and 0.99 in another. Input quality is the difference between a warm up rehearsed properly and one performed flat, and completion at 77 percent of sessions is where that shows. Prediction is the feedforward model that meets the landing before the athlete is aware of it. Coupling names the timing relationships a fatigued athlete loses before losing output, and oscillation is the rhythm underneath the readiness signal a monitoring system samples each morning. The full framework is set out in the Unified Model of Tone.

12Where this sits

How This Page Relates to the Rest of the Library

Seven places the load argument continues, each with the claim that earns the link.

13Questions athletes ask

Questions Athletes Ask

14The sources

References

19 primary sources, each linked to its record. Figures quoted on this page were checked against the published abstract.