Can the ACWR Predict Injuries? What Coaches Should Know Before Trusting the Number

"Keep the ACWR below 1.5, or injuries will follow." If you work in football, basketball or any team sport, you have probably heard some version of this rule. The short answer is that there is not enough evidence to use any ACWR value as a hard threshold for injury risk. A 2020 commentary by Impellizzeri and colleagues explains why: much of the research linking training load to injuries rests on shaky conceptual and methodological ground.

What the study looked at

This is not a new experiment. It is the second part of a two-part commentary on training load and injury prevention, published in the Journal of Athletic Training. The authors review the body of research that relates load metrics to injuries and identify problems that many of those studies share — both in how the questions are framed and in how the data are analyzed.

Importantly, the paper ends with practical recommendations for the people who actually manage athletes' load day to day.

Key findings

The authors highlight four main issues.

1. There is no causal model linking load to injury. Most studies have been run without a clear framework explaining how a load metric would lead to an injury. Without one, an observed association is easily misread as cause and effect. The authors compare it to the classic spurious correlation between shark attacks and ice cream sales — both rise in summer, but one does not cause the other.

2. The ACWR itself has methodological problems. The ratio divides recent (acute) load by longer-term (chronic) load. Because the acute load appears in both the numerator and the denominator, the two parts are mathematically coupled, and using a ratio introduces statistical distortions of its own. These properties undermine its validity as a predictor of injury.

3. Researchers have too many degrees of freedom. How many days count as "acute"? How many weeks as "chronic"? Which metric, and where do you draw the cut-offs between groups? Without a framework, these choices are left open, which greatly increases the risk of chance findings (false positives) and of results that simply confirm what the researchers expected.

4. Reporting is inconsistent. Differences in study design, unclear statistical methods, small samples and selective reporting make results hard to interpret and hard to apply to other teams or sports.

Based on this, the authors recommend that practitioners avoid relying on unvalidated numerical metrics and instead adjust training load using traditional training principles — such as progressing load gradually — while watching how each athlete responds.

How to read this in practice

The paper does not say that measuring load is pointless. Its target is the idea that a single number can predict injuries and should dictate training decisions.

On the field, a realistic takeaway looks like this:

• Do not treat the ACWR or similar values as a red-light/green-light system for planning sessions.
• Interpret load data together with how athletes are responding — their wellness, perceived fatigue and soreness.
• Prioritize the basics, such as avoiding sudden spikes in load.

Keep in mind that this is a commentary, not a trial comparing metrics head to head. The full text is freely available on PMC; this article is a summary of its main points prepared by the ATHCON sports science team, so please refer to the original for exact wording.

Using ATHCON

The authors' advice — adjust load by watching athlete responses — only works if you can collect those responses every day without adding paperwork.

In ATHCON, athletes log sleep, fatigue, soreness and other items on their phones each morning. The app then compares each value with that athlete's own baseline over the previous 14 days and flags the deviation in four tiers: great, normal, caution and alert. Rather than a single ratio or a team-wide cut-off, it looks at change from each athlete's personal "normal." Because training records sit alongside wellness data, you can quickly see whose responses shifted in a week when load went up.

The flags are a prompt to check in with an athlete, not a diagnosis. The decision stays with the coach.

Takeaways

• There is not enough evidence to use an ACWR value as a hard injury-risk threshold.
• Training load and injury research has recurring pitfalls: no causal model, distortions from the ratio itself and too much analytic flexibility.
• Stick to sound training principles and adjust load based on how each athlete responds, not on one number.