ACWR after the 2026 meta-analysis: a signal, not a forecast

Few metrics have divided sports science as sharply as the acute:chronic workload ratio. Popularised by Tim Gabbett's 2016 paper on the training-injury prevention paradox, ACWR compares what an athlete has done recently with what they are used to. It became a standard line in load-monitoring dashboards. It also drew sustained criticism, most prominently from Impellizzeri and colleagues in 2020, who argued that its causal link to injury had never been established and that the ratio itself had statistical problems.

In August 2026, a multilevel meta-analysis in Frontiers in Public Health brought new numbers to the debate. It is worth reading carefully, because it supports neither side's strongest claims.

What the meta-analysis found

Ding and colleagues screened more than 10,000 records, included 41 studies in their systematic review and pooled 16 studies with 797 team-sport athletes in the meta-analysis. The main findings:

• Elevated ACWR was associated with injury, with a small-to-moderate pooled effect (Hedges' g 0.35, 95% CI 0.16 to 0.54).
• Heterogeneity was very high (I² of 95.8%), so the size of the association varies a great deal between studies and settings.
• Injury type was the only significant moderator. The association was clear for contact (g 0.54) and non-contact injuries (g 0.55) but essentially absent for time-loss injuries as a category (g 0.01).
• Rolling-average and exponentially weighted (EWMA) calculations did not differ significantly (g 0.38 against 0.26), and neither did internal against external load, time window or sport.
• Sensitivity analyses were stable and there was no sign of publication bias.

The authors' conclusion is unambiguous: ACWR should not be used as a stand-alone injury prediction tool. It is best treated as a contextual indicator within a multi-marker load management system, and they caution against universal thresholds. They also note that reverse causation is possible. Staff often reduce a player's training when something feels wrong, so the injury and the change in load can share a cause, and causal conclusions cannot be drawn.

ACWR is best read as a question: why has this player's recent load moved away from what they are used to?

Using ACWR without overselling it

For a staff that already calculates ACWR, the meta-analysis is not a reason to delete the column. It is a reason to change how the number is used:

• Never read the ratio without the absolute loads behind it. A ratio of 1.5 on a very low chronic base and on a high one are different situations. Impellizzeri's group highlighted known problems with ratio data, which is why the raw numbers matter.
• Treat bands such as 0.8 to 1.3 as local working ranges, not laws. The authors explicitly warn against universal thresholds.
• Combine it with other markers. Monotony and strain describe how training is distributed across the week, and a daily wellness check-in shows how the player is responding.
• Compare each player with themselves. Returning players, new signings and youth players moving up will often show extreme ratios for good reasons.
• Be honest about what it can claim. An elevated ratio says the recent load is unusual for this athlete. It does not say an injury is coming.

Rolling average or EWMA?

The exponentially weighted moving average was proposed by Williams and colleagues in the British Journal of Sports Medicine as a better way to compute the ratio, because it gives more weight to recent days and decays older ones gradually. In this meta-analysis the difference between the two methods was not statistically significant, which the authors attribute partly to limited power in the subgroups. In practice, consistency probably matters more than the choice: pick one method, use it for every player and every season, and keep the raw loads visible.

The debate is unlikely to end with this paper, and it should not have to. A modest, real association is useful when it is used as one input among several, not as a forecast.

SquadStats calculates ACWR from session RPE with both rolling and EWMA methods, alongside Foster's monotony and strain, a daily wellness check-in and a 0 to 100 readiness score. They sit on the same screen, so the ratio is always read in context.

Sources

1. Ding L et al. Acute:chronic workload ratio and load management for team sports: a multilevel meta-analysis. Frontiers in Public Health, 2026
2. Gabbett TJ. The training-injury prevention paradox: should athletes be training smarter and harder? British Journal of Sports Medicine, 2016
3. Impellizzeri FM et al. Acute:Chronic Workload Ratio: Conceptual Issues and Fundamental Pitfalls. International Journal of Sports Physiology and Performance, 2020
4. Williams S et al. Better way to determine the acute:chronic workload ratio? British Journal of Sports Medicine, 2017