ACWR and Endurance: What Predicts Non-Contact Injuries in Footballers

The full term ACWRHMLD describes a specific version of the ratio: acute-to-chronic workload, computed on High Metabolic Load Distance (HMLD). HMLD is the GPS-derived distance covered at a speed above 5.5 m/s OR with an acceleration of at least 2 m/s², which captures both high-speed running and the high-energy accelerations that don't show up in raw distance metrics. The "acute" period is the most recent 7 days; the "chronic" period is the most recent 28 days. Divide the 7-day mean by the 28-day mean and you have ACWRHMLD.

The intuition: if a player's recent week is heavy relative to their recent month, they're running hot — fatigued, undertrained for the spike, more vulnerable. If their recent week is light, they're undercooked relative to their conditioning baseline and may be detraining. Sweet-spot guidance has typically said 0.8-1.3, with spikes above 1.5 flagged as elevated injury risk. The 2024 paper tested that intuition against actual non-contact injury data and found the picture more nuanced.

The study tracked 23 first-team Bundesliga players (mean age 24.5, VO2max 53.7 mL/min/kg) across a full competitive season. Eleven players sustained a non-contact injury serious enough to keep them out of team training for at least five days. Each was matched against an uninjured player of the same position; the four-week pre-injury window was compared between the two groups.

The standout result: the chronic workload (CW) — i.e. the 28-day running average — was the best non-contact injury predictor, with the last two weeks before injury carrying the most signal. The ratio itself, and the acute (7-day) component on its own, performed worse than CW. The paper is explicit: "Apparently, only the CW of the ACWR is related to the occurrence of non-contact injuries."

The single strongest predictive metric in the analysis was the chronic workload standardised to each player's individual lactate-threshold velocity (vLT). Standardised CW in the last week before injury produced an AUC of 0.81 (95% CI: 0.59-1.00, p=0.022), with a recommended cut-off of 0.04 km giving 78% sensitivity and 80% specificity. AUC of 0.81 is, in the language of sport-science predictive modelling, a strong signal — not deterministic, but well above the random-chance threshold of 0.5 and competitive with the best individualised load metrics published to date.

Most ACWR work in football treats running load as raw distance — kilometres covered, sprints completed, accelerations registered. But a 7.5 km high-intensity day means something very different for a player whose lactate threshold sits at 16 km/h versus one whose vLT sits at 14 km/h. The first player is running comfortably within their aerobic ceiling; the second is closer to their limit, accumulating more physiological stress for the same external load.

The Marshall et al. study confirms this is more than theoretical. Injured players had a significantly lower VO2max than matched uninjured players (53.2 vs 56.3 mL/min/kg, effect size 0.6) — i.e. the cohort that broke down had less aerobic headroom going into the season. Once you standardise running load to each player's individual lactate-threshold velocity, the same external workload reveals very different internal stress between players. The paper's conclusion is direct: "It is extremely necessary to standardize the ACWR using the aerobic capacity. The isolated use of load parameters is insufficient."

For performance teams looking at how to apply this in-season, the paper points in five directions. None of them require new equipment — most clubs in Europe's top five leagues already have the GPS layer (STATSports, Catapult, Chiron Hego optical tracking) and the pre-season lactate diagnostics in place. The change is in interpretation.

• Run pre-season lactate diagnostics for every senior squad player. Establish individual vLT and VO2max — these are the standardising denominators for in-season load.
• Watch chronic load (28-day) more than the ratio. Sudden spikes in the ACWR ratio are still worth noting; persistent chronic load above a player's aerobic-capacity-adjusted threshold matters more.
• Pay special attention to the last week. The strongest predictive window in the data was the seven days immediately preceding injury — that's where weekly load management matters most.
• Standardise externally-measured load to internally-measured capacity. HMLD on its own is useful; HMLD divided by each player's vLT-anchored capacity is more useful.
• Don't rely on subjective effort (sRPE) for objective load tracking. The paper recommends GPS-derived HMLD over sRPE as the workload variable.

The findings come from a 23-player sample at one Bundesliga club over one season. That's typical for elite-football load-management studies — players don't come in batches of thousands — but it limits how confidently the cut-offs (0.04 km, 78%/80% sensitivity-specificity) generalise to other leagues, other player ages, or women's football. The paper is also explicitly about non-contact injuries; contact injuries (tackles, collisions) are a separate problem with different drivers.

The cohort exclusion of goalkeepers is also worth flagging. Goalkeeper running profiles look nothing like outfield profiles, so HMLD and lactate-threshold metrics need their own benchmark for the position before any ACWR cut-off is applied. KiqIQ's broader performance-science pillar covers the goalkeeper-specific physical profile elsewhere; treat this article as outfield-only.