The Acute:Chronic Workload Ratio (ACWR)
Introduced by Tim Gabbett and colleagues in 2016, the ACWR is a simple but powerful ratio that compares your recent training load (acute) to your longer-term training load (chronic). Typically, acute load represents the past 7 days of training, while chronic load represents the rolling 4-week average of prior weekly loads. The formula is:
ACWR = Acute Load (7-day) ÷ Chronic Load (28-day average)
The underlying logic mirrors Banister's classic fitness-fatigue model: acute load represents the fatigue component, while chronic load represents developed fitness. When your recent week of training is proportional to what your body has been prepared for, recovery and performance are optimised. When the ratio spikes, meaning you've suddenly done far more than your baseline, injury risk rises.
A 2020 systematic review (The Relationship Between Acute:Chronic Workload Ratios and Injury Risk in Sports, Open Access Journal of Sports Medicine) examined 22 studies and found consistent support for the association between high ACWR and increased non-contact injury risk. An ACWR above 1.5 was repeatedly associated with elevated injury incidence across multiple sports.
The Rolling Average (RA) Model
The simplest version of ACWR uses a Rolling Average, each of the 28 days in the chronic window is treated as equally important. If you trained 50 km/week for four weeks, your chronic load is 50 km. However, this approach has a significant limitation: it treats training done three weeks ago the same as training done yesterday. This is physiologically unrealistic.
The Exponentially Weighted Moving Average (EWMA)
The EWMA model was developed to address this shortcoming. Rather than treating every day in the time window equally, EWMA assigns greater weight to more recent training loads, and the weight decreases exponentially as you go further back in time.
In practice, this means that if you had a big training week two weeks ago but have since rested, your EWMA-based ACWR will reflect your current, lower fatigue state more accurately than the RA model would.
A 2020 systematic review by Malone et al. (British Journal of Sports Medicine) and subsequent analyses confirmed that EWMA has greater sensitivity than the RA model, particularly at higher ACWR values. When ACWR exceeds 1.5, EWMA-based calculations show significantly higher hazard ratios for injury than RA-based calculations, meaning it captures risk more accurately at the critical danger zone.
Which Should You Use?
For practical day-to-day use, both models are informative, but the EWMA is considered the more physiologically accurate of the two. The key difference in application: the EWMA will react faster to a sudden increase in load and will also "recover" faster after a drop in training, better reflecting real biological adaptation and detraining dynamics.
A 2024 descriptive epidemiology study among 430 recreational runners (Journal of Athletic Training) found substantial differences in which training sessions were flagged as high-risk depending on which calculation method was used, with different methods identifying up to 43% non-overlapping sessions. This highlights the importance of knowing which metric your app or watch is using.
