Although the injury is a complex and multifactorial phenomenon, numerous studies highlighted how less than optimal workload management represents one of the biggest injury risk factors. Therefore, the possibility of excluding one of these factors is an opportunity that trainers, coaches and sports scientists cannot ignore.

3 ACWR calculation models

There are two main models for calculating the ACWR (fig. 6) [15]:

• the “Rolling Average” model (RA) (as explained in Part 1);
• the “Exponentially Weighted Moving Average” model (EWMA).

Fig. 6 Differences in the acute:chronic ratio (ACWR) according to the calculation model used (RA vs EWMA) (Williams, 2016) [15].

The main difference between the two models is the “weight” assigned to each training load within the time frame considered. Let’s consider, for instance, Athlete 3 in fig. 6: compared to RA, the EWMA method assigns a greater “weight” to the loads undertaken by the athlete in the last period (from day 25 to day 28), when the athlete is in a state of fatigue due to a substantial increase in workloads. This leads to a higher acute:chronic ratio on day 28 for the EWMA model (acute:chronic=1.55), which is probably more accurate than the RA model (acute:chronic=1.43), considering recent workloads the most relevant ones.

The “Rolling Average” model (RA)

According to the RA model, the workload carried out during the last week (“acute workload”) in absolute terms (i.e. the total workload) is related to the average workload of the previous four weeks (their arithmetic mean, “chronic workload”). Therefore, in the RA model, workloads carried out both during the past and more recent periods are equivalent: all the training stimuli have the same “weight” over time (for example, the effects of a training session done 28 days ago are similar to those of a training session done 2 days ago). According to this model, there is a linear relationship between workload and likelihood of injury. This is due to the fact that this model does not take the natural and physiological decrease of the effects of a training session over time into account, and how workloads have been distributed over time.

The model which became popular to bypass the limitations of the Rolling Average is the EWMA model [15-17].

The “Exponentially Weighted Moving Average” (EWMA) model

The EWMA model [15-17] assigns a gradually decreasing weight to workloads to which the athlete has been subjected: recent training sessions have a greater “weight” compared to those carried out in the past. This model takes the physiological decrease of the effects of a training session over time into account, hence the non-linear nature between workload and injury. Moreover, the EWMA model more appropriately represents the variations with which workloads have been accumulated over time. Some studies have highlighted how the EWMA model shows greater sensitivity and accuracy compared to the RA model (fig. 7). 

Fig. 7 The relationship between the likelihood of injury and ACWR (RA vs EWMA). Each graph takes into consideration a different workload variable: (A) total distance, (B) “moderate-speed distance” (6-18 Km/h), (C) “high-speed distance” (18-24 Km/h) and (D) “player load”. When the ACWR is greater than 1.5, the EWMA model shows greater sensitivity in detecting an increase in the likelihood of injury compared to the RA model (Murray, 2016) [17].

4 Choice of workload variables

The ACWR model allows monitoring many workload parameters, including variables linked to the “external” load (like distance, energy, the number of accelerations or sprints, distances above speed or power thresholds etc) and variables linked to the “internal” load (like HR, the RPE or the training load) [9, 13, 14]. The heterogeneity of the parameters used in the studies shows that there is no univocal consensus in the literature on which are the most “sensitive” variables to be used in training load monitoring. The choice of the parameters depends on the specific requests of each sport and on the type of injury taken into consideration. For example, in contact sports like rugby, the monitoring of the number of impacts and the injuries linked to them represents one of the most important variables to monitor with the ACWR model. In this way, in sports where hamstrings injuries caused by high speeds are common, like in soccer, one of the parameters to take into account is the distance covered at high speed to which the athlete has been exposed during the weekly training and match.

Even though there are conflicting opinions on which are the variables to monitor through the ACWR model, the experience and data collected in the past seven sports seasons (six Serie A football leagues and one Pro 14 rugby league) lead to two further considerations linked to this difficult choice:

1.  using internal and external workload parameters together seems to represent the best compromise in order to have a global vision of the stimuli to which the athletes are subjected. With the same external load, two athletes can have a totally different perception of the effort. For example, high levels of anxiety or prolonged insomnia can amplify the perception of the effort during a training session or a match. Under these circumstances, the evaluation of the external workloads alone could be inadequate and hide particular stress perceived by the athlete, which represents a possible injury risk [19].
2. the simultaneous monitoring of more variables can provide more detailed information, not only about the type of stimuli the athletes have been exposed to during a training session or a match but also on the probability of the injury risk. While the presence of one spike only (ACWR > 1.5) in one parameter represents a quite “alarming” situation (fig. 8), the presence of more spikes at the same time highlights a particularly risky situation, such as to entail timely measures in future workload planning (fig. 9).

Fig. 8 The histograms show volumes both in terms of distance covered (meters, chart A at the top) and energy consumption (J/kg, chart B at the bottom) by a professional rugby player during a 4-week period (Pro 14 season 2019/20). The orange squares represent the weekly ACWR for each parameter. The acute:chronic ratio for both variables (distance and energy) does not highlight any spike (ACWR between 0.9 and 1.2).

“Is there an optimal range called “sweet spot” within which the likelihood of injury is low, whereas outside it the injury risk considerably increases?“

Alberto Botter
Benetton Rugby Performance Analyst

The fourth week the ACWR is 1.5 (“danger zone”) for the distance and 1.2 (“sweet spot”) for the energy. The analysis of one parameter only (like distance) highlights a possibly risky situation caused by a considerable increase in the distance covered in the fourth week. However, an overall approach which considers more variables detects a pretty safe situation, confirmed by an energy expenditure falling within the optimal range (0.8 < ACWR < 1.3). The greatest distance covered and a limited additional expenditure of energy reflect the specific requests of training coaches: more technical and tactical sessions at low intensity.

Fig. 9 The histograms in each chart show the weekly distance (m), the accelerations (number of accelerations greater than 2.5 m/ s2 of a duration of 0.5 s), the energy (J/kg) and the power events (numbers of events performed at an intensity above VO2max) to which a professional soccer player has been subjected (serie A league 2018/19). The orange circles represent the weekly ACWR for each parameter. The fifth week the athlete was exposed to greater workloads compared to the previous four weeks and so there are four spikes (red circles), one for each parameter (ACWR between 1.65 and 1.82). 

5 Choice of the time window

Another aspect which influences the ACWR model is represented by the time interval chosen for the calculation of the acute and chronic training load. Time windows which are normally used vary from 2 to 9 days for the acute load and from 14 to 35 days for the chronic one. The choice of different time windows leads to a different acute:chronic ratio. One of the few studies carried out on Australian football examined the combination of different time intervals to calculate the acute and chronic load [13]. The study shows that the interval from 3 to 6 days for the acute load and to 21 days for the chronic load is more associated with the likelihood of injury (fig. 10). However, the authors underline that the choice of the best time window in the calculation of acute and chronic load depends on the sport taken into consideration and on the number of training sessions, but mostly on the number of weekly matches which athletes undergo.

Fig. 10 Three different time windows for the calculation of the acute:chronic workload (3:21 days, 6:24 days and 3:18 days) and the likelihood of injury. ACWR at 3:21 days (in yellow) has the highest association (R2 = 0.79) with the likelihood of injury. MSR (m) = moderate speed running (distance covered between 18 and 24 Km/h) (Carey, 2016) [13].

6 Limitations and criticism of the ACWR approach

The growing interest arisen in the last years for the study about the relationship between workloads and likelihood of injury, pointed out some weak spots of the ACWR model.

1) In a study of 2018 [22] carried out on professional soccer players, it appears clear how the ACWR shows a high specificity but a low sensitivity in determining the injury risk. The authors conclude that even though there is a significant association between ACWR and likelihood of injury, this is not to be confused with the capability of predicting the injury, that is to say, that the model cannot identify the athletes who will suffer an injury. Similar results emerged in a study by Raya-González in 2019 [23] who analysed the relationship between internal load (RPE and training load) and the capability of the ACWR model of predicting non-contact injuries in young professional soccer players. No relationship emerges from the study between internal load and injury. Moreover, the ACWR pointed out a poor capacity in predicting the non-traumatic injury.

2) Another criticism of the approach shared by multiple authors is to believe that there is a “cause-effect” relationship between workloads and injury [25]. According to this belief, the ACWR model can explain all injuries. However, like many studies show, the causes of injuries are complex and multifactorial [26-28] (fig. 11). There is no linear relationship between a single and isolated variable and injury, but rather the interaction of more related factors which can provide a more complete picture of the causes of injury.

For example, it has been proven how low muscular levels, a limited aerobic fitness or a condition of emotional stress, as well as previous injuries or prolonged insomnia, are just some of the factors which increase the likelihood of injury [25, 26].

Fig. 11 The injury is a complex and multifactorial phenomenon. The graph shows the interaction between some variables and the risk factors associated with the injury risk (adapted from Bittencourt, 2016) [25].

3) In a 2019 study, Enright and colleagues examined whether there was a relationship between ACWR and likelihood of injury both in terms of the type of injury (muscular, tendinous or ligamentous) and seriousness of the injury (number of training sessions which the injured athlete did not carry out with the team). The results have shown that: 

• the ACWR is not sufficiently sensitive in discerning the type of injury;
  
• there is no direct connection between ACWR and seriousness of the injury.

4) One of the essential requirements for the correct calculation of the acute:chronic ratio is the possibility to daily monitor workloads to which the athlete is subjected. However, the absence of athletes during certain periods of the year or during the competitive season (for example, the summer break in soccer or the gatherings and commitments with the National team during the Six Nations in rugby) makes it extremely difficult to monitor training workloads with constancy. These “time gaps” limit the possibility to use the ACWR model the whole year through, or at least they make it difficult to recover data related to workloads carried out outside the usual training sessions and matches.

5) Lastly, a limit of the ACWR model concerns the athletes who return to play after an injury. Especially for injuries which require more weeks of rehabilitation, the strict respect of the “sweet spot” (0.8 < ACWR < 1.3) or the increase of training workloads smaller than 10% compared to “chronic workload”, will considerably delay the return of the athlete to the team. In this way, increasing the training workloads by more than 10% is often necessary, in order to quicken the process of rehabilitation of the injured athlete (fig. 12) [25].

Fig. 12 Hypothetical relationship between chronic workload and possible changes in the weekly training load. Each bar represents an increase of the weekly workload equal to 10%. Limited increases of weekly workloads (< 10%) are recommended both when the chronic workload is extremely low or extremely high (bars in red). On the contrary, weekly workload increases greater than 10% are well tolerated by athletes who have a medium/high chronic load. These increases may be necessary in order to quicken the rehabilitation process of injured athletes (bars in green) (Gabbet, 2018) [25].

7 Results

Literature supports this association between ACWR and injury. The use of this model represents a useful tool in monitoring workloads within a greater monitoring system, as well as other scientifically proven methods. As a matter of fact, ACWR allows an objective analysis of the stimuli to which athletes have been exposed to and it represents a useful tool in training planning and periodization. The principles on which the model is based are:

• a gradual progression of workloads by limiting elevated variations between them;
• the creation of an adequate chronic workload (the highest, the better) in order to prepare the athlete to the “worst-case scenario”, protecting the athlete from possible spikes and reducing the likelihood of injury.

The choice of training load variables, as well as the time window, are closely related to specific requests of the sport taken into consideration, and to the number of training sessions, but above all, to the number of matches to which athletes are exposed on a weekly basis. However, it seems that the combined use of internal and external workload parameters can provide a more detailed and precise global vision of stimuli to which athletes are subjected.

Although the injury is a complex and multifactorial phenomenon, numerous studies highlighted how less than optimal workload management represents one of the biggest injury risk factors. Therefore, the possibility of excluding one of these factors is an opportunity that trainers, coaches and sports scientists cannot ignore.

ACWR is not a “magic number” but a useful tool in workload monitoring. Like any other tool, the knowledge of possible advantages that this method can provide and the awareness of the intrinsic limits of the model itself, are essential for its correct use.

Gpexe users can check the tutorial section into the web app for more details about how to use the ACWR. 

Read the first part of the article here: https://gpexe.com/2020/06/16/acutechronic-workload-ratio-part-1/

Author: Alberto Botter

References (articles cited in Part 1. are included)

1. Soligard, T., Schwellnus, M., Alonso, J., Bahr, R., Clarsen, B., Dijkstra, H., Gabbett, T., Gleeson, M., Hägglund, M., Hutchinson, M., Janse van Rensburg, C., Khan, K., Meeusen, R., Orchard, J., Pluim, B., Raftery, M., Budgett, R. and Engebretsen, L., 2016. How much is too much? (Part 1) International Olympic Committee consensus statement on load in sport and risk of injury. British Journal of Sports Medicine, 50(17), pp.1030-1041. 
2. Banister, E., Calvert, T., Savage, M., 1975. A systems model of training for athletic performance. Aust J Sports Med, 7, pp.57-61.
3. L.Z.F. Chiu, J.l. Barnes. The Fitness-Fatigue Model Revisited: Implications for Planning Short- and Long-Term Training. National Strength & Conditioning Association Volume 25, Number 6, page 42–51.
4. Gabbett TJ. Severity and cost of injuries in amateur rugby league: a case study. J Sports Sci. 2001;19(5):341–7
5. Drew MK, Raysmith BP, Charlton PC. Injuries impair the chance of successful performance by sportspeople: a systematic review. Br J Sports Med. 2017;51(16):1209–14.
6. Bittencourt NFN, July 2016. Complex systems approach for sports injuries: Moving from risk factor identification to injury pattern recognition-narrative review and new concept. British Journal of Sports Medicine 50(21):bjsports-2015-095850.
7. Griffin A., November 2019. The Association Between the Acute:Chronic Workload Ratio and Injury and its Application in Team Sports: A Systematic Review. Sports Medicine 50(9):1-20.
8. Blanch P, Gabbett TJ, 2016 Apr. Has the athlete trained enough to return to play safely? The acute:chronic workload ratio permits clinicians to quantify a player’s risk of subsequent injury. Br J Sports Med. 50(8):471-5.
9. Hulin BT, Gabbett TJ, Lawson DW,et al. The acute:chronic workload ratio predicts injury: high chronic workload may decrease injury risk in elite rugby league players. Br J Sports Med Published Online First: 2015;0:1–7.
10. Gabbett TJ, 2018. Debunking the myths about training load, injury and performance: empirical evidence, hot topics and recommendations for practitioners. Br J Sports Med.
11. Gabbett TJ., 2016. The training—injury prevention paradox: should athletes be training smarter and harder? British Journal of Sports Medicine, 50(5), pp.273-280.
12. Lolli L., Batterham AM, Hawkins R, Kelly DM, 2017. Mathematical coupling causes spurious correlation within the conventional acute-to-chronic workload ratio calculations. Br J Sports Med. 2019 Aug;53(15):921-922.
13. Carey DL, Blanch P, Ong K-L, et al. Training loads and injury risk in Australian football—differing acute: chronic workload ratios influence match injury risk. Br J Sports Med. October 27, 2016 doi:10.1136/bjsports- 2016-096309.
14. K. J. Weiss, S.V. Allen, M.R. McGuigan et al. The relationship between training load and injury in men’s professional basketball players. International Journal of Sports Physiology and Performance. February 8, 2017.
15. Williams, S., West, S., Cross, M. and Stokes, K., 2016. Better way to determine the acute: chronic workload ratio? British Journal of Sports Medicine, pp. bjsports-2016-096589.
16. Hunter, J., 1986. The exponentially weighted moving average. J Quality Technol, 18(203), p.10.
17. Murray, N., Gabbett, T., Townshend, A. and Blanch, P., 2016. Calculating acute: chronic workload ratios using exponentially weighted moving averages provides a more sensitive indicator of injury likelihood than rolling averages. British Journal of Sports Medicine, 51(9), pp.749-754. 
18. Hawley, J., 2002. Adaptations of Skeletal Muscle to Prolonged, Intense Endurance Training. Clinical and Experimental Pharmacology and Physiology, 29(3), pp.218-222.
19. N. F. N. Bittencourt, W. H. Meeuwisse, L. D. Mendonça, A. Nettel-Aguirre, J. M. Ocarino, S. T. Fonseca. Complex systems approach for sports injuries: moving from risk factor identification to injury pattern recognition—narrative review and new concept. Br J Sports Med 2016;50:1309–1314.
20. Malone S, Hughes B, Doran DA, Collins K, Gabbett TJ., Jan 2019. Can the workload-injury relationship be moderated by improved strength, speed and repeated-sprint qualities?  J Sci Med Sports 22(1):29-34.
21. Cross MJ, Williams S, Trewartha G, et al. The influence of in-season training loads on injury risk in professional rugby union. Int J Sports Physiol Perform 2015.
22. M. Fanchini, E. Rampinini, M. Riggio et al. Despite association, the acute:chronic work load ratio does not predict non-contact injury in elite footballers. Science and Medicine in Football. 
23. Raya- González J et al. Determining the Relationship Between Internal Load Markers and Non-Contact Injuries in Young Elite Soccer Players. International Journal of Sports Physiology and Performance, 2019, 14, 421-425.
24. K. Enright, M. Green, G. Hay, J. J. Malone, 2019. Workload and Injury in Professional Soccer Players: Role of Injury Tissue Type and Injury Severity. Int J Sports Med 2020; 41: 89–97.
25. T. Gabbett, 2018. Debunking the myths about training load, injury and performance: empirical evidence, hot topics and recommendations for practitioners. Br J Sports Med.
26. N. F. N. Bittencourt, W. H. Meeuwisse, L. D. Mendonça, A. Nettel-Aguirre, J. M. Ocarino, S. T. Fonseca. Complex systems approach for sports injuries: moving from risk factor identification to injury pattern recognition—narrative review and new concept. Br J Sports Med 2016;50:1309–1314.
27. P. Philippe, O. Mansi. Non Linearity in the epidemiology of complex health and disease processes. Theoretical Medicine and Bioethics 19: 591–607, 1998.
28. Roald Bahr. Why screening tests to predict injury do not work— and probably never will…: a critical review. Br J Sports Med 2016;50:776–780.