Monitoring What Matters: A Systematic Process for Selecting Training-Load Measures

Abstract

Purpose: Numerous derivative measures can be calculated from the simple session rating of perceived exertion (sRPE), a tool for monitoring training loads (eg, acute:chronic workload and cumulative loads). The challenge from a practitioner's perspective is to decide which measures to calculate and monitor in athletes for injury-prevention purposes. The aim of the current study was to outline a systematic process of data reduction and variable selection for such training-load measures. Methods: Training loads were collected from 173 professional rugby union players during the 2013-14 English Premiership season, using the sRPE method, with injuries reported via an established surveillance system. Ten derivative measures of sRPE training load were identified from existing literature and subjected to principal-component analysis. A representative measure from each component was selected by identifying the variable that explained the largest amount of variance in injury risk from univariate generalized linear mixed-effects models. Results: Three principal components were extracted, explaining 57%, 24%, and 9% of the variance. The training-load measures that were highly loaded on component 1 represented measures of the cumulative load placed on players, component 2 was associated with measures of changes in load, and component 3 represented a measure of acute load. Four-week cumulative load, acute:chronic workload, and daily training load were selected as the representative measures for each component. Conclusions: The process outlined in the current study enables practitioners to monitor the most parsimonious set of variables while still retaining the variation and distinct aspects of "load" in the data.

Identification and Preference of Game Styles in LaLiga Associated with Match Outcomes

“…, very long (0.7 ≤ r < 0.9), almost perfect (r ≥ 0.9) and perfect (r = 1). In the present study, only indicators with a value higher than 0.7 (positive or negative) were considered to define this component [23].…”

Organized Chaos in Late Specialization Team Sports: Weekly Training Loads of Elite Adolescent Rugby Union Players

“…Due to methodological and logistical issues (e.g., participant recording failure and equipment malfunction), it was not possible to collect continuous weekly observations which could have been used to calculate acute:chronic workload ratios or exponentially weighted moving averages (19,43). Future research should aim to assess the week-to-week changes in acute internal and external loads of adolescent rugby union players relative to chronic loads (42).…”

Predictive Modelling of Training Loads and Injury in Australian Football

“…Many of the training load variables collected were likely to be correlated, thus our prediction problem may suffer from multi-collinearity. Principal component analysis (PCA) is a dimensionality reduction process that reduces a large number of predictor variables to a smaller number of uncorrelated variables [21] and has been employed in studies of training load monitoring [29]. Each multivariate model was trained with unprocessed data and data pre-processed with PCA.…”