Determining relative population-specific acceleration intensity thresholds in soccer using game locomotion data: Validation of a new method using data from male youth elite players

Abstract

In soccer, relative population-specific acceleration intensity thresholds are required to create meaningful activity profiles. These thresholds can be derived from the maximal acceleration-initial running speed (amax-vinit) regression line, whose determination has so far required time-consuming testing. The aims of this study were to introduce a new method for determining population-specific amax-vinit regression lines in soccer using game locomotion data and to assess its validity as a function of the amount of data used. The method accounts for both the amount of data used and the distribution of high-intensity accelerations across the velocity measurement range when identifying maximal accelerations in game locomotion data. This is intended to minimize the risk of selecting submaximal accelerations or predominantly maximal accelerations with a positive random measurement error. Game locomotion data were collected from 55 male youth elite soccer players using a GPS-based tracking system. Multiple population-specific amax-vinit regression lines were determined using locomotion data from one to five games per athlete. Furthermore, each athlete completed an acceleration test to determine his test-based amax-vinit regression line. The mean biases for the regression coefficients (i.e., amax-intercept and slope) were estimated and assessed using standardization and Bayesian analysis. Regression lines based on locomotion data from two or three combined games showed trivial biases for both coefficients. However, due to the large uncertainty in the estimates, the chance of equivalence was only assessed as possibly equivalent. The proposed game-based method represents a viable and easy-to-implement alternative to the test-based method for determining population-specific amax-vinit regression lines in soccer. This simplifies the process of determining relative population-specific acceleration intensity thresholds, which are required for creating meaningful activity profiles.