Quantification of physical match outputs has become a routine practice in most team sports with such data extensively reported in the literature. However, the match to match variability of key physical output metrics in some team sports is poorly understood. Hockey is a popular fast-paced intermittent team sport for which limited data exists with respect to longitudinal analysis of match to match variability and analysis of the influences of situational factors on match outputs and within player variability of physical outputs. Therefore the aims of this thesis were to 1) determine the variability in match outputs in international hockey and 2) if variability exists; determine causes of the variability and the implications for applying physical output data from matches into a training setting for international hockey teams.
Twenty-one male international hockey players were monitored using 10Hz GPS units over the duration of an entire international season, resulting in 339 data sets from 29 international matches. Mean physical outputs were determined for key metrics including absolute and relative measures (m/min) of total distance covered, low (<11km/h), moderate (11.01-19km/h) and high (>19.01km/h) speed distance covered across the whole team and by position (strikers, midfielders and defenders). The within and between player coefficient of variations (CV), smallest worthwhile change (SWC) and likely worthwhile change (LWC) (study 1) and the effect of rank of opposition, result and score margin of matches on physical output and variability of outputs (study 2) were determined and compared.
Significant difference in mean physical output values were observed between strikers, midfielders and defenders for many metrics with strikers reporting the highest relative distance of 139m/min. There were high amounts of within and between player variability (CV: 27.8-42.5%) observed across all positions for high speed running (>19.01km/h). Relative distance covered typically displayed the lowest variability across all positions (CV: 5.5-10.9%). Smallest worthwhile change and likely worthwhile change values were also lower for relative distance (1.1-2.8% and 9.3-11.3%) than for high speed distance (7.2-8.5% and 34.6-49.4%) respectively. When analysed by situational variables there were typically non-significant changes to mean outputs with the exception of relative distance and high speed relative distance for midfielders which was increased by 5.9% and 22.9% respectively when playing high ranked teams over low ranked teams. High speed relative distance for defenders increased for highly ranked vs low ranked teams (30.6%), decreased for a win vs a draw (38.3%) or loss (36.6%) and decreased for a big loss vs a close game (34.5%) or big win (43.1%). Within player variability remained relatively unchanged by situational variance across all positions when compared to overall within player variability.
At moderate and high velocities a large degree of match to match variability exists, whilst for relative distances and low speed distances variability of match to match physical outputs is reduced. There appears to be minimal influence of situational variables on within player variability suggesting that the rank of opposition, result or score line of the match does not consistently have a similar influence over the GPS-derived physical outputs from match to match. Significant differences are evident in physical outputs between positions in hockey, suggesting the need to train each position in relation to the demands of their respective position. However within positions significant differences are not as common when analysed by situational variables. The use of likely worthwhile change values based off a ~10% change in relative distance across all positions may be an appropriate method of observing changes in mean outputs that are ‘likely’ substantial in matches and training.||en_NZ