Table 1.
Presentation of variables used in the automatic algorithm.
Fig 1.
Number of players on the court during a game.
The pregame is in yellow, the halftime break is in blue and the postgame is in green. Yellow arrows show time-outs periods.
Table 2.
Conditions to define starting and ending time of pregame, game or halftime.
Fig 2.
Average Y position of players on the bench.
First half the bench was positioned in negative Y and second half in positive Y.
Fig 3.
An illustration of the average Y position of players on the court at the beginning of a match.
The defensive phases are shown in yellow, for negative average Y values, because the bench is positioned in negative Y (see Fig 2) and the offensive phases are in the purple zone (positive average Y values).
Table 3.
Threshold for defensive specialisation.
Table 4.
Quality of the data validation set recognition for handball phases.
Fig 4.
Percentage of time spent in each displacement direction by position.
Fig 5.
Boxplot of normalised distance per position.
Highly significant differences between positions (p < .001) except Goalkeeper and Pos3_Def where it is significant at p <0.05.
Fig 6.
Boxplot of normalised Accel’Rate per position.
Strong significant differences between positions (p < .01) except and Backs_Both—Pivot_Both where it is significant at p <0.05.
Fig 7.
Mean ± SD of time spent per displacements direction (in %) for each game phase.
Fig 8.
Boxplot of normalised distance per game phase.
Highly significant differences between groups (p < .001) except Def_transition and Off_transition where it is significant at p <0.05.
Fig 9.
Boxplot of normalised Accel’Rate per game phase.
Highly significant differences between groups (p < .001).
Fig 10.
Plots of time played per position/phase of play with 95% confidence interval.
Fig 11.
Plots of normalised distance per position/phase of play with 95% confidence interval.