The expected value of possession in professional rugby league match-play

Abstract

This study estimated the expected point value for starting possessions in different field locations during rugby league match-play and calculated the mean expected points for each subsequent play during the possession. It also examined the origin of tries scored according to the method of gaining possession. Play-by-play data were taken from all 768 regular-season National Rugby League (NRL) matches during 2010-2013. A probabilistic model estimated the expected point outcome based on the net difference in points scored by a team in possession in a given situation. An iterative method was used to approximate the value of each situation based on actual scoring outcomes. Possessions commencing close to the opposition's goal-line had the highest expected point equity, which decreased as the location of the possession moved towards the team's own goal-line. Possessions following an opposition error, penalty or goal-line dropout had the highest likelihood of a try being scored on the set subsequent to their occurrence. In contrast, possessions that follow an opposition completed set or a restart were least likely to result in a try. The expected point values framework from our model has applications for informing playing strategy and assessing individual and team performance in professional rugby league.

History, Evolution and Future of Big Data and Analytics: A Bibliometric Analysis of Its Relationship to Performance in Organizations

“…Chen, Cheng and Hsu, 2013;Song et al, 2013). Studies in cluster seven used big data analytics in sports to analyse the evolution of gameplay in Australian football (Woods, Robertson and Collier, 2017), the rela-tionship between practice and injury in American football (Wilkerson et al, 2016), and the possession value (Kempton, Kennedy and Coutts, 2016) and match demands in rugby football (Hogarth, Burkett and McKean, 2016). Finally, the two studies in cluster eight used machine learning to predict the performance of brain-computer interfaces (Halder et al, 2013;Hammer et al, 2014).…”

Explaining match outcome and ladder position in the National Rugby League using team performance indicators

“…Although relinquishing ball possession, longer kicks push an opposition closer to their goal line. This is an important consideration, as Kempton et al 15 noted that ball possession closer to an opponent's goal line (within 20 m) was likely to increase the likelihood of scoring a try. Given this, it would be of value for future work to examine the placement of kicks performed during game-play, as this may offer a deeper insight into the explicit offensive strategies successful teams implement to optimise their likelihood of scoring.…”

“…Five of the 13 modelled team performance indicators were included within the CI tree(Figure 1), these being 'try assists' (root node), 'all run metres', 'line breaks', 'dummy half runs', and 'offloads'. Nine terminal nodes were grown; numbers4,5,8,9,10,12,15,16, and 17. ****INSERT FIGURE ONE ABOUT HERE****Following the branches to the right of the root node (>2 try assists), node number 11 partitioned the data on 'all run metres' at a count of 1340m.…”

Influence of Physical Characteristics and Match Outcome on Technical Errors During Rugby League Match Play

“…This finding is somewhat surprising, given that the highest number of try's are scored following a turnover, and the likelihood of conceding points increases the closer to the goal line the turnover occurs. 18 However, successful teams have a slightly lower completion rate, compared with less successful teams in their defensive zone. 17 These results may reflect that when a team is winning a game, and there is less "scoreboard pressure," they may be more expansive with the ball, which although may lead to more points being scored, it may also result in a greater number of attacking errors.…”