ANZ #Netball Grand Final 2016 (2)



Earlier this week I shared some data from a very exciting ANZ netball grand final between the Queensland Firebirds and the NSW Swifts.
This post provides some additional thoughts about the flow of performance in the game. I draw upon some research in handball and basketball to explore this performance.

Random Walk

It seems strange to be talking about walking during a dynamic grand final. But walking is the start of my thinking.
Should we anticipate that in evenly matched teams that the score will alternate according to centre pass possession?
Champion Data produced this graphic of the difference in goals between the two teams during the course of the game:
Final Scoreline
My reading of this graph is:
The Swifts are clearly ahead in the first quarter of the game and the Firebirds bring the game back in the second quarter. There is a period of stability and close margins. Firebirds start the third quarter with a surge and the Swifts bring it back into a close game. There is another Firebirds surge in the fourth quarter but the Swifts are able to respond and bring the game back to a draw at the end of regular time. Firebirds extend their lead in extra time but once again the Swifts close the gap again until the final two goal margin in the concluding segment of extra, extra time.
Whenever I see a court game where teams have alternate possession (with slight variation in netball), I think back to a paper written by Martin Lames (2006) about handball. In it he discusses the concept of a random walk:

The development of the score in handball may be perceived as two interlaced random walks. Each team has a probability p to score at ball possession, P(1)=p, and a probability of q=1-p not to score, P(0)=q. …It becomes obvious that the processes are dynamic, we have phases where almost each ball possession leads to a goal but we find also periods with no goal scored. In some phases the two teams perform at the same level, in other phases there are differences.

Martin’s use of the concept of a random walk formalises how we, as observers, might understand the “process formed by successive summation of independent, identically distributed random variables” (Gregory Lawler and Vlada Limic, 2010).
More recently, Leto Peel and Aaron Clauset (2015) have discussed scoring trends in basketball. They note that the rules of the game add a non-independent dimension to game playing, namely, “a forced change in ball possession after each scoring event”. This is the case with netball too with some variation.
Leto and Aaron note that a scoring pattern in which “a score by one team is more likely to be followed by a score by their opponent” is described as anti-persistent.  They add that momentum occurs when the leading team has a higher chance of scoring again.
With these thoughts in mind, I had a closer look at anti-persistence in the four quarters of the final:
There are some very clear anti-persistent phases in each quarter. Scores alternate.
There are some good examples of momentum too.
The Swifts’ start to an away grand final:
The Firebirds in the fourth quarter:
The game ends in regular time with a Swifts’ comeback as a momentum shift.
Extra time has it all:
At the end of the extra time there is a turnover of possession whilst the Swifts are attacking. This stops the anti-persistence in scoring. The Firebirds score and have the centre pass after scoring for what will be the final play of the game.

Coaching and Performing

I have really enjoyed working through the data about scoring events in the Grand Final.
I did so with coaches and players in mind.
The patterns of play in the final raise for me some important questions about how we prepare for the intensity of competition. From this final alone, there are some outstanding scenarios to scale for training at any level of competition.
My analysis is a lapsed-time analysis. It has encouraged me to contemplate what messages (if any) I would share with a coach in real time. This would include, I think, any conversation about substituting players.

Recording, Analysing, Sharing

I do see an unequivocal place for analysis and analytics in netball. I am mindful that both are services to coaches and players.
There are some fascinating insights in the literature about scoring patterns in invasive court games. Researchers have tended, to date, to focus on basketball and handball.
This post has looked to extend the conversation to netball. A third post will explore some of the mathematical opportunities we have to consider random walk behaviour in netball.

Photo Credits

#Friendships (Susan Pettit)
Frame Grab (ANZ Championship video)


Lames, M. (2006). Modelling the interaction in game sports–relative phase and moving correlations. Journal of Sports Science and Medicine, 5(4), 556-560.
Lawler, G., & Limic, V. (2010). Random walk: a modern introduction. Cambridge: Cambridge University Press.
Peel, L., & Clauset, A. (2015). Predicting sports scoring dynamics with restoration and anti-persistence. In Data Mining (ICDM), 2015 IEEE International Conference Proceedings (pp. 339-348). IEEE.



Please enter your comment!
Please enter your name here