Off to Merici

On Monday, 14 May, I have an opportunity to visit Merici College, Canberra at the invitation of Dr Kathryn White.

Kathryn is a science and engineering teacher and runs a series of visiting speakers for Merici students to share with them current Australian science research.

My topic will be performance analysis in sport. I am using only data from women’s sport for this presentation. Given the school’s connection with Australia rules football, I have designed some activities linked to AFLW data.

I am hoping that this can be a flipped, BYOD event. I have prepared a Google Slides presentation with links to the AFWL 2018 season data.

The Google Slides presentation.

There is a Speaker Deck version of the talk too.

 I hope to connect any students keen to develop their analysis skills with Jocelyn Mara at the University of Canberra.

Chance and goal scoring in association football

Martin Lames (2018) observed in a recent paper, Chance involvement in goal scoring in football – an empirical approach:

If one accepts the selected chance variables as indicators of uncontrollable or not planned aspects of scoring, the results of this study can be taken as empirical proof for the existence and widespread prevalence of chance involvement in goal scoring in football with many interesting consequences.

Martin looked at goals scored in the Bundesliga (n=875) and the Premier League (n=1056) in the 2011-2012 season. Martin defined chance in goal scoring “as goals involving uncontrollable or not planned aspects”. His analysis led him to identify six ‘chance variables’ (and their frequency):

  • defence involvement (22.5%)
  • long distance shots (13.1%)
  • rebound (9.5%)
  • post/bar (5.2%)
  • deflection (4.8%)
  • goalkeeper involvement (2.5%)

Martin undertook a lapsed-time analysis of each goal scored from video recordings of the goals in the two leagues in the 2011-2012 season.

The analysis was contextualised:

In order to study possible influences on the rate of chance goals, additional variables that are frequently cited as influencing match performance were collected per goal. The game situation was classified in both open play and set plays such as corner, free kick and penalty. Time of the match was characterized with the six 15 min intervals and stoppage-time. Standings were recorded from the perspective of the scoring team and with the number of goals scored so far. The location of a match (home or away) was registered, as well as the ranking of the goal scoring and receiving teams in the season’s final table.

Martin noted that “the global chance score for the Bundesliga was 46.9% and for the Premier League 47.2%.

In his conclusion, Martin noted “chance is associated with neither time-line nor home advantage”.

He adds:

The involvement of chance in constitutive aspects of the game, creating chances and scoring, may be assumed to be typical for the game rather than exceptional. (My emphasis)

This quote took me back to Charles Reep and Bernard Benjamin’s (1968: 585) conclusion:

An excess of shots by one team does not mean that, by chance, the other side will not get more goals and thus win the match. All this is so far removed from current soccer beliefs and tactics that general acceptance of the random element has been inhibited (though one of us, C. R., has shown that a successful style of play can be built upon it). It seems, however, that chance does dominate the game and probably most similar ball games. (My emphasis)

Martin includes Charles and Bernard’s paper in his literature review as a ‘classic study’. Amongst other references, Martin cites his own early work in this area (1999) and the follow up to the 1968 paper by Charles Reep, Richard Pollard and Bernard Benjamin (1971). In that latter paper, the authors note “since the chance of a goal being scored is not constant throughout a match, but is affected by prior scoring and other factors …” (1971: 624)

I D Hill (1974) also looked at chance in association football. He observed in his introduction:

I find it difficult to imagine that anyone, who had ever watched a football match, could reach the conclusion that the game was either all skill or all chance. That both skill and chance are involved seems too obvious. (1974: 203)

He concludes “using the classical significance test approach an inference can be made that skill does play some part in football”. (1974: 204) (See also, M J Maher’s (1982) analysis of a season.)

I am delighted that Martin has refocused discussions about chance in association football. There is a rich literature to explore that invites us to consider within game and between game ‘chance variables’ as they apply to individual teams as well as leagues as a whole.

Photo Credits

Strafraumszene (Mightymightymaze, CC BY-NC 2.0)

Malouda (Ben Sutherland, CC BY 2.0)

Commonwealth Games 2018 Netball Tournament (#GC2018Netball)

There were 38 games played in the 2018 Commonwealth Games Netball Tournament. Just one of these games, Scotland v Barbados, required extra time to determine the winner of the game.

A box plot (using BoxPlotR) of the 37 games played in regular time is:

Centre lines show the medians; box limits indicate the 25th and 75th percentiles as determined by R software; whiskers extend 1.5 times the interquartile range from the 25th and 75th percentiles, outliers are represented by dots; width of the boxes is proportional to the square root of the sample size. n = 37 sample points. Winning teams are shown in light green, losing teams in light blue.

Overall, the median profiles and ranges in the 37 games were:

Photo Credit

Scotland v Barbados (Aaron Hurle, Twitter)