With the move into Overs/Unders wagering this season, I've taken a special interest in the history of aggregate scores over the past few weeks. In this blog we'll review that history from a team, era and team pairing viewpoint.
The red line in the middle of each box plot denotes the median total score for games in that era, which we can see rises across the first five eras of football used here, only to decline a little in the most recent era spanning 2000 to 2015. Each box's top and bottom (known as "hinges") mark the upper and lower quartiles - that aggregate score above (or below) which only 25% of games fell. The distance between them is one measure of the spread of aggregate scores in that era, another perspective on which you can obtain from the broadness or narrowness of the violin plot that surrounds the boxplot. For example, you can see that the spread of aggregate scores in the 1960 to 1979 era was much greater than in the current era.
Next, let's introduce teams into the analysis and create a similar chart looking at aggregate score by team by era.
Despite the fact that I've allowed the Total Score axes to vary for each era, at this scale it's difficult to discern other than large differences in team aggregate score profiles within an era. Port Adelaide in the 1980 to 1999 era, for example, was clearly a team involved in generally lower-scoring games, and the same could be said for Fremantle in that same era.
To highlight some of the differences across teams in the current era, let's focus solely on this time period, essentially producing a blown-up version of the bottom strip of the chart above.
I've made two other adjustments in producing this chart, colouring each team's violin plot to assist in its visual delineation, but also ordering the teams based on their median aggregate score. So, for example, the Western Bulldogs appear as the first team on the left of this chart because they have the highest median aggregate score (200 points per game) in this era of all 18 teams. Conversely, Fremantle appear at the far right because they have the lowest median aggregate score of just under 175 points per game, more than 4 goals a game below the Bulldogs' median.
The chart also gives us an indication of how variable have been the aggregate scores in games involving particular teams. Hawthorn, for example, has a smaller spread than, say, Essendon or Richmond, while North Melbourne's spread has been relatively small too.
We'll finish by looking at the average aggregate scores for all pairs of teams across the 2000 to 2015 period, this time in the form a heat map, coloured based on the magnitude of the average aggregate score for that pair.
The Western Bulldogs stand out in this chart too - they're the rightmost column - because of the high average aggregate scores they've produced with almost every team. Only against the Gold Coast, who they've played just 7 times, have they generated an average aggregate score below 175 points per game.
Amongst the pairings that have occurred on a relatively significant number of occasions, the highest averages are for:
- Brisbane Lions v Carlton: 211.2 points per game (25 games)
- Essendon v Western Bulldogs: 210.8 points per game (24 games)
- North Melbourne v Western Bulldogs: 209.5 points per game (24 games)
- Carlton v North Melbourne: 206.5 points per game (21 games)
- West Coast v Western Bulldogs: 205.8 points per game (25 games)
The lowest averages are for:
- Collingwood v Sydney: 168.6 points per game (24 games)
- Richmond v Sydney: 168.5 points per game (25 games)
- Fremantle v St Kilda: 168.4 points per game (25 games)
- Brisbane Lions v Sydney: 168.4 points per game (28 games)
- Sydney v West Coast: 167.3 points per game (26 games)