Meanwhile, despite the Adelaide Oval redevelopment boost, the AFL has also stagnated recently in terms of this bellwether indicator of the competition's economic health, which the leagues care about greatly.
Alternative systems for setting the season's fixtures to improve attendance have been considered recently by both leagues. However, they persist with the status quo, with each team playing some other teams twice each season and the remaining teams once. I have shown previously, in a 2011 Economic Modelling journal article, just how this affects the fairness of the fixture.
In both cases, the methods used to determine exactly which teams play each other once or twice seem somewhat of a black box. The teams are given some opportunity to request an extra match-up against their favoured rivals, but there is also a fair degree of randomness to the selected league-wide set of combinations.
There is some degree to which the fixture is already skewed to target attendances, because the AFL, for example, always matches local rivals – such as West Coast and Fremantle twice, as well as some high-profile Melbourne-based teams. However, this accounts for but a handful of the 153 possible opponent matchings, so therefore we could go a lot further with this strategy.
More scientific solution
In joint work with Jordi McKenzie (Macquarie University) and Stephan Lenor (Heidelberg University, Germany), we sought to find a superior, more scientific solution. Being economists, we set as the sole objective maximising attendances, simply by way of the best selection of opponents to meet each other a second time, subject to existing constraints.
As a base, we estimated "hedonic" regression models for match attendances, using home-and-away match data from seasons 1997-2010, pre-dating the recent expansion. From this, we calculated a panel of "rivalry effects" – one for each pairing of teams, based on games played between them over the sample period. These effects also account for other game-specific variables influencing attendance, such as weather, round, venue, standings and day.
Finally, we employed sophisticated mathematical optimisation techniques – specifically "binary integer" programs – to match each team to the set of opponents to play twice such that the sum of all rivalry effects is maximised. By extension, aggregate attendance is itself maximised.
Some of the optimal match-ups identified were as predicted – eg, Collingwood-Essendon – though others were surprising, such as North Melbourne-Fremantle. Our methodology is flexible enough to revise the match-ups with newer data – after all, the fixture should evolve over time – as some non-traditional rivalries suddenly emerge and later dissipate (Sydney-West Coast 2005-09; Hawthorn-Geelong 2009-14).
In summary, our optimally engineered fixture is estimated to increase attendances by almost 7 per cent above the current level – matches where teams meet for the second time – and more than 9 per cent better than a totally randomised fixture. All other things equal, this translates into about an extra 100,000 paying spectators a season, based on 2015 figures, and about $3.5 million in greater ticket and complementary revenue.
Furthermore, these additional attendance benefits can be reaped with the league having neither to compromise any other scheduling constraint nor fundamentally change anything, such as the length of the season. It is also politically tractable, because no stakeholder group would be worse off. In fact, teams and their fans would barely notice any difference operationally, other than getting to play their team's "pound-for-pound" preferred opponents more often.
Add to all of this that the smaller-drawing teams would probably benefit from higher attendances more proportionally than the game's economic superpowers – Collingwood's fixture is pretty much attendance-optimised already – thus contributing towards much-needed equalisation measures designed to preserve competitive balance as much as practicable.
Our results have been submitted to AFL executives for consideration, while our research article itself, presented earlier at three universities and a conference in the United States, is due to appear in the June 2016 issue of the academic economics journal Review of Industrial Organisation.
Anticipate same result
While replication of our study for the similarly structured NRL fixtures is still in progress, we anticipate the same result. That is, there is considerable scope for improving fixture policy to increase attendances.
If senior sports administrators care about getting the most out of their competitions' resources, optimal fixture policy reform is a compelling low-adjustment-cost reform to deliver better returns for their constituent teams and fans.
Liam Lenten (Twitter: @llenten) is a senior lecturer in microeconomics at La Trobe University.