Thursday, 7 November 2013

The Fix Is In

As we have done in previous years, the coming seasons fixture has been released and needs our review. And for those interested, our previous draw analyses are here... 2013 and 2012.
And we also have a treatise on how the draw should function on a special page "Reform: The Fixture"

Draw Designation
As was the case in 2013, the AFL have made another attempt to balance the draw. Using some complex rules that link clubs in three brackets (top 6, mid 6 and bottom 6) to each other.
While these are functional for the coming season, we wonder what happens when teams are not in the correct brackets? Will the AFL forego its regularly scheduled blockbusters? Or is their enough flexibility to allow them to dance on regardless?
[Hint: the latter... just look at the mandatory derbies twice every year]

Analysis by the AFL's own website writers suggests some level of balance was targeted, yet struggles to find any genuine examples.

It does highlight other 'quirks', such as
- Carlton not leaving Victoria until round 11.
- Hawthorn having nine six-day breaks, and Geelong, Carlton and North Melbourne each having eight.
- The Western Bulldogs closing the season with five games from six at the Docklands.
- St Kilda and North Melbourne playing just twice at the MCG, while the GWS Giants play three games there.
Rather oddly, the article also mentions (as a measure of fairness of the draw) that
"An unprecedented nine Collingwood games will not be shown on free-to-air TV, including five on the trot from rounds 18 to 22."
Which may be a measure related to equality of brand exposure, but perhaps not exactly equality and balance in the fixture.
And cynics (like us at the FootyMaths Institute) may also say shifting Collingwood off free-to-air TV is just a jim-dandy of an idea for Foxtel subscriptions.
You remember Foxtel... those guys who forked out record sums for AFL broadcast rights. And now need to fund that expense.
Yeah. Them.


The Fairfax Treatment
As always there are a myriad of ways the draw can be interpreted.
The Age has again gone with the same system, totaling points based on finishing position, with added penalties for road trips, match turn arounds and travel.

While that method worked to a point in last years system (as per our blog post), this year there are some key anomalies of that systems results versus what the AFL are trying to do.

Working down the list, a top 6 finishing team in Richmond is given the 4th best draw. If the system is right, we should see smiles on Tiger faces.

At the other end, Brisbane cops a big hit. As the 2013 season 12th placed team, to get the 4th toughest draw is a bit unfair.

All that said, the Age system does calculate the best of the draw to the bottom 6 sides, with (lucky) Richmond, Adelaide and North Melbourne mixed in among them.


FootyMaths Rankings
Using our rankings data (calculated for each and every Premiership season game since 1897) we can create a few measures on the draw as below.

System 1: Rating against your opposition
Firstly, we can accumulate the oppositions teams rankings points for each team and order the teams accordingly.
These rankings are based on this calculation [1]:
Sum (Opponents rank) + sum (Home advantage if away) - sum (Home advantage if home)

This gives the table as at right which is a real mixed bag for 2014:

- The easiest of the draw is given to 2 of the sides that finished top 6 last year (yellow highlighted Richmond and Sydney), and one of the sides that finished bottom 6 (grey highlighted Western Bulldogs).

- At the other end of the spectrum, the teams with the toughest draw are mostly non-top 6 teams.


If we use that quite simple calculation to determine if the AFL has scheduled the 2014 season for balance, the only conclusion is that there is no clear pattern or sign of balance.

The next measure takes the above data and refines it further.


System 2: Rating against expectation
As we also used last year, a more accurate reflection of ranking the draw incorporates
- average quality of the opposition
- expectations of a 'win' for each game
- expected end of season ranking and an expected end of season ranking for a perfectly even draw
- difference between those two totals.

This produces the table below.

This method, of considering not just who is played, but where and what likely outcomes are expected produces the handicapped result the AFL was after.

The better of the draw is to the lower placed teams, with the bottom 6 teams from 2013 (shaded grey) all getting the best draw rankings 1 through 6.

At the other end, we see the toughest schedules being handed to 4 of the top 6 teams. Escaping the worst of it are Richmond (who should count themselves lucky to be closer to the easier ledger then the harder side) and Sydney.

Port Adelaide, who starting 2013 so well with 5 straight wins, receive a light draw. As does the other SA team, Adelaide. It will be interesting to see which of these performs better in 2014.

Overall, our rankings data analysis shows there is indeed some degree of separation of clubs based on past performance. The handicapping appears to be right this season.


Closing Comments
The AFL have tried to create a "fairer draw", but we still consider it compromised, to say the least. The attempt is really to reduce the possibility of lop-sided games and also to increase blockbuster games and drive revenue.

True balance comes from thinking beyond the base dollars and fear of embarrassment of score lines.
As we mentioned before, we have already proposed our approach to balancing the competition. It is as fair a system as you can have with 18 teams and 22 games, and the schedule for the following season is transparent and gerrymander free.

[1]: For the sake of determining 'home matches' the below was used:
- any Victorian team playing an interstate team in Melbourne/Geelong is considered a home game.
- any interstate team playing any one outside their state is considered a home game.
- Victorian teams playing against other Vic teams, and the local state derby's are (of course) considered neutral.
- Vic teams that play less than 3 home games in any single non-Victorian location are accorded a neutral status (Example: Melbourne playing in Darwin etc).

- Teams playing 4 or more games in any single location are accorded a home game status (Example: Hawthorn at York Park). With GWS playing 3 games in Canberra, they miss that home status.

No comments:

Post a Comment