Sunday, 3 December 2017


The AFL fixture was released a few days back so it's time, once again, to check how the draw stacks up when the FMI rankings are applied. It's the now annual and traditional... and at least sixth time the FMI ranks have been applied to the coming seasons fixture.
All here to assist you the loyal and sole reader, in understanding who gets the armchair ride, and who gets the draw of death.
All the past years analyses are listed here, per year: 20172016201520142013, and 2012.

AFL's Fixture 'Balance'

As per recent rulings, the AFL once again have decided to 'balance' the fixture by grouping teams of similar strength together. Then with the allowance of slotting in a few derby's and other marquee games, they season is set.

Once again, there is limited balance in this method. The system exists to create more games of interest for the broadcast partners. By creating more matches between 'better' teams, they generate better prime time content. This looks on paper better for fans, but the reality is that its about profit.

The handicap system is described in this old blog post, where Stevo gets that its a handicapping

There is no balance to the fixture at all.

2nd Tier Media Assessments

Last season, the FMI assessment of the fixture was joined in the sub-major media blogging world by three other website's and methods of assessing the fixture. As at time of writing, in late 2017, only one of these three has re-joined this discussion.

Currently there is no analysis by HPN, who instead have been working on their own spiffy new website and player assessment tools.

Also, Troy over at WFD has also forsaken fixture analysis, and this is most likely to drifting off in high spirits after his Tigers won that flag thing that people win at the Big Dance... its all a bit confusing.

Thankfully, the doyen over at MoS is still looking at the fixture.

Matter Of Stats
The 'recognised symbol of  excellence' in footy data analysis has posted his 2018 analysis here. Tony over at Matter Of Stats dives deep as always to render a few takes depending on how you slice it

MoS uses not only the traditional 'who does each team play' method to pull out one version of a schedule strength measure, but also looks at the effect of not playing some teams again.

Under the first scenario (about half way down the post), there are some major anomalies to the AFL's intent to generate their handicap.
Placing in the top 6 hardest fixtures are, surprisingly, Fremantle, St Kilda and Essendon.These three all finished 8th or lower, yet appear to have a tough schedule by the MoS analysis.
Sydney, a top 6'er of 2017 seen to shake out with the 15th hardest (i.e. an easy) fixture. Geelong and Adelaide also represented toward the easy schedule spectra. Interesting.

The secondary analysis of the 'missing schedule' (the effect of not playing teams twice) last year showed the AFL going closer to its stated aim. This year though, the 'missing schedule' analysis produces effectively the same outcome as the first analysis - the same teams with the harder fixture, and Sydney sitting with an easy schedule, and again Adelaide and Geelong in proximity.

The three listed above are great resources in the 'second tier' media space - those that blog, pod, tweet or facebook with zero to minimal corporate backing (and dollars).

I guess, sadly (...maybe) joining this tier is former Fairfax stalwart Rohan Connolly. Rohan is now striking out into his own sphere with the Footyology site, and as per the last few years at Fairfax, he has given the Connolly Method another go over the 2018 fixture.

As per last year he rather simply applies points based on ladder position which means Adelaide (15w, 1d) is rated higher than Geelong (15w 1d). Oh yes, finals. So then instead we have North Melbourne rated 2 points higher than Gold Coast (both 6 wins) and 3 points higher than Brisbane (1 win less).

That said, the overall outcomes are interesting.
At the foot of Rohan's post is a neato graphic (which can't be imported here) that finalises his fixture assessment into an easy to digest snack.

On that chart, the hardest schedule resides at the bottom with West Coast snaring the dubious honour, with Adelaide and Richmond in the top 3. As you would expect.
Fourth and fifth are St Kilda (mirroring the earlier listed analysis) and Melbourne.

There are no major surprises at the easy schedule end of the spectrum, though Collingwood with the third easiest should leave no excuses for them this coming season.

FMI Calculations - 2018 Scheduled.

In doing the FMI schedule assessment over the last few years, the outcomes have tended to follow closely the official AFL handicapping process very well. Particularly in the last three seasons.

2015: "...all in all, it looks like the AFL have achieved their outcome, so ... well done!"
2016: "the KPI seems to have been met and met well this year. A few exceptions to the rule, but a decent synergy between the objectives and the outcomes."
2017: "So overall, its another win for the AFL in its programme of handicapping its own field."

So what do you think the outcome of 2018's schedule will be?

Well, once again, applying the FMI rankings of each team to the future schedules of 2018, and making allowances for HGA and other details as mentioned in previous posts, the table below does the talking.
5St Kilda111054
8West Coast61155
10Port Adelaide71221
11W Bulldogs/FFC101089
15Brisbane Lions18766
16North Melbourne15938
18Gold Coast17760

The first point that highlights that the AFL schedulers have achieved (to a very high degree) the handicapping of the draw is to look at the 'Ladder Position' column, All those low ladder positions in the bottom six is a big big tick.
The top 6 from last year at the top of the list is also a tick, but probably needs a 'please show your working out' next to it as well.

Those two teams outside their grouping is a slightly worse result than the situation last year, but still a very acceptable effort in handicapping.

St Kilda fans should feel most aggrieved is the FMI analysis position, and West Coast, Adelaide and Port Adelaide get the best treatment from AFL House. Particularly Adelaide.

So while the handicappers have basically done their job and the fixture is tilted as per preferred scenarios from HQ, this year's fixturing also shows another subtle measure of improvement.
In the past seasons the FMI analysis of the fixtures has  also noted the overall spread of the values in the 'Tough Draw Index'.

Over the past few years the spread has moved as per below.
   2017:    8.65
   2016:    8.90
   2015:    8.98
   2014:    9.15

So that is year-on-year improvement in keeping the scheduled opponents closer.
And the fixture for the coming year has a 'Tough Draw Index' spread of of 5.04, which is a major step change for the efforts of balance.

This step change could be from an effect of the FMI rankings points being closer together this year (by approx 100 points), but the scale of the 'Tough Draw Index' change relative to the scale of the FMI rankings points change suggests there has been more achieved at scheduling level as well.

So with that metric in mind, the 2018 schedule is possibly the best presented by the AFL in the last 5 seasons, as it achieves its primary aim of handicapping, with the added bonus of relative balance as well.

FMI Calculations - 2018 Alternate Scheduling.

Last year, on a separate blog post, the FMI ranking system data was applied to the other FMI proposal of the fixture as per this old post. The basis of that post is that the fixture should be set so that each team plays home and away against teams throughout the full ladder, and not handicapped as per the AFL system.

Under this system, the below home and away pairs are formed in three groups, and the remainder of the 22 round season is played against the other teams outside the group. Travel balances are also made in those non-group games.

2Adelaide6West Coast5Sydney/SMFC4GWS
3Geelong7Port Adelaide8Essendon9Melbourne
4GWS12Hawthorn11St Kilda10W Bulldogs/FFC
5Sydney/SMFC13Collingwood14Fremantle15North Melbourne
6West Coast18Brisbane Lions17Gold Coast16Carlton
7Port Adelaide
10W Bulldogs/FFC
11St Kilda
15North Melbourne
17Gold Coast
18Brisbane Lions

Setting up a schedule to mimic this season proposed above, the below 'Tough Draw Index' ranking of the fixture can be determined.
Immediately the eye is drawn to the 'Ladder Position' part of the table which suggests Collingwood (13th) now have the 4th hardest schedule, or that 4th placed GWS are benefitting from this system with the fifth easiest schedule.

But this is a wrong-headed analysis, because you should be looking at the 'Tough Draw Index'. These numbers say there is not much difference between Collingwood's schedule or any other teams.

3West Coast61155-1000.66
6Port Adelaide71221-1000.54
12Gold Coast17760300-0.18
13St Kilda111054200-0.18
16Brisbane Lions187660-0.37
17W Bulldogs/FFC101089-100-0.63
18North Melbourne15938

Most clubs under this system improve their schedule, with a few getting a slightly harder outcome, but not by any significant standard.

Looking at the spread across the whole 18 teams, it has been reduced considerably. The above FMI analysis of the 2018 schedule has the spread on the 'Tough Draw Index' is 5.04 points.
Under the FMI manifest above, it is brought right down to 2.91 points.

Comparing the same process over the last two seasons, it is clear that a balanced fixture creates

   2017:    8.65 (AFL format)    4.67 (FMI format)
   2016:    8.90 (AFL format)    3.09 (FMI Format)

So under the FMI split of teams, the are three data points that shows a much more balanced fixture than that the AFL has set. The simple rule below produces the best outcome for an 18 team, 22 week competition where you cant play each team home and away.
Split them so that each group is composed of two from the top 6, two from the middle 6 and two from the bottom 6.

Simple. Transparent. Fair. Balanced.

It really is time that the leaders at League HQ to make a move, set a fairer draw and reject input from the broadcasters.

No comments:

Post a comment