Friday, 4 November 2016

2017 FIXTURE REVIEW

With the recent release of the AFL fixture, it comes time to continue the series of FMI fixture analyses, now in its sixth edition.
As per the last analyses in 2016, 2015, 2014, 2013, and 2012, the system applied looks at each teams ranking points and measures them against their opponents to determine an index. A few other factors are also employed, and this years analysis follows the precedent set in past years.



2016 Post-Analysis Confirmation

As an example of the past analysis and determinations on fixture effects on teams, the North Melbourne of 2016 experience is an ideal determiner of performance of the FMI data.

The FMI analysis in a broken down form as per the post 'Further Fixture Analysis' marked up North Melbourne's 2016 in four slices. Their rounds 1-5 and rounds 6-10 fixture was determined to be the 3rd easiest and 2nd easiest over those respective time periods.
On the back of those 10 games, North won 9.

The second half of the season was analysed over rounds 11-17 and rounds 18-23, withe split fixture for North providing the hardest and 6th hardest calendar.
Through rounds 11-17, where North had the hardest fixture by the FMI test, North won 1 of six games, defeating a disinterested Richmond on a cold Hobart night. In the last 6 rounds, North won 2 only, both against non-top 8 teams. The other 4 losses were all against top 8 teams.

Now, if you think this reflects more on North Melbourne and their list than the FMI rankings and fixture analysis, then fine, lets just call it a coincidence.


What would be worth doing, is a bit more of a detailed review. Another day.



AFL's Fixture 'Balance'

To call the current policy employed by the AFL as a fixture balance mechanism would only draw to attention your lack of thinking on the issue.
The system is purely predicated on creating 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.

Once again, the description and process the handicap system that the AFL work to sits in this blog post, where even Stevo 'gets it'... the handicap of the fixture is spelled out clearly.

There is no balance to the fixture at all.



2nd Tier Media Assessments

The community of data analysers and bloggers out there in the blog-sphere and working for non-traditional media (the now famous '2nd tier media', as described by Knucklehead Newman) often produce better content and data than the 'old' media.
Usually very quickly and for free. With minimal adverts or paywalls. Maximum content on a different plane to what regular media is focussed on. You know... the sort of match reports and 'what the WAGs are wearing' fluff that the little paper specialises in.

Here are some of the alternate analyses of the 2017 fixture for you to also consider.


HPN
1. The team at Hurling People Now have another good and detailed post on their analysis. A different tack is employed, and the fixture analysis that includes the 2016 ladder positions (about 3/4 of the way down the post) shows that the AFL handicap system has worked for about two thirds of the teams. 

And as per last year, they again find GWS as having a hard schedule, and the defeated Grand Finalists ranked down the table.
A fine post and much content from the trade points experts at HPN.


MoS
2. Not referred to in last years post, the 'recognised symbol of footydata analysis excellence', the Matter Of Stats blog, has posted a 2017 fixture analysis as well. And yes if you click that link you can see links to the the previous 2 years analyses as well.

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. A top 6 team in Adelaide is found to have the easiest run of all teams. Also among the top 6, Geelong and the Western Bulldogs are also placed outside the 6 hardest of handicap. Of the bottom 6, Fremantle appear to have the third hardest fixture and three other teams from hat bracket a re also harshly dealt with (Essendon, Carlton and Brisbane). 

The secondary analysis of the 'missing schedule' (the effect of not playing teams twice) shows the AFL going closer to its stated aim. In this study the correlation of intent to outcome is a lot closer with only Adelaide slipping out of its 'top 6' bracket, and Carlton and Fremantle sitting outside their 'bottom 6' bracket. Much like HPN, its a two thirds compliance result.


TWF
3. The third measure for you to also read and consider is that done by Troy over at The Wooden Finger.
Much like the FMI analysis below, its geared primarily around Troys power rankings of teams, and it is also into is sixth season of deployment. I am unsure if there are any longer running blog-based analyses than those by TWF and FMI.

TWF finds an outcome similar to MoS, with Hawthorn at the top of his tree for schedule toughness, and also bookending the other 16 teams, Adelaide with the easiest run. In fact in both MoS and TWF, the same five teams are tipped to finish with the five easiest fixtures, though with some differences in positioning.
At the other end, there is a stronger correlation to HPN with 4 of the 5 teams to have the worst schedules common between the two.



Mainstream Media

As always, the Fairfax assessment is headed by Rohan Connolly and his annual review of the fixture, the short breaks, some kms traveled, and a weighting depending 'interstate' teams played.

And once again comment should be made on adding points to teams according to ladder position and the other weightings. Reversing the ladder position and allocating points is always odd, but given the even-ness of the top 7 and other disparities between teams, its hard to see it being a fair distribution.

For example, how are Fremantle (4 wins) only 1 point distant from Brisbane (3 wins), and yet only 1 point from Gold Coast (6 wins)?

The same can be said of teams who finished higher than North Melbourne. They finished at least 4 wins distant (e.g. West Coast who would have finished 7th post their Elimination final loss), yet are only 1 point different to North.

Maybe using wins next year could improve the analysis. And really why not? Wins are a better reflection when you have teams finishing several games apart, and for finalists you can add in the wins gained there (and another for a bye through to the Prelim).
You know it makes sense!


The eventual output from the Connolly method is as at right, with Essendon, Brisbane and the Gold Coast as some of the bottom dwellers receiving easy calendars. Richmond, Carlton and most surprisingly Port Adelaide also collect a good run in 2017.
Out of last years bottom six, missing the easiest of schedules is Fremantle, who get shuffled up to 11th easiest.

At the other end of the spectrum, there is correlation with the 2nd tier media types, with GWS, Hawthorn, Sydney and the Bulldogs figuring prominently.
There is more disparity with what the AFL was intending to achieve with the top 6 though by this analysis. Of those 6 teams, 4 have the hardest fixtures, but Preliminary finalist Geelong get a nice gift by the AFL with the 7th hardest schedule.

Adelaide get an even better run (as also noted by the '2nd tier' team), and under the Fairfax system, they get the 10th hardest / 9th easiest run.
No excuses Crows!

Faring not well by Rohan's methodology are West Coast and St Kilda. Its a fair enough outcome for the Eagles, though. While not 'top 6' material, they wound up the season in 7th with 1 more home and away win than the eventual Premier, and level on wins with GWS and Adelaide.

St Kilda on the other hand is a different matter. Thy finished in 9th, 4 games behind West Coast and 5 behind Geelong (5 after the H&A season, plus a finals win to the Cats as well), yet end up bumping the Cats down the tough fixture list.
A touch harsh on the Saints by this method.



...And Finally (&c) - FMI.

The interesting thing from the last two years fixture assessments were how similar they were once completed.

In 2015, the summary was
"...all in all, it looks like the AFL have achieved their outcome, so ... well done!"

And from the 2016 analysis, it was thought
"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."

Of course, in both cases, the outcome and KPI was to handicap the fixture such that the higher placed teams on the previous years ladder received a harder set of games, and those finishing lower also received an easier schedule.


But what of the 2017 fixture? This analysis applies the FMI rankings to the fixture and calculating the Tough Draw Index based on looking at the following inputs.
- the opposition strength,
- any home/away bias, and
- expectations of a 'win' for each game,
- an expected end of year rank
- an expected end of season ranking for a perfectly even draw, and the
- difference between those two ranking weights.

Working through the calculations yields the table below for 2017.
FIXTURE BALANCE - 2017
RNKTEAMLAD POSFMI RANK PTSSUM OPP'N PTSHOME / AWAY BONUSTOUGH DRAW INDEX
1GWS413782479204.98
2Sydney/SMFC214282428604.33
3W Bulldogs/FFC11272245191004.10
4Geelong31316246211003.78
5Hawthorn5130524650-3003.03
6West Coast71325242831002.80
7Adelaide6140123973-3001.26
8St Kilda91017247611001.25
9Collingwood121096242792001.13
10Port Adelaide10118223982-200-0.04
11Melbourne111023244090-0.34
12North Melbourne81175237520-0.83
13Fremantle1690224435-100-2.38
14Carlton1488623942200-2.73
15Richmond1397223697100-3.02
16Gold Coast1585624177-100-3.32
17Essendon1869324078200-3.60
18Brisbane Lions1763024202-100-3.67


Its another triumph for the AFL in the off-Broadway and niche entertainment value of handicapping the fixture!

Much like last year, there is only one occurrence where a team is listed as having a harder schedule than their handicap bracket. This year though its a much closer fit with West coast and Adelaide out by a single position, unlike last year where Richmond were 4 places removed from where they should be.

Also worth noting is the decrease in the team spread this year, and in fact year on year since the handicapping has been in play.
The spread has come down to 8.65 for next year (calculated as the absolute difference between #1 ranked GWS on 4.98 and #18 ranked Brisbane at -3.67).
The previous years were:
   2016:   8.90
   2015:   8.98
   2014:   9.15

So in essence, the tweaking of the fixture is improving, and the locking of more teams into the 6 x 3 brackets the AFL sets, as well as to reduce the overall discrepancy over the total competition both highlight this.

As before the only teams outside their brackets are West Coast and Adelaide, with Adelaide getting the better edge of that result.

Otherwise, to find any team 'hard done by' you can only look inside the brackets for oddities.

Take for example North who finish 8th to be 2nd in the mid 6 bracket, and get the easiest of the fixtures for that bracket. A benefit for North there.

On the counterpoint to that, three teams have all suffered similarly. GWS, Collingwood, and Fremantle who all are allocated schedules tougher than their final ladder positions.



So overall, its another win for the AFL in its programme of handicapping its own field. It would make the VRC's handicapper proud.

Further analysis will follow-up this post in due course, including an assessment on the calendar under the FMI system of balancing the fixture, as well as under a 34 round system.

No comments:

Post a Comment