The Roar
The Roar

AFL
Advertisement

Forecasting this year's AFL season from following the footy

The Swans head to Melbourne to take on the Bulldogs in a grand final rematch. (AAP Image/Julian Smith)
Roar Guru
1st March, 2017
3

Being a maths geek and an avid footy fan, I’ve kept a myriad of stats on not just Australian but also American and Canadian big-league football, as well as Division 1 college football in the US.

I’ve had good fortune in both the AFL and ESPN tipping competitions over the years, typically placing in the top 5-10 percent.

In the CFL, which I follow in less detail, I’m only in the top third over the last two years. And to prove how useful statistical knowledge is, I’m below average in the AFLW competition this year, since I had no idea ahead of time what the teams would be like!

We use a system like the chess world’s ELO rating system for all of the sports we cover (except college football), and our AFL system accurately picked 143 of 196 games last year (there were two ties in our ratings).

The system’s actually pretty basic in principle: the difference in two team’s ratings (allowing for home field and a couple of other factors like injuries and turnaround time) should predict the final score differential.

Whatever the difference between the prediction and the real outcome turns out to be is divided by eight, added to one rating and subtracted from the other.

Here are the starting ratings for each team for 2016 and coming into 2017.

Sydney Swans – started 2016 season at 63.2; starts 2017 before the JLT series at 81.3.
GWS Giants – from 44.9 to 75.1.
Adelaide Crows – from 61.3 to 73.8.
Western Bulldogs – from 55.9 to 69.9.
Geelong Cats – from 52.3 to 68.1.
West Coast Eagles – from 82.1 to 62.5.
Hawthorn Hawks – from 88.0 to 59.8.
Collingwood Magpies – from 45.8 to 55.5.
Port Adelaide Power – from 63.2 to 54.5.
North Melbourne ‘Roos – from 61.4 to 50.4.
St Kilda Saints – from 28.0 to 48.5.
Melbourne Demons – from 33.8 to 42.6.
Carlton Blues – from 21.1 to 33.3.
Richmond Tigers – from 61.4 to 32.8.
Gold Coast Suns – from 31.1 to 31.3.
Fremantle Dockers – from 57.7 to 30.3.
Essendon Bombers – from 22.3 to 21.4.
Brisbane Lions – from 26.3 to 8.9.

Advertisement

An average rating is 50.0. In fact, with the AFLW starting from scratch this season, we started the eight teams with that fifty rating and have gone from there. As of week four, here’s where the eight teams stand: Adelaide (60.3); Brisbane (55.3); Carlton (54.2); Melbourne (53.0); Collingwood (45.9); Western (45.6); Fremantle (43.6); and GWS (42.1).

At this point, of course, it’s a lousy predictor – the system’s only batting .500 so far. But it’s a long-term system, and it should continue to become more effective as time goes by.

melbourne-demons-afl-womens-2017

We track records back into the early reaches of the VFL, and there are some important patterns that are worth noting if you’re planning on trying to predict the future.

As is often noted, since 2008 there have been either two or three teams that dropped out of the finals. Amazingly, the exact same number has moved into finals! Who’d have thunk it?

What’s less well known was that in the last ten years, six of the 14th-placed finishers jumped all the way into the finals the next year (and St Kilda almost made it last year, missing eighth by percentage).

In comparison, only five ninth-placed teams have made the piddling leap into finals the next year.

Advertisement

Fremantle last year became the first top-of-the-ladder team to miss finals since 1993’s Bombers placed tenth in 1994.

By contrast, the wooden spoon winners have often made finals. West Coast went from 2010’s spoon to fourth in 2011, and in consecutive years, Melbourne in 1997 and Brisbane in 1998 both jumped from 16th into top four appearances the next year.

Richmond almost joined this club, placing last in 2007 before jumping to (of course) ninth the next year.

Here’s something worth tracking for 2017 and beyond, teams that have a higher percentage than their record would predict usually place higher the next year. It happens two out of every three times – and teams that have a lower percentage than their record would predict virtually always place lower the next year (over 85 per cent of the time).

The higher percentage rule is really only effective looking at middle of the pack teams who stray from the even scoring of the 100 per cent mark. Teams at the top or bottom of the ladders are going to vary a great deal more than those in the middle just by nature – that defines their competition more than their own prospects.

But if this theory holds for 2017, Port Adelaide should make finals this year (which I truly don’t think they will. But do I have the guts to go against my own metric?)

Nobody else was significantly off centre in 2016.

Advertisement

The lower percentage rule works all over the place, although when it fails it tends to fail towards the bottom because, let’s face it, they’re already a bad team at that point. If this holds, Hawthorn should drop precipitously (and I think they will – and so did they, hence the radical trades they made in the off-season).

What do I see happening in 2017?
My site’s forecasts don’t stray that far from everyone else’s. It’s pretty easy to see Greater Western Sydney as being the odds-on favourite for the upcoming season. By almost any metric, they’re head-and-shoulders above the competition.

Only injuries or a run like the Western Bulldogs had last finals series should stop them from at least being there at the end.

I’m still thinking the Doggies have the edge over Sydney in balance across the pitch, and the other three teams who have an outside shot at the title in my estimation are Geelong, Adelaide, and maybe the West Coast Eagles.

Their reps as ‘flat track bullies’ is completely deserved, and the only reason I include them is if somehow Sam Mitchell changes that culture from within this season.

I don’t think the Cats have the defenders nor the Crows the midfield depth to defeat enough of the top three to make the grand final.

Rory Sloane Adelaide Crows 2016 AFL

Advertisement

The fight to be knocked out in the first week of finals comes down to St Kilda, Melbourne, Hawthorn, and Port Adelaide. The Saints and Demons are on their way up; both look much better already this summer, and the ‘two-new-teams-in’ rule works to their advantage.

I have more faith in the Hawks than the Power, so I’m placing these teams in the order listed, and when the dust settles I don’t think it’ll be a percentage difference that keeps ninth from eighth this year, unless Hawthorn’s new boys gel faster than expected.

Below that, it’s hard to know what Essendon will look like this year. So far this summer, the signs are good, certainly better than Fremantle, whom the Eagles annihilated on their first hit out, and Collingwood, who beat the Dons only because the latter tired out in the fourth quarter of their first game out.

I was ready to drop both North Melbourne and Gold Coast well below those three, but they’ve both showed strong signs of life this spring. The Kangas have found some youth that seems to work together well. I heard a commentator suggest that getting rid of four legends was the only way they’d be able to force themselves to find out if they had the future here or not, since the vets might never be outplayed.

The Suns picked four usable parts with those top ten draft picks (and the injuries have only hit one Day so far this season). Richmond has looked decent this pre-season and might move up a couple of places if injuries hit the Suns and others.

So, put these six teams bunched up in slots 11-16 in this order: Essendon, Collingwood, Fremantle, Gold Coast, Richmond, and North, based on the quality of teams. When you see the records, you’ll realise how the schedule adjusts that when they actually play the games.

The bottom two are pretty easy. Carlton looks set to take a step back this year through youth, in order to take two steps forward in 2018 and beyond. And even if they take two steps forward with a new coach this year, Brisbane has a long way to go to climb out of the cellar this season.

Advertisement

chris-fagan-brisbane-lions-afl-2016

I played out the season hypothetically and came up with this projection.

GWS (18-4)
Western Bulldogs (17-5)
Adelaide (17-5)
Sydney (16-6; tougher schedule)
Geelong (16-6)
West Coast (14-8)
St Kilda (14-8)
Melbourne (13-9)
Hawthorn (12-10; but they could win 14!)
Port Adelaide (10-12)
Essendon (9-12)
Gold Coast (8-14; injury dependent as always)
Collingwood (7-15)
Richmond (7-15; some winnable games early)
Fremantle (6-16)
North Melbourne (5-17; but they could win more if the new kids gel soon)
Carlton (5-17)
Brisbane (4-18).

My 2017 finals brackets
They should run to form: GWS and Western will host and beat Sydney and Adelaide respectively in two exciting qualifying finals; Geelong and West Coast eliminate the two newbies in Melbourne and St Kilda respectively, just on the experience factor alone.

Second weekend, it would be West Coast going to Adelaide and getting beaten, and Geelong going to Sydney and… well, I could see that one going either way, frankly, depending on the health of each team.

Third weekend, the big question becomes whether the top two teams (Giants and Bulldogs) get rusty having played just once in four weeks? If not, it’ll be a GWS/Bulldogs final. They should be the class of the AFL this year.

Who wins the flag? If it’s close towards the end, I’d lean to the defending champs who’ve been there before.

Advertisement

If GWS can get a three-score lead going into the fourth quarter, as they should, they’ll definitely win.

Any lead and they should win, but I wouldn’t feel good about being in a close game with the Bulldogs after last year’s September magic!

close