How will the top eight look come September?

By Gordon P Smith / Roar Guru

The two longest finals participation streaks in the AFL belong to the Hawthorn Hawks and the Sydney Swans, who have each competed in the eight-team finals bracket for seven consecutive years, from 2010-2016.

Presumably, sitting at a combined 3-11, those streaks will likely end this year.

The next longest streak belongs to North Melbourne, with three straight finals appearances in 2014-16.

Those Kangaroos are 14th at 2-5 and presumably making other plans for September.

The upshot of this is that the closest thing to a ‘veteran’ club in the 2017 finals might very well be Adelaide, or the Bulldogs, or West Coast, with two consecutive berths under their belts.

Geelong and GWS will make their second straight this year, and three ‘new’ teams will probably join them, very possibly Port Adelaide and Richmond being two of them.

Neither team is a stranger to September action, although September winning might be more foreign to these players. The two pre-season darlings, Melbourne and Saint Kilda, are still very much in the hunt at 4-3, and while finals would be a novel experience in either club’s recent history, they each carry players on their roster who can share their background with premiership teams.

Round 7, when finalists often solidify, is a good time to glance at some meta-numbers for each of these nine clubs.

As it has for several years now, my ‘Spreadsheet of Death’ is tracking basic stats, our own ‘ELO-Following Football’ rating system, and trends of both reality and perception.

This means watching the tipping change gradually as the punters and the casual fan gradually realise that Hawthorn 2017 is not the same as the three-time premiers, that Richmond and Adelaide really are that good – and that Sydney really does have some issues to deal with.

In addition, the spreadsheet tracks what I call the ‘meta-Brownlow’, the Player of the Year as determined by accumulation of as many other ‘PotY’ and ‘team of the week’ and ‘3-2-1’ rating systems as I can track: the AFL itself, best and fairest, the game coverage, The Age, the AFLCA and AFLPA, SuperCoach and other fantasy ratings, and yes, The Roar.

Besides following individual point totals, it’s interesting to see how the team totals of these votes compare to the ratings and ladder positions: often, they’ve shown a possible future trend that hasn’t manifested yet in the win-loss columns.

Adelaide: In six weeks, the Crows moved from an ELO-FF rating of 73.5 (50 is average) to 90.0, the highest since Sydney before the 2016 GF – and then dropped 12 of those rating points on Saturday, most of them during that surreal 64-0 first quarter.

Their cumulative player vote total is also first place, as is their point differential in reality and predicted through ratings, so they should maintain a solid grasp on the ladder’s top rung. Rory Sloane sits first overall with 255 points in our meta-Brownlow race, and Rory Laird is ninth, with Taylor Walker, Eddie Betts, and Sam Jacobs all in our top 25.

Geelong: Why are we still surprised when the Cats win? Perhaps because we’ve long expected them to do what the Hawks and Swans did this year. Their rating was third two weeks ago; now, though, they’ve dropped 16 points in two weeks, sit just seventh, and lie 12 points below where they ended 2016. But their record still has them holding a double chance right now, and their player vote totals sit third overall.

How do they do this? Mirrors? Well, yes, if Paddy Dangerfield and Joel Selwood are mirror-images of each other.

At fourth and fifth overall, they’ve carried this club for a year and seven games: only Mitch Duncan joins them in the top seventy in our meta-Brownlow at 38th.

GWS: I love Ryan Buckland‘s ‘the team who shall not be named‘ moniker for the Giants.

Not even in the Hawks’ recent heyday, or the early 2000s when you just knew the Lions would find a way, did we ever have that feeling of inevitability that the Giants give you in 2017, even after the hiccup in week one, and the injury-pressed loss.

Most punters have them winning every game, and nine players in our meta-Brownlow top 80 explains why a real Brownlow winner won’t come from western Sydney: too many players split their votes too many ways. (Toby Greene, Jeremy Cameron, and Josh Kelly all sit within a few points of each other in our top 30.)

Melbourne: They started the season with a rating of 43.8; currently, it’s 43.9. All of their prediction numbers are spot on their actual record. But they’re only 2-5 against-the-spread, and Clayton Oliver is the highest-ranked player in our meta-Brownlow, barely top 20, with nobody else top 70.

With their overall player vote total at eighth, and currently sitting on the deadly ninth rung, losing winnable games against Geelong, Freo, and Hawthorn means that while they’re close to being 7-0, their margin for error is (ahem) Gawn.

Port Adelaide: Before Saturday, their two losses were to the two best teams in the comp, which is why they’re still a clear third on our rating table. Last week’s 76.0 was their highest rating in three years.

Bettors haven’t figured out how strong this team is, or just don’t believe it yet. But all the signs are there, and Ollie Wines is holding the largest sign, currently in sixth place at 146 points.

With Robbie Grey back in the top ten again, and about a dozen in the top 125, they’ve got a strong team that could earn a double chance this spring.

Richmond: The Tigers are for real. They are not going to win it all – the Crows showed us why – but Dustin Martin (second in our meta-Brownlow), Alex Rance, Trent Cotchin, and Jack Riewoldt are all top 40 players; Dion Prestia, Toby Nankervis, and Shaun Grigg lead a crew of the second tier all playing top 100 footy, and the team is still 15 rating points up from the start of 2017.

Except for a trip to Spotless in Round 9, they should be favoured in every game until July. This might be the year they win a final, especially if they can finish top six.

St Kilda: Their rating had dropped from an average 49.2 to a below-average 45.5 until the Trampling in Tassie; now, they’re fourth. No player was in the top 40 of our meta-Brownlow count, not even Sir Nick Riewoldt, until Hawthorn; now Riewoldt, Jack Steven, and the surging Dylan Roberton all are. While the writers kept picking the Saints, the punters had jumped off the band wagon.

Whether they return depends on the next month: will they beat the teams they should beat (Carlton and the Kangas the next two weeks), and split the next two with Western and the Crows? Or go 1-3 and fall out of a tight pack in finals contention?

West Coast: Elliot Yeo leads the way and sits third on our meta-Brownlow ladder with 157 points, but he, Luke Shuey, and Josh Kennedy can’t do it all. (Sometimes, Jeremy McGovern has to help.) They’re 15-4 in their last nineteen: three losses at the MCG and one where they stepped in front of a rampaging September legend about to be written.

Their only road games before July are Essendon and Gold Coast, by which time they might be 11-2. Even nine or ten wins sets them up for finals. (And then?)

Western Bulldogs: We all had them top three this season (alongside the Swans – whoops – and the Giants). Are they a top-three team? They’re currently in a pack at 5-2, but their rating is sixth at 61.2 with the Eagles, Saints, Power, Giants and Crows above them. In player voting, they wallow in ninth. Marcus Bontempelli is doing his part, in 11th place in the meta-Brownlow, but it may be a bad sign that next comes a defenceman (Jason Johannisen).

If they come through trips to West Coast, Geelong, and essentially a neutral game with the Saints with three wins and reach the bye at 8-2, they’re golden. Go 2-1, and they’ll make finals. Lose two, and it’s uphill.

Quickly looking at the other nine, Brisbane started a distant last in our ratings at 9.5, and is back down to a 9.1, still last. Carlton, on the other hand, is up five to 34.2, and might top last year’s seven wins.

Collingwood’s numbers drives home the theme of the Magpie season: “if only they could kick straight!” They’ve kicked 75.96; had they flipped goals and behinds in Rounds 4-5, they’d have a winning record.

I expected Essendon’s ELO-FF rating to continue to drift upwards towards 40, but they’re back down to 25, barely above their start value of 22.

Fremantle may drift up towards average but won’t surpass 50. Nat Fyfe and David Mundy sit top 20, but no one else is top 50. Gold Coast’s rating is higher than both, their player vote totals exceed both and sit them in 10th overall.

They’ve been consistently undervalued by punters this season, but if they can avoid another plague of injuries, anything’s possible.

At one point, Hawthorn had no projected Brownlow votes through four rounds for any player, an unheard-of ‘achievement’. They still have no players in our meta-Brownlow top 40, with Tom Mitchell the closest at 71 points.

Their rating fell by half to below 30 after Round 4, and still sits at 29.9. At least the Kangaroos were expected to lose this year but they’re back in the ELO-FF top eight at 57.5, above Richmond and Melbourne.

Sydney’s Big Four last year (Franklin, Hannebury, Kennedy and Parker) were all top 20 in the meta-Brownlow last season; right now, none are even top 35. They’ve lost 36 rating points since 2016.

The Wanderings expects to see Adelaide and GWS on September 30th, with Geelong and Port reaching Prelim Finals. Richmond, West Coast, and Western also seem likely finalists, and if the Saints bring the bettors to their side, they’ll play in September as well.

The Crowd Says:

2017-05-10T05:31:19+00:00

Don Freo

Guest


Marto, Marto, Marto... I'd reject every one of those predictions...maybe the Dogs will be there somewhere. Saints will struggle to make the 8. Tiges will just plod around 6 to 12. Suns are quite good. Suns and Freo are fielding fit stars again. Freo will be top 4.

2017-05-10T04:07:23+00:00

Marto

Guest


Put a line through GWS , Adelaide, Port, West Coast for the GF win... It`ll be a Melbourne team that wins it.. Again... .FREO SWANS SUNS LIONS are garbage. The rest choke away in big games or are too young... DOGS ST KILDA will win it..Maybe TIGERS

2017-05-09T05:46:48+00:00

Pope Paul VII

Guest


Now I'm just hurt on behalf of 80s legends like Kym Hodgeman and Peter Jonas.

2017-05-09T04:29:57+00:00

Dalgety Carrington

Roar Guru


Accompanied by a fading relevance in the 80's.

2017-05-09T04:23:07+00:00

Dalgety Carrington

Roar Guru


Probably not one complimentary to umpires or the MRP.

2017-05-09T04:19:34+00:00

Dalgety Carrington

Roar Guru


Yeah, I get the idea of the process. I started to put together a bit of a beast of a database a couple of years ago, trying to equip it with predictive power, essentially trying to draw team performance across a number of metrics, including key personnel available. It all came to a bit of a screeching halt as the scope/ambition of it went beyond my time capacity and available data precision. Getting more details on how you arrive at the final figure would be nice though in engaging with the findings.

2017-05-09T04:12:37+00:00

dontknowmuchaboutfootball

Guest


I'm intrigued to see what kind of analysis we might get using an NWA-following football rating system.

AUTHOR

2017-05-09T04:01:44+00:00

Gordon P Smith

Roar Guru


(I know. I feel particularly stupid because it's RIGHT in the spreadsheet, and it took me a year of watching him play to get it right originally!)

AUTHOR

2017-05-09T03:59:44+00:00

Gordon P Smith

Roar Guru


Dalgety, the system regresses the rating towards the norm between seasons by 30%, which just seems to be appropriate after experimentation over the last several seasons. But with particular care towards Essendon this season, I was watching to see if in a special case like that, they needed a higher regression. As stated (amidst all the spelling and numerical errors I let creep into the article - gaaaa!), I expected the Dons to move up quickly this season, but since Round 1's win over what was still an over-rated Hawks team, they've actually dropped a couple of points. To give you an idea, GWS was obviously very lowly-rated until 2015, and even at the end of that 11-11 season, their rating was only 35. During 2016, they started at just under 40 and reached 78.2 by round 9. They finished the year just over 80, regressed to 75.8 to start 2017, and are now at 76.3. Hawthorn fell from a high of 89 in early-mid 2016, ended the season in the high sixties, and are already below 30. Wish I could tell you the system has more behind it than just comparison testing over the years, but that's the joy of computers: test lots of versions and keep the one that best fits the data. It generally matches but doesn't beat the best of the public pundits over the course of the season. But no, Port's rating has no portion from its hot start three years ago at this point, for example.

AUTHOR

2017-05-09T03:47:01+00:00

Gordon P Smith

Roar Guru


Hmm. You are absolutely right, both of you. That's just a screw up.

2017-05-09T02:46:14+00:00

Pope Paul VII

Guest


North have retro progressive rocky history on their side Dal. In 1975 they started 0-4 and didn't enter the top 5 until rd 14 at 7 - 7 before winning the thing. Starting similarly here, they could do it again if they master the tricky bit of winning enough games.

2017-05-09T01:08:41+00:00

Dalgety Carrington

Roar Guru


I may have missed it somewhere else, but I presume your system isn't progressive with retro/homage sound having its heyday in the late seventies and early eighties?

2017-05-09T00:25:44+00:00

gyfox

Guest


Gordon........it's Robbie GRAY!

2017-05-09T00:04:06+00:00

Stewie

Guest


Kind of agree with that first bit Don. Footy analysis HAS to use stats and data. It's how they are interpreted, manipulated, and how in depth you go which determines the quality of analysis. Where most people stop is those basic measures (i50s, scoring shots, possession numbers). Cam's great article looks a bit deeper into the quality of those inside 50s. It's annoying that the AFL doesn't release those "advanced" stats to the general public for free. It just promotes dumb arguments about free kick counts. They can learn a lot from the NBA on that.

2017-05-08T23:15:14+00:00

Cat

Roar Guru


I think Paul was refering to: 'Melbourne and Saint Kilda, are still very much in the hunt at 4-3'

2017-05-08T22:25:11+00:00

Don Freo

Guest


Compare all this reference to ratings numbers with Cam Rose's "Port Should Have Won" article...Stats versus footy analysis. When you say Yeo, Shuey and Kennedy can't do it all, you make no mention of their consistent gems who are doing just as much. Sheppard, Barrass, Hurn, Masten ( yes, Masten) are doing plenty as well. That's your reason for ranking them beliw Port. There is no way Fyfe and Mundy should rank ahead of Hill, Hill, Sandi, Blakely or Neale at Freo based on this season's performances so mentioning Fyfe and Mundy as their only significant rankees makes no point.

AUTHOR

2017-05-08T22:23:58+00:00

Gordon P Smith

Roar Guru


Sorry - I was writing in American, and it didn't translate to actual English accurately. (Mostly sarcastic. We do abuse the language terribly.) Dan, my error for saying "next" in there; thanks for pointing that out... Paul, while I didn't say 3-4, I see why you thought I implied that with the 7-0 referent. Thanks for catching that.

2017-05-08T21:37:07+00:00

Paul Thompson

Guest


And also Melbourne are 3 - 4,not 4 - 3

2017-05-08T21:17:54+00:00

Dan

Roar Rookie


St.Kilda have Carlton and Sydney at Etihad next 2 weeks Gordon, not Carlton and North. The North game is after Adelaide

Read more at The Roar