Anatomy of an upset – how do AFL teams overcome the odds to spring a big win?

This season has seen some pretty surprising upsets as well as a general pattern of unpredictable results. As we’ve said before, there’s a wide spread of non-terrible teams – 17th is currently two games outside the top 8, where last year that gap was 6 games. Most teams this year are capable of causing a boilover. We thought we’d take a look at some of the bigger upsets this year and see if we can identify the sorts of things that lead to or characterise the upset results.

Before we jump in to talking about upsets, here’s our current ratings of each team on strength in the three parts of the ground:

Round 12 ratings

Big movers this week are Essendon and Sydney who both jump two spots from unexpected big wins. At the other end of the scale, North, who looked quite good early despite not getting results, finally drop from the fringes of the eight to the edge of the middle tier of teams.

So, what goes on when a poorly-rated team beats a highly-rated team? This post will make no pretence of trying to predict such upsets, after all, if we could do that they wouldn’t be upsets or we’d be very rich betting on them. What we’ll instead do is see how our ratings can offer some clues as to what happens in some upsets.  We have tried to break down upsets into four broad categories. Our inside-50 based  strength ratings potentially have something to say about the first three categories.

1 – Midfield dominance

Generally, the easiest path to victory is to dominate territory, and in turn dominate the number of opportunities your team has to score at the expense of the opposition.

The Gold Coast win over Geelong was, intriguingly, driven entirely from the midfield. We wrote recently about how the Suns could (could!) make finals, and their midfield starting to put some wins on the board is a big part of why.

We rate Geelong overall as a fairly indifferent midfield, in terms of the team’s ability to secure an inside-50 count advantage, and this was a game where they were comprehensively outplayed in that department. While the Suns aren’t world-beaters in this area, they are better than a lot of the ordinary sides – and not much in raw numbers behind the Cats. In this game Gold Coast had 71 inside-50s to 48, Witts dominated Rhys Stanley in the ruck, and they had a small 40-36 clearance advantage led by David Swallow. They laid more tackles, had a ton more possession including more contested ball. Ablett sent nine kicks inside-50 himself.

Only the Cats’ greater forward efficiency (scoring from 58% of their entries) and a bit of Gold Coast inaccuracy prevented a worse rout.

North Melbourne’s huge win over Adelaide, while a shock in terms of Adelaide’s underperformance, was generated by the predominance of inside-50s. Adelaide’s weakest line is their defence, and whenever teams have been able to pepper them enough, they’ve looked vulnerable. Both North and Adelaide scored from around half their inside-50s (that’s a pretty high level of forward efficiency or defensive weakness) meaning that by dominating the inside-50s, North assured themselves victory. They seem to have achieved that dominance through forced turnovers, centre clearances, and simple Adelaide errors.

We still have to throw our hands up in confusion at why Adelaide’s midfield were so impotent on that blustery Hobart afternoon, particularly in the first quarter, but there’s a limit to how much any statistics can identify things like confidence, effort, and simple cascading team-wide shock. However, we can see from that, and also from Melbourne’s win over the Crows, that when the Crows are beaten in the middle, teams can score against them with relative ease considering what a good side they have mostly been.

2 – Stylistic superiority

This refers to a situation where a lower rated team can use its limited strengths to grind out a surprise victory.

Fremantle lost to Brisbane on the weekend and a big part of why was Brisbane’s ability to actually help out their decent forwardline with good supply – something they have rarely been able to do in 2017 as the worst-rated midfield on inside-50 differentials.

Fremantle, rating stronger overall, had been looked pretty balanced across the board – almost equally below-average in all areas. Brisbane, by contrast, are more lopsided, they have had what we consider a decently efficient forward line, but they have been starved of inside-50s and hammered in defence.

In this game, perhaps aided by the absence of Sandilands and the pain-impaired performance of Fyfe, Brisbane actually found itself winning in the middle and dominating the inside-50s battle handsomely. Their forwardline (including through goals from midfielders) accordingly put on a big score through their greater opportunities.

Sydney and the Bulldogs has been an interesting battle for a couple of years now. Opinions differ on whether this was an upset – ladder position would say so, but some ratings systems had the Swans ahead through either lingering 2016 strength or home ground advantage.

Note that the gap has closed in our ratings as a result of this game. The weight of the win resulted in a major recalibration of ratings for the two teams’ strengths this week. This was particularly in our Bulldogs midfield rating dropping by about 5% after their apparently stronger midfield lost the inside-50 count 71-39. That imbalance could easily insert this match into the “midfield dominance” category, but interesting things also happened at either end.

The matchup looked a contrast to last year’s grand final which pitted a stronger Dogs midfield against a superb Swans defence. Things have changed this year – the Swans are rating as average or below-average across the board, while the Dogs’ already mediocre forward potency has regressed and the midfield struggled more, as the backline takes up more of the slack.

Sydney unexpectedly dominating the midfield battle gave the game a huge margin, and this was the first time Sydney had won that inside-50 count in the last four encounters. However, the Swans also scored more efficiently (48% vs 42%) from the inside-50 opportunities given.

Basically, the Swans’ average forwardline did better against a decent defence than the Dogs’ terrible forwardline did against Sydney’s average defence. The Bulldogs needed to win the inside-50 count by a bit, not just get close, under those circumstances. They did so in the previous three matches and carried the day, but things turned dramatically here.

3 – Incredible efficiency

This refers to a situation where a lopsided team uses its dominance in at least one part of the ground to overwhelm the opponent. Instead of a perfect storm of opposing strengths and weaknesses, the underdog instead relies on their own strength in an area to snatch the win.

Referring back to our strength ratings, Essendon’s poor control of the midfield appears to be the only thing holding them back from the upper reaches of the ladder, as they loom large within both 50 metre arcs (2nd and 3rd for forward and defensive efficiency, respectively). That lack of control means their efficiency usually only has limited impact – an efficient defence under seige still concedes scores, and an efficient forward setup can only do so much with little opportunity.

Essendon have won the inside-50 battle just twice this year – by one over West Coast in Round 9 and by 16 against Port Adelaide last week. In both games Essendon won by more than 10 goals. Essendon v Port was especially polarising, as a Port Adelaide accustomed to winning plenty more inside-50s found itself beaten in that area.

Logic indicated that if the Bombers could get close to breaking even, or better, in the inside-50 battle, then they would win the game. They did more than that. They were also more productive up forward than Port when they got the ball up there, leading to the slaughter. Port Adelaide’s denial of space and movement in transition was going to be crucial, and it completely evaporated.

Whenever a team has a single overarching strength – think Adelaide’s forwardline, Richmond’s defence, Collingwood’s midfield – there would seem to be the potential for these sorts of results of something else gives them a chance to really press that advantage.

4 – Unnerving accuracy

This last one should be fairly self-explanatory, and we have seen a few examples where teams have gotten close but only took the chocolates due to unusual accuracy or inaccuracy relative to expectations. We are big fans of Rob Younger’s Figuring Footy as a guide to these games. Of all upset victories, these are probably the hardest to predict in advance.

On form, Carlton shouldn’t have been close to GWS last week. The Giants are in an injury-weakened state – they have maybe ten players missing from their best-22 and their reserves were pumped by 171 points due to the lack of available players. Despite this, they’d been getting results until they ran into this ragtag group of navy blue misfits.

As impressive as Carlton were in dragging GWS into a close battle, Figuring Footy noted that the Giants would still normally have won from the shots they created. Anyone who watched Toby Greene can attest to this, of course:


Another example of this type of game was West Coast’s dour win over Port Adelaide where West Coast, underdogs going in, needed an accuracy differential to complete their raid on Adelaide Oval.

All-in-all we are no closer to identifying upsets, but we reckon the clubs with lopsided strengths in different parts of the ground might be the ones to watch for unexpected results. Collingwood, Geelong, Essendon and the Bulldogs are among the most lopsided clubs, and perhaps will continue to be among the more unpredictable.

We might also expect more surprises for the Crows and Port Adelaide if they get on the wrong end of an inside-50 count, and even the lowly Lions to snatch another couple of results if they can get the ball forward enough.


One thought on “Anatomy of an upset – how do AFL teams overcome the odds to spring a big win?

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s