Let’s rate team strength (and do a round 6 preview)

For the past few months HPN has been looking at how to evaluate efficiency and movement.  In the last few of our weekly review/previews we have been dropping little nuggets here and there – scoring shots per inside 50, goals conceded per opposition inside 50 – to try to look at the statistical trends of successful (and unsuccessful) teams.

We’ve decided to take a fairly simple and reductionist view, using a small number of ratios to measure strength in each main area of the ground. We’re going this way because our chosen indicators correlate quite well with success over recent years, whereas many other statistical categories seem to be indicators of game style as much as strength. Generally speaking, finalists come from the top end of the ranking for indicators.

In brief, in our ratings we are concerned entirely with comparing a team’s inside 50 counts and inside 50 scoring efficiency with that of their opponents.

First up, here’s the table:

R5 rankings.PNG

Midfield Score

Of the three simple measures, our Midfield Score is the most simple of all. It is purely a team’s inside 50s divided by their opponent’s inside 50s. This tracks how often the midfield is able to put their team in a position to score, and how much they are able to prevent their opposition from doing so. It is a pure measure of movement, of ball position and ultimate opportunity.

Of all of the statistics that are publicly available, this inside 50 ratio appears to correlate with success as much as any other except for actual goalscoring. We investigated adding other factors in with just inside 50s, such as clearances, tackles and hitouts, but this would both dilute the pure measurement of movement and doesn’t seem to add any more clarity around team ability. Some very good teams are pretty indifferent in the clearance or hitout stakes, for instance.

As can be seen below, after goals, inside 50s are the second most influential statistic on team success according to the esteemed MAFL, who as we’ve said before, is very good at this kind of thing:


Of course, it is a misnomer to suggest that only midfields (and midfielders) can contribute inside 50s, however for clarity of labeling we are suggesting that inside 50s can only occur from the “middle of the field” so in a movement sense, they’re reflective of the stage in between scoring and defending scores. This is also reflected in our other measures, which measures the place on the ground and not necessarily the position.

Offensive and Defensive Scores

These two scores look at efficiency once inside 50 – they are goals per inside 50 and scoring shots per inside 50. The logic is straight forward. We’re fairly certain that goals (scoring and stopping them) are the most important thing in football. We also know that lopsided inside 50 counts correlate pretty well with more successful teams. Therefore, the more efficient a team is turning their inside 50s into scoring, the better their forward strength. The less efficient a team’s opponents are with inside 50s, the better that defense.

Note that since it’s goals/scores per inside 50, a lower number is better for Defence and a higher number for Offence is better.

The reason we’ve used both goals and shots in combination is to try to account for some of the luck effects that occur with scoring accuracy. The key data item missing from that is any indicator of scoring opportunity quality or expected accuracy from scoring shots. Not all inside 50s and not all scoring shots are equally advantageous, and we would expect quality teams to gain better quality entries. This data doesn’t exist for the general public such as ourselves, so we can’t adjust forward and defensive strength by the quality of entries and opportunities. So we will just plough ahead with what we have.

We considered incorporating marks inside 50, as this correlates quite well with success for most teams. However, there seems to be a stylistic element to the importance of marks inside 50 – in 2016 Adelaide and the Bulldogs for example score heavily without being great on this measure and in 2015 Geelong did not convert a great marks per inside 50 rate into correspondingly high scoring shots. In the end we wanted to keep things relatively simple and figured that if you’re taking marks inside 50, you’re getting scoring shots anyway.

Opponent adjustment

Given it is early in the season, teams’ schedules have been very uneven. As can be seen in the table, we have therefore adjusted each team’s raw midfield, offensive and defensive ratios by the strength of their opponents in these areas. That is, their Midfield Score (inside 50 ratio) is compared to their opponents midfield. Since the two measures are the converse of each other, the Offensive Score (goals/scores per i50) is matched up with opponents Defense (scores/goals conceded per i50).

For example, we can see that the Gold Coast have faced the weakest roster of opposition midfields (in terms of their inside 50 ratios) and have faced the leakiest defences. Thus the Suns’ Midfield and Forward Scores get scaled down. Conversely, North Melbourne have faced an efficient-scoring set of opponents so their Defence Score is improved.

The final columns convert each measure to a percentage of the league average and finally an all-in rating tells us overall team quality on these three key measures of their ability to do well on these measures which correlate highly with team success.

Currently this set of rankings tells us that the Swans rate the best overall based entirely on their defensive strength (causing inside 50 inefficiency) and midfield strength (getting a lot more inside 50s than their opponent. They’ve been able to compensate for average or below average forward potency through sheer quantity of inside 50s and a stingy defense.

GWS are a suprise packet – even adjusting for schedule, they’re sitting second right now, rating well above average on their ability to defend inside 50s, create their own, and convert them to scores.

Fremantle might be real bad

Whilst no team has started 0-4 in the AFL era and gone on to make finals, no team as good as Fremantle is has been in this position in our opinions.

Not our finest moment. So let’s walk that back a bit. We still stand by our observation that their earlier opponents mostly look fairly strong, but it turns out after a loss to Carlton at home that Fremantle are also bad in absolute terms, not just relative to their early opponents.

Starting a season 0-4 is very bad, starting it 0-5 with three of your best three players out for the foreseeable future is heading towards disastrous. Freo’s next three weeks see them playing Adelaide in Adelaide, a rising GWS at home and Hawthorn in Melbourne. There is very little margin for error left for the Dockers here but even that might be overselling it. Right now perhaps a better goal for Freo might be to avoid the 1928 Hawthorn Piss Poor Performance Plate, or even just avoiding the worst season by a defending minor premier (as outlined by Matt Cowgill for The Arc).

Melbourne are average

Welcome to the annals of the average, Melbourne

Here are some words that long-suffering Melbourne fans have being dying to hear for the last 10 years: the Dees are average. Literally average.

Looking above to our quality ratings, the Demons sit 9th of 18, right in the middle of the pack for their Midfield, Offense and Defense Scores. They’re creating inside 50s, converting them, and restricting opponent inside 50 conversion at roughly middle of the pack levels.

Their weaker set of opponents drags down their midfield rating slightly at this stage, but their extremely (and uncharacteristically for this year across both preseason and the real stuff) poor showing against Essendon is 1/5th of this average currently. If that insipid performance turns out to have been an anomaly, the spoils of being truly average (slipping into 8th) may well be theirs for the taking.

Let that sink in for a second. Rather than bringing up the rear of the competition yet again, MFC has risen up through the dregs of the early season competition and hit a level of glorious mid-table mediocrity that has been long imagined but rarely seen.

Hot Take of the Week

Mason Cox’s debut is more impressive than Jarryd Hayne’s achievements according to Patrick Smith

We hate defending Jarryd Hayne hype here. We initially assessed him as a chance to get onto a practice squad and he exceeded this by playing real games. We then observed the actual reason why he wasn’t playing later in the season wasn’t form or stupid coaches but to do with contract flexibility and practice squad rules. We like to think we’ve been cautious realists on him, at least compared to guff like Patrick’s “He’s doing bugger all, what’s he done? He went over there and fumbled.”

But the thing is, Hayne has been attempting to break into the biggest and most competitive sporting league in the world in a manner unprecedented in that league’s history.

The story of international players in the AFL is a fascinating human interest story, but we have to remember that the reason it happens is to do with the structure and talent pool of our league. We’re a small country with a limited talent pool, and it’s in clubs’ interests to leverage against that by exploiting the vast resources of surplus underpaid athletes in other countries for whom the AFL can offer a competitive payday. The AFL Draft makes it hard to source talent freely in Australia. The salary cap means it’s tough to take punts on too many project players within the cap. The rookie list means international players are a free hit.

Mason Cox is the latest talented prospect in a long line. He’s very large, seems to have a good brain, will probably be successful. But a goal and a dozen hitouts and possessions against a shellshocked and remnant Essendon team isn’t necessarily world-beating. We’re not even sure yet that he’s the most promising American basketballer so far (remember Jason Holmes?) but at any rate, let’s just remember that the AFL is not the NFL.

Things to watch this week

1. GWS v Hawthorn

Hawthorn have been lucky in their close games this year (as opposed to last year when they went 1 and 4 in a performance that was typically meaningless when September rolled around). GWS are shaping as very very good.

By our ratings the Giants are well above average in all areas of the ground, and while Hawthorn look strong in midfield strength and forward efficiency, they’re below average at defending opposition inside 50s. Unless Hawthorn’s midfield can shut GWS down, GWS will get their chances and Hawthorn probably need to improve on the ability they’ve so far shown to stop teams converting their forward forays into scores. This should be a fascinating game to see where these two teams – both hard to get a read on in 2016 – truly sit.

2. North Melbourne v Bulldogs should be about contrasting strengths

In brief terms, North Melbourne’s success seems all about their extremely high efficiency in turning inside 50s into scoring. The Bulldogs’ seems to be all about a dominant midfield and an efficient defence.

We observed earlier in the year that the Bulldogs were conceding a lot of marks per inside 50, but that hasn’t actually translated into conceding a lot of scores from total inside 50s, with the Bulldogs well above average for inside 50 to score conceded ratio. This would suggest that they’re very proficient at defending the ground ball whenever marks aren’t taken, and that marks are crucial to scoring against them.

At Etihad, conditions won’t play a role but the interplay of the Bulldogs’ increasingly undermanned (and already undersized) defence with the North Melbourne tall forwards will probably be crucial. The Bulldogs’ drastically better Midfield Score (ie inside 50 ratio) suggests they can limit North Melbourne’s forward opportunities, while giving the Dogs’ forwards plenty of their own and they’ll need this to happen in order to win.

Look to see if the Bulldogs midfielders can prevent Goldstein’s dominance from translating into forward ball movement. If they can do that, we should see a lopsided inside 50 count but North Melbourne making better use of more limited opportunities.

3. We have no idea what might happen between Richmond and Port Adelaide

Both these teams have been pretty disappointing relative to preseason expectations. Both have a number of injuries, players coming in and out of their teams, and seasons that are quickly becoming unsalvageable. Both have large, angry fanbases demanding answers and change, and whoever loses will likely go into full-blown psychodrama crisis mode. Both teams even have had hot tempers boiling over into brawls and cheap shots.

If you want some numbers to think about, they look pretty similar. Both teams have roughly average forward efficiency and are well below average in both generating inside 50s and defending against scores from them.

However, between the impact of confidence and desperation and the impact of team changes we don’t think this game is going to be won and lost on paper or by anything statistics can measure. It could be a fun and spectacular mess.




Leave a Reply

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

WordPress.com Logo

You are commenting using your WordPress.com 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