How Close Has the 2017 Season Been?

(This piece originally appeared in edited form in The Guardian as www.theguardian.com/sport/2017/aug/08/a-compelling-2017-afl-season-but-exactly-how-close-has-it-been-so-far)

The 2017 season has been a close one, with any team a genuine chance of dominating and maybe even toppling any other team on a given day. More than once, a team near the foot of the competition ladder has defeated a team near the top, and we sit here at the end of Round 20 with the final 8 far from decided.

But, just how close has it been: quantitatively, what can we say about the closeness of the 2017 home-and-away season in an historical context?

Victory Margins

One obvious overall measure of the closeness of games in a season is the average margin of victory, with closer seasons marked by smaller average victory margins. The average, however, can be distorted by the occasional blowout victory – roughly speaking, in a 200-game season, a 100-point victory adds half a point to the average – so we might also want to have a look at the median victory margin as well, which is defined as the margin exceeded in exactly 50% of the games.

The chart below records the average and median margins in home-and-away games for every season from 1897 onwards.


 What’s immediately apparent from this chart is how aberrant 2017 average and median victory margins have been in the context of recent history. You’d need to go back to 1976 to find a home-and-away season with a lower average margin and to 1998 to find one with a lower median margin (the 2007 season had the same median margin as 2017 has so far).

That said, if we do look back to the 1976 season and earlier, we see that average and median margins similar to or lower than those we’re seeing in 2017 were commonplace. In that sense, 2017 is something of a return to the VFL era, at least in terms of game margins.

Now the relative safety of a lead – and ‘comfortableness’ of a victory, if you like – depends not just on its absolute size. In years where, say, teams only scored an average of 50 or 60 points per game, such as we saw in the early parts of VFL history, a 20-point victory margin would imply a level of dominance much higher than in a year where teams scored an average of 90 points per game, such as they do now.

It’s instructive then to look at average victory margins in the context of average scoring.

In the chart below we do that for each season, colouring each season’s label based on the era to which it belongs, with each era generally of about 20-years duration.

We see then that the 2017 season, whose label is filled with grey, is unusual in the context of the modern era not just in terms of average victory margins, as we’ve already noted, but also, relative to all but the most recent years, in terms of total scores. Across all seasons since 2000, only 2014, 2015 and 2016 have seen lower average total scores per game.

So, although victories have, on average, tended to be narrow this season, they might be considered a little more comfortable than they would have been had they been recorded between 2000 and 2013 or in other seasons where total scoring was higher.

A Closer Look At the Modern Era

Averages and medians provide a useful basis for comparing two distributions such as victory margins across seasons, but they can hide interesting and important differences.

Let’s look then at the detailed distribution of victory margins for each of the seasons in the modern era.
 
In this chart we see the number of occasions on which a particular victory margin was recorded in a home-and-away game during the season in question.

What’s striking about the 2017 distribution in this chart is the relatively small number of very large victories and the relatively large number of very small victories. Both of these characteristics help to lower the average and median margins.

That suggests another basis on which we might legitimately assess the closeness of a season – on the proportion of ‘close games’, which we’ll define as those won by less than 2 goals, and the proportion of ‘blowouts’, which we’ll define as those won by 6 goals or more. Close seasons will have relatively more of the former and less of the latter.

The chart of these proportions for every home-and-away season appears below. Eras are reflected here, as earlier, by label colour.

 
In the context of the modern era then we see than 2017 is marked partly by a relatively high proportion of close games, but far moreso by a relatively low proportion of blowouts. Only slightly more than 1 in 3 games this season has been decided by 6 goals or more, while the average for the period since 2000 is almost ten percentage points higher.

That 1976 season that we noted earlier as the most recent with a lower average margin, achieved that result mostly by having an even smaller proportion of blowouts – under 30% in fact – though it also had a higher proportion of close games than 2017 so far.

Team Statistics

Which teams then have been contributing most to the dearth of blowouts and the plethora of close games in 2017?

The chart below shows the counts for each team of in close games and blowouts, and their records in both types.

 
West Coast have been involved in the highest number of close games this season, winning three and losing six. Adelaide have been involved in fewest, their draw the only game this season for them decided by less than a goal.

Adelaide have also been involved in the highest number of blowouts, racking up a healthy ten and two record in these, while Collingwood and Carlton have been involved in just three blowouts each, Carlton losing all three, and Collingwood winning two and losing one.

Team statistics for a single season, based as they are on a small number of games, are subject to high levels of sample variance, so it’s misleading to use them to claim, for example, that some team or other is “good at winning the close ones”. We can get a better idea of how prone teams are to being involved in close games and blowouts, and how they tend to fare in those games, by expanding our view to take in the entirety of the modern era.

 
The Kangaroos, it turns out, have been involved in the highest proportion of close games during this period, with nearly 1 in 4 of their home-and-away contests having been decided by less than 2 goals. Their record in those games is just marginally better than 50%.

Across the same period, they’ve also been involved in relatively few blowouts, second only in percentage terms to Sydney who’ve played in only 150 of them, winning an impressive 116 or just over 77% of them.

The Swans have fared less well in close games, however, having played 92 of them and winning only 44%, while the Cats have fared far better having won 61 and drawn 5 of their 95 close games – a 67% winning record. No team has a better winning rate in close games over the entire modern era.

So, what do we make of 2017?

As we started by noting, average (and median) victory margins in 2017 have been at historical lows, at least in the context of the last 40 years or so and certainly in the context of the modern era. Total scores though have been down a little as well compared to the early parts of the modern era.

Close games have been a hallmark of 2017, and have occurred at a rate that’s quite high by recent standards, but what’s been even more distinguishing is the relative absence of frequent blowouts. That, more than the preponderance of close games is what’s driven down average victory margins this season.

But, with 27 games still to be played in the home-and-away season and finals places a genuine possibility for at least 13 teams, there’s still a chance for 2017 to offer up a few more nail-biters yet.