Sunday, July 29, 2018

AFL Power Rankings: Round 19 2018


I feel like taking the week off writing something detailed this week.

An odd thing I noticed about the rankings though: nearly all of the Victorian teams are rated higher by my rankings than they are on the current AFL ladder. Which means of course that nearly all of the non-Victorian teams are rated lower.

Three of the top five sides on the current ladder are non-Victorian: West Coast (second), Port Adelaide (fourth), and GWS (fifth). On my rankings all of the top five sides are Victorian.

There's no Victorian bias in the rankings formula! (At least not intentionally …) And in some areas of the rankings the teams are all pretty close together anyway. But maybe the rankings suggest that teams like Melbourne, Hawthorn, and Geelong - currently filling out the bottom three spots in the top eight - are more within striking distance than you may think.

Sunday, July 22, 2018

AFL Power Rankings: Round 18 2018

The Sydney Swans – consistently a good team in recent years, and one of the best teams coming into the 2018 season – are in a spot of bother.


The Sydney Swans have been a good team for most of the past fifteen years. So if they have a bad few weeks, there is a fair amount of faith they will come back. Even when they lost their first six matches last season they were still considered a decent side. They proved to be even better than that – ending up as one of the best teams in 2017 – but coming from so far back made it a hard path to the premiership.

This year the Swans have been generally good again. However some worrying signs are starting to emerge for one of the AFL’s most reliably high-performing sides.

The Swans had an absolute shocker this week … but maybe we shouldn’t have been quite as surprised as we were

On the weekend, the Swans lost by four goals to the very lowly-ranked Gold Coast Suns. After adjusting for Sydney’s home ground ‘advantage’ (we’ll get to this) and the strength of the opposition, this is estimated to be the equivalent of an 11 goal loss against an average side.

That is the Swans’ worst ‘adjusted net margin’ since losing to Melbourne by 73 points in Round 17 2010. In terms of the rankings it puts them now just above average, and in eighth place.


Sydney has lost the most ranking points of any team in the past five matches, although until this week’s disaster they had still been going OK in terms of their margins. However in two of those losses – against Richmond, and Geelong – the Swans’ relative accuracy in front of goal may have made things look closer than they were.

Sydney kicked 11.1 against Richmond, meaning they had 12 scoring shots to the Tigers’ 23, and they got the ball inside 50 only 70 per cent as much as the Tigers did. Against Geelong the Cats kicked a highly inaccurate 8.23, meaning they had 31 scoring shots to the Swans’ 14. The Swans got the ball inside 50 less than 80 per cent as much as Geelong did.

Hence if the rankings were based on scoring shots rather than margins we would have already seen Sydney start to dip. On Matter of Stats’ rating system based on scoring shots the Swans were already considered just an average team going into this round.

Away sweet away    

An unusual aspect of the Swans’ 2018 season is that they have been really good away from their home ground, and pretty bad on it. In terms of wins and losses they have three wins and five losses at home (not including their ‘neutral’ match against GWS), and seven wins and one loss when playing away.

In terms of net margins adjusted for home ground advantage and opposition strength, Sydney has an average adjusted net margin of -17 at home (see chart below), which includes losses to Gold Coast and Adelaide. Playing away their average adjusted net margin is +29, including wins against West Coast, Geelong, and Hawthorn. It’s all a bit weird … (I’m sure someone can try to explain what is going on, but I’m not going to attempt it here.)



Given the Swans’ horrible home and great away record I’ve seen some Sydney supporters suggest that they are better off finishing seventh or eighth, and then trying to win four away finals to win the premiership. And indeed Sydney has been a very good team ‘on the road’ this year. However, as last year demonstrated for the Swans, even for a very good team having to win four straight finals is not the ideal path.

Sunday, July 15, 2018

AFL Power Rankings: Round 17 2018

The Brisbane Lions have not only been the most improved side in the AFL over the past few weeks, they may also have been its best.


It has been a good few weeks for the Brisbane Lions. The Lions have been a below average to relatively poor side for almost a decade. On the weekend however they beat finals aspirant Hawthorn by over five goals in Tasmania, where the Hawks have been highly successful. The week before they thrashed Carlton by 65 points, this in turn followed a 55 point beating of Fremantle in Perth.



Even after adjusting for the relative weakness of their opponents the Lions have the highest total adjusted net margins over the past few weeks (see chart above). If this was the start of the season people may well have been falling over themselves to declare Brisbane the big improvers of 2018.

Brisbane has still been below average across 2018 as a whole. Their three weeks directly preceding this ‘hot’ streak were poor, as they were comfortably beaten by Essendon and GWS at home, and well beaten by North Melbourne. The Lions also got destroyed by Richmond earlier in the season (a match in which they kicked only two goals), were easily beaten by St. Kilda, and lost to the bottom-ranked Gold Coast Suns. On the bright side, they did have another strong win against Hawthorn earlier in the year.

The Lions’ improvement has been largely driven by their forward and defensive efficiency. They have amassed 51 goals over the past few weeks and conceded only 25 goals, despite getting the ball inside 50 about as much as their opponents (see table below). AFL teams in 2018 have averaged about 1.5 points per inside 50. Over Round 15 to 17 Brisbane has scored an astonishing 2.25 points per inside 50, and conceded only 1.20 points per inside 50 against.


The Lions’ defensive performances these past few weeks is even more amazing when you consider that their best defender, Harris Andrews, has been missing through injury. Andrews’ absence has been ably and surprisingly covered by Josh Walker. Despite missing the past three matches Andrews still has the league’s second-highest number of ‘one-percenters’ (spoils and other useful things) in 2018, averaging almost 12 per game. Walker has done a pretty good Andrews impression, recording 37 one-percenters over his past four matches. He had only four one-percenters in his four matches before that, and has never averaged more than three per game in his previous six seasons, having played the majority of his career as a forward.

Up forward for the Lions Eric Hipwood has been the star, with 11 goals and 14 marks inside 50 over the past three weeks. But a fair bit of the help has still come from the midfield, including Dayne Zorko’s five goal assists against Carlton – a career and league-wide season high – from just seven inside 50 entries. While the midfield hasn’t been outstanding in terms of the amount of times they have got the ball into attack, they have been pretty good at setting up scoring opportunities when they do.

Is this a turning point for the Lions? At the least – as long as it’s not just that they play well against the Hawks – they might become average again. Given how the 2010s have generally gone for Brisbane that would be something in itself.

Sunday, July 8, 2018

AFL Power Rankings: Round 16 2018

Richmond may well be further ahead of the second best side than any team this decade. However, that is not because they are the best side of the 2010s – it is because so far they lack any strong rival. 


Last week, Jake Niall of The Age wrote that Richmond is currently further ahead of any other side in the AFL than any team has been this decade. That includes the Hawthorn side that won three premierships in a row from 2013 to 2015, and the powerful Geelong and Collingwood sides of 2010 and 2011. Is he right?
Note that Niall is not saying that Richmond is the best side of this decade. Niall’s claim is just that the gap between Richmond and the second best side is the largest of this decade. And while it’s not completely clear that he is right about that, his claim at least looks to be reasonably defensible.
Richmond is currently 12 ranking points, or two goals, ahead of the second best side (Geelong) on the Power Rankings. Based on the end-of-season Power Rankings, since I started them for 2010, only Geelong in 2011 had a higher point gap over its nearest rival. (See chart below – note that I have re-based historical rankings to match my current method.) It’s also debatable whether Geelong was generally that far ahead of second-placed Collingwood that year; the Cats trailed the Magpies by a fair distance right up until the end of the home-and-away season.

Some other ranking systems, such as the scoring shots system used by Matter of Stats, also have Richmond ‘dancing on their own’. As Niall himself notes, there is certainly time for another team to catch up, but at the moment the gap between the Tigers and everyone else looks pretty high.  
That is not to say though that Richmond is the strongest side of the decade – indeed far from it. As I wrote about earlier this season no team in 2018, including the Tigers, looks to be as dominant as the greatest sides of the 2010s – Collingwood and Geelong in 2010 and 2011, Sydney in 2012, 2014, and 2016, and Hawthorn from 2012 to 2015. They’re not far off some of those sides, but they are probably not catching up to that historically great Cats side from 2011.
However, each of those ‘great’ sides had another side that was really good as well. Collingwood and Geelong had each other in 2010 and 2011. Hawthorn had Sydney in 2012 and 2014, West Coast in 2015, and to a lesser extent Fremantle in 2013 as worthy rivals. In 2016 and 2017 both Sydney and Adelaide were really good, along with the 2016 GWS side and Richmond’s premiership team last season. (The other premiership side of that era – the 2016 Western Bulldogs – were really good when it mattered most.)
In 2018 though, so far the second best side is by far the weakest of this decade. Second-ranked Geelong are a good side, but 2010 is the only other year since the Rankings began that they would even finish in the top three. It’s not even clear the Cats are the second-best side. They are currently seventh on the ladder, and reasonable arguments could be made for every other side in the top eight.
To be rated as ‘very good’ a team obviously needs to often put together very good performances. So far in 2018 only Richmond has done that (see table below). Other teams have had really strong streaks that have seen them considered ‘premiership contenders’, such as West Coast in Rounds 7 to 9, and Melbourne in Rounds 8 to 10. But only Richmond has made wins like their eight goal win against Adelaide on the weekend seem routine, albeit so far only in Victoria.

Of course that does not guarantee that Richmond will win the premiership. As Niall notes perhaps the last side to be this far ahead was Geelong in 2008 – one of the strongest teams ever – and they famously lost the Grand Final to Hawthorn. 
It is rare for a side to have more than a 50 per cent chance of winning the flag up until Grand Final day. Richmond is generally considered to currently have about a 30 to 40 per cent of winning the premiership. In other words, there is still a far greater chance that some team other than Richmond are premiers. There may not be another really good team in 2018, but there are still a bunch of good ones.

Sunday, July 1, 2018

AFL Power Rankings: Round 15 2018

Essendon may not have been the most inconsistent side in 2018, but they have definitely tended to have their best performances against the better teams. 



It does feel like Essendon has been the most – or at least one of the most – inconsistent sides in the AFL this season, doesn’t it? They shocked West Coast in Perth and GWS in Sydney, and they have also beaten Geelong, Port Adelaide and most recently North Melbourne. However they have also provided bottom side Carlton with their only win of the season, and lost to the Western Bulldogs and Fremantle. But have the Bombers actually been the most inconsistent side of 2018?

Maybe not. If we take the standard deviation of Essendon’s net margins in 2018, adjusted for home ground advantage and opponent strength, then the Bombers have only the fourth largest variation in their performances (see chart below). Melbourne, Carlton, and Adelaide are the three teams that have a slightly larger variation in their adjusted net margins.
A particular feature about Essendon’s inconsistency though is that they’ve done well against the top teams, and badly against the lower teams. Melbourne’s inconsistency is a result of them bashing up on the lower sides, but struggling against the higher ones. Adelaide’s inconsistency may well be largely a result of injuries. Of the three most inconsistent sides Carlton’s performances are the most similar to Essendon’s, in that they did OK against top sides such as Richmond, Sydney, and West Coast, but lost at home to Gold Coast and Fremantle. But Carlton still lost the matches they did relatively well in, as they aren’t yet good enough to win the games where they play relatively well against the good sides.
Essendon can, and six of their seven wins have come against teams that are above average or at least average (see chart below). In contrast, three of their losses have come against relatively bad sides, although their loss against Carlton may have had a fair deal to do with their inaccuracy in kicking for goals. The stark difference in the Bombers’ performances against the good and bad sides is probably where our feeling of their inconsistency comes from.

Next week I think I’ll write about the Age’s Jake Niall’s claim that Richmond are farther ahead of the second-best team than any side has been in the past decade. My ranking points above indicate a big gap between the Tigers and everyone else. Like Niall though I’ll argue that has as much to do with Richmond lacking a strong rival as the Tigers being a very good side, and also like Niall that being a very good side doesn’t guarantee a premiership.