Round 19 – which state is the most productive “footy factory”?

This week we thought we’d take a look at where AFL players are from. The chart below is a pretty good illustration of the most fertile state breeding grounds. We’ve also included Ireland:

players per million.PNG

From here we can see that Victoria contributes the most AFL players relative to its population while NSW is only three times as productive as the island of Ireland.

 

Here’s the basic data:

Location Population Current AFL players AFL players per million people Avg games by player
New South Wales 7,670,700 40 5.2 53.6
Victoria 5,996,400 429 71.5 72.2
Queensland 4,808,800 45 9.4 58.4
Western Australia 2,603,900 131 50.3 75.7
South Australia 1,702,800 115 67.5 79.8
Tasmania 517,400 30 58.0 52.7
Australian Capital Territory 393,000 8 20.4 30.3
Northern Territory 244,000 12 49.2 45.4
Ireland (all) 6,378,000 11 1.7 24.0

We can see then that Victoria has obviously produced the most currently listed AFL players, and also produced the most per head of population. However, South Australians and Western Australians have played more games on average than Victorians:

average games by loc

And as a result the per-capita contribution of football to the AFL is higher by South Australia than by Victoria:

games per capita

Based on this, we’d call South Australia the purest footy factory in the country, producing many players and players who do well in their careers.

The more successful average careers of SA and WA seems to suggest that marginal Victorians are more likely to be recruited than similarly marginal players from elsewhere. That might be a recruiter bias towards the TAC Cup or clubs leaning towards hometown recruitment.

By contrast, the ACT shows as an emerging football base, currently having twice the players per capita as Queensland and four times NSW, but with a much lower number of games per player. 7 of the 8 have played a senior game, but the ACT players are mostly quite young and new to the league. This is a new situation, as we’d be surprised if there have ever been 8 Canberrans in the league simultaneously before.

When we distribute players and their career games by club we see some interesting effects by club, with Gold Coast having the most Tasmanians and Carlton leading the way in international players:

countbyclub

And here’s each club list by the location of origin of their games experience:gamesbyclub

The two most successful Victorian clubs of recent years (Hawthorn and Geelong) have the lowest current share of Victorian-origin players. Only St Kilda is as non-reliant on Victorian experience, thanks to Riewoldt, Gilbert and Armitage from Queensland being among their oldest heads.

The cosmopolitanism of Geelong and Hawthorn probably isn’t a coincidence. It suggests these clubs have gained a competitive advantage by looking elsewhere more than other Victorian clubs have done. This will have been especially important in the time of expansion, as we can see that GWS in particular have a strong Victorian representation, by far the highest share among interstate teams, representing their capture of a lot of elite talent from that state since inception.

Hawthorn, notably, has nearly as high a share of NSW players as GWS, drawing a number from western NSW as well as two Langfords from Sydney.

Finally, we can also use this chart to illustrate the potential  dangers of the go-home factor for each club:

homestate.PNG

Most Victorian clubs have a higher predominance of hometown players most non-Victorian clubs. Aside from West Coast as an outlier, it’s a completely bifurcated distribution. Structurally then, it looks as if the AFL pool and distribution of clubs more or less guarantees clubs inside Victoria are more able to draft their local players and be less exposed to go-home risks.

A note on the source: This is a custom count of players we have developed, starting with the FanFooty website and making a number of corrections to the “State of Origin” item in each club list. Edits consisted of things like:

  • assigning the correct state to NSW-origin Murray Bushrangers (the TAC Cup team covers north of the border too)
  • Correcting the state location of certain clubs (eg Dylan Shiel of Edithvale-Aspinvale is not from NSW, and Eastlake is in ACT not NSW)
  • shifting some players away from Victoria who only spent their late teenage years there (eg Harbrow, Hawkins, McGlynn, Betts)
  • Calling Phil Davis a Canberran because he started out at Marist and only moved to Adelaide around age 14 and that makes him forever a Canberry.

Due to our local knowledge these changes mostly impact ACT and NSW players, but we assume the larger states with more regular recruitment channels are largely accurate. If there’s further inaccuracies it’s probably mostly in Tasmanians and Queenslanders who are wrongly attributed to Victoria due to spending their last pre-draft years there.

Attempting to quantify goal creation

We are of course limited at HPN to publicly available data, but we try to push what we can do with that into new areas. This week we thought about using the goal assist statistic to identify efficient creators of goals. We assume that a lot of goal assists come in the form of passes to players inside 50, so we’ll use that as a comparison:

goal creators

This measure is basically a quick experiment and clearly isn’t a perfect indicator. We seem to have identified that forwards are the most efficient producers of goal assists, as they get them while getting relatively few inside-50s.

Franklin stands out as being the only key forward in the list of most inside-50s per goal – it shows that he is playing quite high and bombing the football into a forward line like the others in the top list tend to do. Compare the four Adelaide forwards who all have high numbers of goal assists per each inside.

Effectively what we seem to have identified is that most players who get a lot of inside-50s aren’t necessarily getting a lot of really high quality ones, rather they’re creating opportunities from sheer quantity. Patrick Dangerfield is the exemplar of this approach. That’s unsurprising, even very efficient teams tend to score goals off well under a quarter of their inside-50s.

What to look for this week

  1. Geelong and the Bulldogs should be won in the midfield

gee v wb.PNG

Both Geelong and the Bulldogs have, by our ratings, very strong midfields that generate good inside-50 differentials. Both get plenty of the ball (1st and 4th in disposal differentials, 4th and 1st for contested possession differentials), Geelong is much better in the ruck (6th vs 16th for hitout differentials) but the Bulldogs are the better clearance team (1st vs 4th in clearance differentials) which suggests losing in the ruck never bothers them.

Given Geelong seem to have the edge offensively and defensively, the Dogs are going to have to sustain a positive inside-50 differential to overcome them.

2. A two-Richards forward line?

The loss of Callum Sinclairto the Swans, in addition to the loss of Kurt Tippett, has had commentators talk about the damage done to Sydney’s ruck prospects. However, only twice this season has Sinclair led the hitouts in a Sydney game. This was against against GWS in round 12 when Tippett was injured early, and in round 15 against the Bulldogs when it’s entirely likley Nankervis still attended and lost the majority of the ruck contests.

The rest of the time, Sinclair has been a secondary ruckman and played other structural roles instead. Sinclair’s loss will be a detriment to Sydney’s already shaky forward structures. He is sixth at the club in contested marks and fifth in 1%ers, illustrating his role as a marking target who creates contests even when not taking marks, especialoly in moving the ball up the wings.

Judging by team selections, Sydney’s first port of call will be hoping Nankervis and Naismith can cover the forward structure role as well as ruck.

However with the continued absence of Sam Reid, if the Naismith-Nankervis pairing doesn’t prosper, a dearth of forward targets means over the next month we might see both recent forward convert Richards brothers, both Xavier and Ted, playing tandem roles that Sydney would prefer played by Callum Sinclair and Sam Reid.

Advertisements

Round 18 ratings – Hawthorn march onward

Hawthorn, who started the season slugglishly and who we’ve mostly rated adrift from or within the top pack but not above it, have pushed up another spot in our rankings this week thanks to their demolition of Richmond. Notably, they’ve pushed their defence rating (efficiency of opponent inside-50s) above the league average in the last two weeks after looking relatively weak in this measure in earlier performances.

r18 movement

West Coast slip in the ratings despite winning thanks to a really lopsided inside-50 count hurting their midfield rating. We use inside-50 ratio as out measure of between-the-arcs strength because it correlates strongly with victory – as The Arc vividly demonstrated this week.

The Bulldogs were also harshly punished (a 1.1% drop in strength rating is big for this stage of the season) after their loss to St Kilda. This was a perfect storm of:

  • breaking roughly even in inside-50s hurt their midfield rating
  • poorer inside-50 efficiency further hurt their forward strength rating
  • St Kilda’s relatively high inside-50 efficiency hurt their defensive rating
  • the addition of a weaker opponent hurt their opponent strength adjustment

It was a bad night for the Bulldogs’ aspirations all around.

Finally, here’s another look at the ratings for each team’s opponent set this year:

r18 opponent ratings

Most teams have played most sides, some have played teams twice. Adelaide have faced the hardest draw due to not yet having played Essendon or Brisbane, while West Coast have had the weakest schedule so far on the back of doubling up on Brisbane and not facing GWS. These are fairly modest differences.

Round 18 – It turns out the Lions really do have 99 Problems

Jackson wrote in the Guardian today that Brisbane have 99 problems but Tom Rockliff ain’t one and by our count it doesn’t look like that headline is particularly hyperbolic. We’ve gone through and attempted to enumerate what might constitute the Lions’ 99 problems:

  1. Defenders nearly all children
  2. A new draftee is their number one forward target
  3. Haven’t led at half time all year
  4. Very poor accuracy early in the season killed some chances to get a couple of wins
  5. Gold Coast took a third of their base:lionscrowds.PNG
  6. Gold Coast took probably the most fertile part of their local recruiting zone
  7. Cumulative $15m losses since 2008
  8. Promising 22-year old player forced to retire due to repeated head trauma.
  9. Currently have only four top 20 picks on their list
  10. That time a coach grabbed a promising young player by the throat at half time of a reserves game
  11. Gabba too big to extract any capacity premium from gate and membership
  12. Daniel Merrett’s balding spot a constant reminder of his/everyone’s mortality
  13. AFL never gives Fitzroy fans enough Melbourne games
  14. Sacked their captain Adcock last year who has gone on to contribute to a top 8 side
  15. Josh Walker somehow spending time in the reserves
  16. Might actually be tanking
  17. The fact you have to pronounce Josh Schache “Shackey”
  18. Still cursed from treatment of Daniel Bradshaw
  19. Still cursed from treatment of Michael Rischitelli
  20. Traded Lachie Henderson for Brendan Fevola
  21. Frequent storms in Brisbane hamper crowds
  22. AFL refuses to give any priority picks
  23. Unable due to financial situation to spend the full salary cap
  24. Unable due to financial situation to spend heavily on football department
  25. First-time coach Justin Leppitsch does not seem to have any senior assistant coach
  26. Lack of a Key Position Whatever who is above league average.
  27. Trent West still on list
  28. Dayne Beams’ near-constant injury issues
  29. Claye Beams not as good as Dayne.
  30. Ciaran Hanley not as good as Pearce.
  31. Lost their zone pick just before being able to draft Nick Riewoldt
  32. Expensive redundancy packages for CEO and other key staff ate into football department expenditure
  33. Training facility a poorly lit, flood-prone basement universally regarded as worst in the league
  34. Gabba facility also about to lose its adjacent car park
  35. Go-home factor
  36. Inability to make “next best available” drafting decisions due to go-home factor
  37. Heavy reliance on Lions@Springwood social club revenue
  38. Support from pokies revenue a common AFL moral failing but a failing nonetheless
  39. Failure to look outside of their backyard for coaching talent
  40. Lots of injuries every year raises questions about conditioning, training
  41. Non-traditional market means a large fan-base but almost entirely casual and easily disengaged
  42. AFL refused to fund the Lions to use the salary cap underspend they banked last year
  43. AFL keeps changing the rules regarding academy selections
  44. Queensland has a lot of bad opinions by bad people that might scare young players back to Victoria
  45. 3am lockout laws
  46. Unavoidably large travel load contributes to wear and tear on players making them leave or keep getting injured
  47. Play in a state where there are no votes in the government giving them any money or help with development
  48. “Miracle on Grass” high wore off
  49. Reigning Michael Tuck medallist Daniel Rich can’t shake a tag
  50. Never reconciled the two sides of their history and the traumatic circumstances of their birth into a cohesive identity that satisfies all parties
  51. Reserves bad for development due to frequently being uncompetitive
  52. Reserves bad for development due to being stacked with development talls
  53. Reserves bad for development due to top-ups often being juniors or substantially weaker players
  54. Little local media exposure to rebuild support
  55. Most people in Brisbane can only name players from 15 years ago
  56. Very difficult to watch Lions away games near the CBD of Brisbane
  57. Few to zero advocates, defenders or insiders within the mainstream footy press
  58. Football club either leaky or journalists feel comfortable making stuff up about them
  59. That time they floated the idea of having a real lion at games
  60. The stupid inflatable lion they run into then back out of when they take the field
  61. Did not lose the balance of close games between 2012 and 2015, never labelled “clutch” in spite of this fact.
  62. Success of the Broncos and Roar driving causal fans away
  63. Access to local talent still heavily dominated by Broncos
  64. Rockliff’s alleged ball-hog tendencies magnified in popular imagination due to current era’s proliferation of fantasy football games
  65. Majority of recent vaunted midfield recruits injured (Beams, Christensen, Bell)
  66. Ryan Bastinac apparently playing but useless
  67. Mitch Robinson unable to do it on his own
  68. Retired Bernie “Gabba” Vegas in favour of less engaging “Roy”
  69. Heavily in debt, over half owed to the AFL
  70. Don’t travel well
  71. Don’t play at home well
  72. Demoting the premiership-winning coach of their NEAFL side to assistant coach only a year after winning that flag
  73. Shane Woewodin’s reserves recaps suggest he may not be able to speak in intelligible sentences
  74. Widespread pundit view since they won three flags that must mean everything is their fault now
  75. Constant uncertainty about the location of future training base
  76. 24-hour infinite footy opinion vortex unkind to all struggling teams
  77. The Pineapple Hotel used to be a much better pub
  78. Nowhere suitable to play preseason fixtures in Brisbane
  79. Q-clash an almost perfectly terrible name for the Queensland derby
  80. Believe. Belong. would be a better slogan for Geelong than Brisbane.
  81. Still entirely unclear where or how the Lions will host their womens team
  82. Callum Sinclair’s trick knee
  83. Poor umpire understanding of stretcher protocols
  84. Youngest team in the AFL
  85. Only four players on the list have played over 100 games
  86. Only one player on the list has kicked more than 100 goals for the Lions (Josh Green)
  87. Only one player on the current list is older than 30 (Merrett) and one more older than 28 (Martin)
  88. Deputy chairman Leigh Matthews picking unwise fights with the Suns about their very existence
  89. Horrendous momentum-killing early fixture in 2016
  90. Lack of hope narratives become self-fulfilling
  91. Again: Their training rooms regularly flood.
  92. Reigning NAB Cup champions in perpetuity, but do not flaunt this fact
  93. Can’t get the ball or use the ball: Second last for contested disposal differential, last for uncontested disposal differential, last in effective disposal differential
  94. Second last for inside 50s in 2016
  95. Second last for contested marks in 2016
  96. Second last for marks inside 50 in 2016
  97. Second last for goal assist differential in 2016
  98. Historically terrible defensive season in 2016, conceding an average of 130 to top 8 sides
  99. Despite protests to the contrary, reports that Tom Rockliff might leave are major a problem

In fact, in our humble opinions, saying that the Lions only have 99 problems might if anything undersell the issue.

Things to watch

1. Toby Greene has evolved as a player, but is still kind of a butcher

“Cleaver” is the nickname that HPN bestowed upon Greene early in his career, named for Toby’s ability to butcher the ball with turnovers (not so much for his resemblance to Cleaver from Rake, nor for his weapon of choice which is a crutch rather than a comically large chopping tool).

This season, Greene has completely turned his game around as a small forward who occasionally drifts into the midfield, rather than the other way around. Greene has been at or near the top of the AFL clanger list for much of his career, but he’s doing it in a slightly different way now. As an undersized forward often used as a marking target, Cleaver engages physically on contests but often gives away the free kick because he’s just a little bit too physical.

Where as in the past this might cost GWS in the middle of the ground, it is now mitigated by his placement in the forward line and the strangling force of the Giants set-up.

The rise in Cleaver’s clangers this year (from 3.45 per game last year to 4.8 this year) is almost exactly accounted for by his rise in free kicks against (2.13, up from 1.27 last year). Greene is a very good player, and maybe verging on elite, however if he could keep working on this part of his game could be even better.

2. Geelong and Adelaide looks a ripper

In the HPN ratings, Geelong and Adelaide sit first and third at present, and very close in overall strength. Both sides are incredibly strong in two different areas; Geelong in the midfield and Adelaide up forward.

ade gee

In recent weeks both have also been on differing trajectories, with Geelong shedding a bit of ratings strength in recent weeks and Adelaide firming a little further as their midfield steps up. The match, like so many among top sides right now, is pivotal for each club’s fate – Geelong trying to secure a top-4 spot and Adelaide two home finals.

We’d suggest the result will hinge on how much more effectively Adelaide can use the fewer inside-50 opportunities they’re likely to gain. We’ve seen Geelong exposed by fast-break football in recent weeks, this looms as a very tough matchup for them if their midfield can’t assert some defensive dominance.

HPN’s Round 17 AFL Ratings – with a top 8 IN FLUX

ratings17

We open this week’s rating post with perhaps the most significant change to the HPN ratings in the last month, and one we have slowly been tracking. North Melbourne, after opening the season on a 9-match undefeated streak, have sunk down to 9th in the ratings and face a not-significant chance of missing the real-life finals as well. As we outlined last week, Port stood a real chance of leapfrogging North in the rating with a victory, as they have done.

Elsewhere, ladder-leaders Hawthorn have snuck above West Coast in the battle for 6th spot, and are well within sight of the Dogs in 5th. The top 4 remains quite congested with the gap between Geelong in first and GWS in fourth shrinking a tiny bit more.

St Kilda managed the biggest gain (1.3%) this week with their execution of a dismal Melbourne side, who slid an equal -1.3% in the ratings. Richmond fell behind the Saints (down to 13th), with the Saints coming into mathematical contention for the real-life top eight. Finally, Carlton moved above Gold Coast for the critical battle for 15th in our ratings.

With the season winding down, we thought it would be useful to have a look at how the three plot charts look, and how the 2016 teams compare with previous Grand Finalists.

2016 midoff

Adelaide’s offensive rating is historic, as we’ve pointed out before. No other team has been that strong in that area before, so it’s hard to look for historical precedents for their strength. GWS, Geelong, West Coast and Hawthorn firmly sit within the zone of former premiers, with North and Port on the fringes of the runner-up area. Sydney and Melbourne sit slightly below these teams, one good performance away from re-entering the fray. The Dogs are notably absent, with their poor offensive strength (and phenomenal midfield strength) leaving them on an island alone. Bringing up the rear is Essendon and Carlton, sides that have been approaching average defensively but relatively weak in the other two areas.

 

2016 middef

On this chart Sydney appear to be in a league of their own in relative defensive strength, with no other team matching them over our sample period. Geelong and the Dogs remain firmly in the GF zone, while GWS, Adelaide and West Coast are around where previous grand finalists have finished up. North and Port sit around the outlier grand final teams, whilst the Hawks sit just below the defensive strength you’d normally expect from a premier (expect that to rise slightly in the final few weeks). On their own at the bottom of the chart is Brisbane, who are in contention for the worst defensive team of our sample to date.

2016 offdef

Again, we see the absolute outlier that is the defensive strength of Sydney, but one that is lacking in the other area (offensive strength). A team with offensive strength this low has not made a grand final before. The cluster of “defense focused” grand finalists (mostly every past Swans team plus a few coached by Lyon, Malthouse and Worsfold) all had a more efficient forward line than Sydney do in 2016. And most of that cluster lost their grand finals.

Teams sitting firmly among premiers for offensive and defensive strength are GWS, Geelong, West Coast and Adelaide. Port and Hawthorn are not very far behind. The Dogs sit behind the outlier “defensive” grand finalists, and North look well off the pace. Additionally, we see a stray Lions team at the bottom of the chart, with Essendon and Carlton also trailing the pack.

 

Round 17 – close games are still basically random

Given the AFL’s insistence on scheduling Thursday night games, Round 17 has already begun with Hawthorn pipping Sydney in yet another close win (and a fourth last-moments defeat for the Swans).

A bit of a narrative has emerged from that game regarding Hawthorn as a “clutch” team and we want to just subtly dissuade our readers from buying too much into the idea that either: a) the best teams have a talent for doing better in close games or b) winning close games is associated with subsequent premiership success.

We stand on the shoulders of Matter of Stats here who has pretty conclusively demonstrated that successful and unsuccessful teams win close games at similar rates. As Tony puts it: the evidence for the proposition that successful teams win more than their share of close games is weak at best. He’s also had a look and found very little effect from narrow wins or losses on subsequent performance.

What we want to contribute is a couple of graphs and a discussion of some specific teams. Firstly, here’s a scatter plot of the records of every team in every season of football since 1997.

Note that close games are defined as margins 12 or less. We figure that anything that could be decided by two more, or less, scores qualifies as a close result and pretty close to random.

closegames.PNG

It turns out that, yeah, the hypothesis that close games are random is pretty close to the mark. The R-squared on this is garbage, as you’d expect, but more relevant is the number of high quality teams over the years who have struggled in games with narrow margins. Since 1997, ten premiership sides have won better than half their close games. Eight won less than half and the Crows in 1997 went at exactly 50%.

Hawthorn themselves last year went 20% (1 and 4) on their way to a premiership. This of course presents a problem for the “clutch” team narrative. In previous years they went 25% (1 win, 3 losses) in 2012, 60% (3 wins, 2 losses)) in 2013 and 100% (4 wins) in 2014. It’s almost as if these results are a bit random.

The thing with close games is that they’re largely a coinflip. And no matter how many times you flip a coin, the chances it comes up tails next time is 50%. Geelong in 2014, for example, had unprecedented luck, going 100% from close seven matches during the regular season, which represents an almost unprecedented level of “clutchness”. Yet when it came finals time, they crashed out in straight sets including by just 6 points in another coinflip match against North Melbourne.

Perhaps the most startling example of a truly great side struggling in close games, and yet still winning when it mattered, is the all-time great Brisbane Lions team of the early 2000s. Across their premiership years, they lost nearly all of their close regular season games. They won 2 games by 12 or less,  (both in 2003), had a draw, and lost 9 by 12 or less. That’s a win-rate of 21%.

The fact is that good teams do a lot worse in close games than they do in non-close ones, and vice versa for poor teams:
close nonclose

 

To emphasise – even the worst teams win a third of their close ones, and the top teams only win about 60% of them. This is despite a significant auto-correlative effect in these categories. That is, winning close games will move teams into a higher win/loss bracket on this chart than they’d otherwise be. So a lot of the teams with 80%-plus records are only distinguished from the next bracket by how their close results broke. In fact only Essendon 2000, Geelong 2007 and Collingwood 2011 are in the 80%-plus bracket solely on the back of bigger wins.

Hawthorn in 2016 stand as one of the strongest premiership favourites even if they often haven’t been as impressive as previous years. Their luck in getting wins from close games rather than losses has helped strengthen their ladder position for a charge at the flag.

But that doesn’t mean they are more likely to win the flag because of a special talent for close games. Such a talent cannot be demonstrated statistically and is undermined by Hawthorn’s jumpy close game results in previous seasons. Such a talent is unlikely to exist given the weight of evidence that close match results are a random walk regardless of underlying team quality. After all, the historical evidence is that the premier is about as likely to be a team with a 50% or worse record in close games.

Tall Defenders Can’t Tackle

With everyone talking about the potentially upcoming sequel to Space Jam (imaginatively titled Space Jam 2), HPN thought it could contribute to the hot-once-again genre of the sports film. You see, we were pouring obsessively over the numbers this week, as HPN does, and we found a disturbing trend:

least tackles.PNG

Tall Defenders Can’t Tackle.

This is a list of the lowest tackles per game, minimum seven games played so far in 2016. As you can see, tall defenders dominate this list. This is primarily because their duties don’t often require them to lay tackles, and instead to operate mostly in the air. Similarly, look at the list of mostly grounded players that dominate the in tackles-per-disposal list:

most tackles.PNG

Following Occam’s Razor thismakes sense – talls work in the air, smalls work on the ground. Except for Shane Mumford, who is a human battering ram.

Call Disney, get Dreamworks on the phone. We finally have a sequel to the much ballyhooed Woody Harrelson and Wesley Snipes film “White Men Can’t Jump”. It’s about a tall defender who lays a bunch of tackles in games of “street football” and leads his team to success in an underground football tournament. It’ll open huge. Call our agents.

Or maybe not.

(Note: one of us wants it known that he utterly disowns this joke)

Things to watch this week

1. North vs Port matters a lot

This could actually be the most important match of the remainder of the season. It’s a massive call given the intra-top 8 clashes coming up, but it could still be true. If North Melbourne beat Port Adelaide this weekend, then the top eight are all but locked in. Sure, there will be jostling for positions elsewhere, but the gap between the September haves and have nots for 2016 will be all but carved in stone.

But if Port Adelaide wins? Things get a little bit interesting. The gap becomes 2 games and not much percentage, with North yet to play the Bulldogs, Hawthorn, Sydney and GWS among top-7 sides, while Port face GWS, Sydney and Adelaide.

Per our ratings, they look evenly matched:

north port

The teams match up oddly next to each other. Port Adelaide have lost much of their tall firepower, struggling mightily in the ruck and with marking targets. By contrast, the strength of North comes in their talented ruckman (Goldstein) and stronger forward line. Port has a relatively good defence, especially in recent weeks, while the Roos struggle a bit touch more.

In spite of the ruck contrast, both sides have nearly identical midfield strengths according to our metrics (inside 50 ratio and clearance ratio). Both sides prefer to use by foot instead of hand, and both sides limit opposition uncontested marks quite well.

2. Sydney are probably about to drop to about 7th

Such is the cutthroat nature of the extremely tight top part of the ladder, that having just dropped another close game with top spot on the line, Sydney can instead expect to sit 7th by Sunday night.

The reason for this is the peleton behind them all have pretty soft draws. Assuming results go as expected, the Crows (vs Collingwood) and Bulldogs (vs Gold Coast in Cairns) will jump Sydney on wins.

West Coast (vs Carlton), Geelong (vs Fremantle) and GWS (vs Brisbane) all sit a game behind Sydney and within 3.5% of them on percentage. If those teams can score over a third more than their opponents, they should all push their percentage past the Swans’ 135.50% and therefore leave the Swans ahead of only North Melbourne within the top 8.

 

Round 16 Ratings: The 2016 Season Is Tighter Than Any In The Last Decade

Sorry for the clickbait headline, but through 16 rounds this statement is more fact than mere hot take. According to the HPN ratings system, applied historically back to 1998, this top four is the tightest since at least 2003.

Our ratings are expressed as in-season relativities, telling us how far above or below the league average each team is on some key measures of strength (related to inside-50s and scoring efficiency). We can represent multiple seasons of ratings like this:

ratings
And it’s not just the top four that is tight – the top seven have the smallest separation since 1998 (the first year that our ratings track). No team has cleared away at the top; and there are no major gaps for most of the top bunch. As indicated in last week’s column, there is a little bit of a gap between the top seven and North Melbourne in the ratings, but even this compares favourably with previous years.

The reasons for this are plentiful. Hawthorn, at the head of the pack in previous years, have been hit hard by injuries and retirements/delistings in the last twelve months. Last year’s ratings leaders West Coast appear to have had their playing style somewhat figured out this year, and look especially vulnerable away from the West. Sydney have improved on last season with a phenomenal defensive structure, and the long awaited rise of GWS is no longer long awaited. And that’s before we get to the historically dominant Adelaide forward-line and the Bulldogs’ midfield strength.

Neutral footy fans have never had it better.

movement.PNG

After several surprising results last week, the top four in the HPN ratings are now separated by less than 2.5% – which could be overhauled with another surprising result or two. Sydney have risen to second, and GWS have slipped to fourth. Geelong also fall back hard into the pack.

Hawthorn, despite remaining clearly on top of the ladder, slipped a place to seventh. That may end up embarassing us come September but we stand by it based on the season to date. They’ve generally been doing enough to win, having the third strongest midfield rating (ie they have a good inside-50 differential) but pretty middling in both the forward and defensive arcs. There is every historical reason to assume they will step up come finals time, but that’s a historical view, not one based on the statistical evidence presented to us this season.

It’s a testament to the closeness of the season that the previous tworeigning grand finalists are both in contention but not streaking away.

This week has been shaping as a critical one in determining the eighth place in these ratings. The recovering Port Adelaide meeting the free-falling North Melbourne. If Port can manage to continue North’s slow march this week, it gives them a not-insubstantial hope of chasing North down on the real chase for the final spot in the finals, given their respective runs home.

Round 16 preview – A look at Buddycentrism

This week we’re going to take a look at the question of how reliant teams have been on particular goal kickers. The specific spur for this is the talk of Sydney being overly Franklin-centric but the question is worth looking at more broadly – how much heavy lifting do each club’s top goalscorers do?

First let’s look at total goals scored:sumhogs

One thing that shows up clearly here is the sheer dominance of the Crows’ four forwards. Collectively they have outscored everyone else’s best 4 goalkickers and even the top 6 of every club except West Coast and Hawthorn. They’ve also outscored the entire Essendon Football Club. Hawthorn have a top 5 (Gunston, Breust, Rioli, Puopolo, Sicily) with similar output.

Below is the same information presented as proportions.

percentage hogs

Sydney do indeed show up as quite relatively reliant on the game’s best forward, with Franklin scoring a full 27% of their goals though him. Notably, 3 of the other 6 Swans players in double-figures for goals this year (Heeney, Tippett, McGlynn) won’t be playing against Geelong while Papley comes in for the first time in a while. While Franklin is entirely capable of winning games off his own boot, this lack of secondary output still strongly suggests a problem. Sydney will want to start finding more goals from alternative avenues over the next few weeks while waiting for Tippett and Reid to become available.

However, we do caution against the idea that only Sydney have a concentrated goalkicking output. Looking at the degree of concentration of goalkicking at each club among those who have scored goals this year, calculated using Gini coefficients (the same thing used to measure income inequality in countries), we can see successful clubs occupy a number of points along this spectrum:
goalsgini

(1.00 would be a perfectly concentrated goalkicking side, where one player kicked every goal, while 0.00 would be every player kicking the same number of goals.)

We can see the league’s top sides sitting at different points along this spectrum. Geelong, North and the Bulldogs have a relatively even spread of goalkickers in their ranks. however, Hawthorn and West Coast, despite their lesser reliance on their number one target, still show about the same degree of concentration in their goalkickers overall. They, like Sydney, seem to have relatively defined avenues to goal.

There may be a suggestion in this chart that there is such a thing as too much spread – bottom sides without any dominant goalkickers fill out the ranks of the league’s most egalitarian sides. Overall though, we don’t think we can make inferences solely from how concentrated a club’s goalkicking is. We’d suggest the issue for Sydney is more about overall efficiency and production, rather than the count of goalkickers – the concentration of goalkicking in Franklin is more symptom than cause here.

Getting serious about concussion

The AFL just released its Injury Report for 2015 (PDF). We can learn from it such things as:

  • clubs average about one new injury per senior player per year (37.7 to be exact)
  • groin and quad issues are recurring much less often in recent seasons
  • with an incidence rate of 0.7 per club, any more than one ACL in a season is unlucky for a club
  • hamstrings are the most common injury in the AFL.

But clearly  the most startling observation is to do with the recent trends in concussion:

concussion.PNG

Without any serious increase in concussion occurrence, games missed per season from concussion shot up to 4.2 per club in 2015 from 1.3 in 2013. The rate of games missed was only 0.3 games per club in 200 – that’s a 1400% increase in games missed compared to a decade a go. In other words, we’ve gone from 5 games missed league-wide to over 4 games per club.

Things to watch this weekend

1. Battle of the Clones

Working from statistics, if you were to describe North Melbourne and West Coast, it would be: dominant ruck divisions open up play for a kick-heavy midfield which targets a plethora of active tall and small forward targets. We’ve got a number of team-based statistical measures that we track at HPN, and they largely indicate that both sides are the same freaking team.

clones.PNG

Naitanui’s absence does create a point of difference, as West Coast’s back-up rucks (as The Arc pointed out here) are not in the same league as Goldstein . Which makes sense; considering NN is a gamebreaking former All Australian and the other ruckmen are not.

West Coast, however, have had an edge in the midfield. The Eagles have a 4% edge across the season to date in clearances won, and a 10% advantage in inside 50 ratios. Across a standard game between two teams that are so tight, the extra couple of clearances and handful of inside 50 opportunities could be the difference between a tight win or loss. The question is how dependent this was on Naitanui initiating things.

It’s quite possible that if the Eagles do win this game their place in the eight will be relatively safe, and the Roos could come under threat from the Power (despite their loss last night). Currently West Coast sit two games clear of Port, and North three clear, however North and Port play each other next week. Next week we will dive into the run home for all three sides, but this is something to definitely keep in mind when watching this one.

2. Tragic Number and Magic Number

If Brisbane lose to the Suns this week, the Lions will officially be the first team knocked out of mathematical finals contention this year. CONGRATULATIONS! If they somehow manage to win their game, and the Eagles lose to the Roos, then the Dons have a chance to get the “first eliminated” mantle instead.

With 8 games remaining and a 32-point gap between the Eagles and the bottom sides, it’s still possible with some heroic assumptions about margin (think ten goal wins and losses every week) for the Lions or Dons to catch the Eagles.

If the Eagles manage to win this week however, both Brisbane and Essendon will be knocked out. Their Tragic Number will hit 0. No doubt their focus will then finally move to next season.

At the other end of the table, the Magic Number for the Hawks is now just 2, with the Hawks sitting 5 games clear of Port Adelaide with 7 games remaining. We’ll keep a Tragic/Magic Number watch in this place for the rest of the season as teams start clinching and being eliminated from contention.