I’ve just found out about this: American Footy Star! It’s a genuine talent search wrapped up in a shiny reality TV show, and I love it. Putting NBA basketballers in the AFL was once purely hypothetical but now may be a reality.
It will increase the chance of recruiting dextrous, strong, tall ball players who can’t quite make the NFL / NBA. It will also take another tiny step to building a global appreciation for AFL.
The AFL is exciting, and it’s a shame to only share it within the few states of our Commonwealth. I want to see awareness of footy spread! One of my favourite AFL results of recent times was when Nauru, the world’s smallest republic, finished third in the 2008 world AFL championships. Footy players can come from anywhere, and they should.
The sport is free-flowing and has so few rules and constraints. Footy is a game of improvisation. It’s not like cricket, which requires a few highly refined movements. AFL takes all body types, from Tony Liberatore (163 cm) to Aaron Sandilands (211 cm). It uses 120 kg big men like Sandilands, and 70 kg wisps like, well, just about any rookie.
Australia is a small country, and innovation in recruiting is necessary. The success of Irish, a Canadian, Papua New Guineans, a Sudanese and a Fijian in the AFL shows only the glimmeringest glimmer of the eventual potential of casting a wider net. Meanwhile, taking yet more ‘skinny but promising’ lads out of Melbourne’s private schools is looking less and less like a sure-fire winner.
But recruiting is not the only aspect ripe for change. We need innovation in coaching as well.
Australian footy analysis generally consists of looking at things that are evident to the eye: ‘zone vs man-on-man’; stacking defence vs privileging attack; kicks vs handballs; corridor vs wings; Ablett vs Judd; numbers of tackles. It’s like having a lot of data on prices, but no model of supply and demand.
We know that Geelong gets 487 possessions to Brisbane Lions 338. But is this a cause or an effect of their winning? It’s like knowing that we have 487 businesses in our economy, but not knowing which ones are contributing the most in employment, wages and productivity. Let alone cracking open the black box to find out why some of them are so productive.
Check this out. It’s an article that tries to explain what in America is called the Ewing Theory, but in Australia could be called the Richardson theory. (Much as the Knicks won more often without their star forward, Richmond’s win-loss record indicates they were more likely to lose when their best player was on the field.) The article is full of hard science. And it’s not unusual.
MIT physicists are applying their models to sport, because there’s big bucks in helping teams win. The nerds recognise that optimising paths through a complex multi-node system is akin to logistics planning. That separating random fluctuations from underlying performance is not so different to doing stock market analysis. That determining what strategies to run, and in what proportions, is not unlike the game theory of international relations.
Also, check this out. It’s a New York Times article about an underdog NBA player who is brilliant because he is some sort of statistical savant and it’s my favourite article ever written. It hints at some of the most highly guarded analysis going on in US sports.
There’s patterns on a footy field. Randomness makes the patterns hard to spot, but we know they’re there because some individuals and teams are able to excel. Players that are lauded for ‘good decision making’ are ones who have spotted some of these patterns. A good coach can spot some patterns. But until patterns are objectively measured and tracked through time, human bias is going to corrupt their interpretation and application.
What’s the right level of analysis to find these patterns? Should we look at:
- Players individual actions?
- Spatial arrangements of groups of players?
- What part of the field the ball is in?
- How the ball is moved (forward, back, sideways, short, long, by foot, by hand, quickly, slowly)?
- The mix of types of players?
There’s enough footy played that we can get a massive sample. If we can apply the right frameworks, discernible effects will be found. We need proper analysis of what’s happening. Not just descriptive statistics, but models that can predict the success of various approaches, before they’re attempted…
It’s unlikely these will crack the game wide open, but they could be a more efficient path to improvement for a team that can’t afford to recruit.
Really all I’m adovcating here is the enlightenment. The application of hard science to the shaman-esque prescriptions of the Malthouses and Sheedys.
Aussie Rules football is an amazing game, and the free-flowing nature of it means that it is ripe for creative thinking. No doubt bringing in skilled foreign players will let creativity flourish at a micro-tactical level. But what will really revolutionise footy is bringing in some foreign techniques that will add creativity at a strategic level.
Share your thoughts on all this. Are you excited about the prospects of Yankees playing our ball game? Do you think letting nerds loose on statistics could revolutionise football strategy? Please comment below!
Thomas the Think Engine 
I like the idea of pushing innovation with further analysis of the patterns, but some of the best players to watch certainly don’t exercise much of their gray matter whilst making supposedly intelligent play. And also I think the most interesting things would not be the analysis of the skills but maybe the lack of skills of team mates or the opposition (maybe like economics’ game theory – the influence of others around you affects your performance), i.e. The Demon Cycle of Boom & Bust, or just Richmond 2010.
Also I’d like to add 2 points to paragraph #5:
1. North Melbourne has been captained by a Korean, Peter Bell.
2. Most of the best young Victorian players don’t come from Melbourne’s private schools but from the country (I haven’t done the stats, but that is the impression I get).
Thomas, if you could combine your two loves of economics and Aussie Rules, you should get someone to pay you for that analysis (and live in Melbourne)…
LikeLike
I’d like to add another link, to a blog that covers this topic handsomely: http://thesportseconomist.com/wordpress/2010/05/12/lebron-stumbling-on-losses/
But if you’re only going to click one link it should be this one, which i mentioned above: http://www.nytimes.com/2009/02/15/magazine/15Battier-t.html?pagewanted=1&_r=2
LikeLike
I like this previous post that starts to get at some of the “Battier Stats” of AFL:
https://thomasthethinkengine.wordpress.com/2009/06/30/out-for-the-count/
LikeLike
Hey – I like this post. I often think about this when I’m at the footy and one of the gentlemen sitting behind me starts to hurl abuse in the direction of the field. I’ve made some attempts to get to the bottom of this, but there’s only limited capacity to do so given the stats available are a bit patchy (and, as I think you’ve observed before, don’t always measure the right thing).
Examples :
– People get edgy when anyone kicks to Bret Thornton. This is justifiable in that Thornton has a shit record for contested posessions and is below average for rebounds from 50. However, Thornton has the Blues’ best stats for contested possessions and is one of the top players for spoils. I assume he gets a game every week becasuse he’s less likely to cause a turnover when they play that stupid game of trying to “slow the game down” by playing kick-to-kick in front of the opposition goals.
– Chris Judd has shit record for uncontested possessions, but no one seems to notice. Possibly becasue he has the highest rate of contested posessions in the league.
– When Fev was a Blue, he always presented the problem was that his team mates would avoid easy options because everyone had devotional levels of faith in his ability; even in cases where he had a pack of defenders using him as a climbing frame. In economics sense, this is a bit like a “three way duel”.
LikeLike
I’m very late to this party, five years late in fact.
But Richmond and Richardson is not the only example of the Ewing theory.
Richmond demonstrates the reverse Ewing theory twice: once with Deledio, and once again with Jake King.
Richmond has won only two games of AFL football without Brett Deledio, both times against the Brisbane Lions in Brisbane. Granted, Deledio has only missed about ten games in his career, and Richmond aren’t known for their sterling winning record, but it’s remarkable regardless.
And for some strange reason, Jake King playing in the Richmond forward line would make Richmond much more likely to win. It was weird and strangely compelling to watch the Richmond forward line to try ascertaining exactly what King would do to make it function better.
And just to add a personal pet peeve of mine when it comes to AFL analysis: sometimes the reason a team lost is simply due to poor kicking for goal. There’s often no further analysis required. Sometimes the team is playing sterling football, except its forwards can’t kick straight.
LikeLike
INteresting! I’d like to see an analysis of Geelong with and without Paul Chapman. His career coincided with the Cats best years, his absences were often marked by losses and now he’s gone the Cats are out of the eight.
Agree on the importance of executing your shots. Every other kick on the ground ends in ambiguity. The shot on goal is the only one with a binary impact on the only metric that matters. Commentators are too lenient on players who miss and they wouldn’t do the same to midfielders who fail to lay tackles, etc!
LikeLike