The Australian public is bombarded with countless statistics during every AFL game: contested possessions, uncontested possessions, marks, handballs, one percenters, the list is seemingly endless.
However how well do we use these statistics to gain a better understanding of our game?
Metres gained is the trendy statistic being thrown about in today’s AFL. You will be hard pressed to find a television or radio broadcast that doesn’t mention it.
However, the metres gained statistic is flawed and is a prime example of how Australian sports lag behind international counterparts in the fields of analytics. To remind everyone let’s see how metres gained is calculated.
Metres gained is the distance between where a player takes possession and the point of the next player’s disposal.
While it is impressive to see a player average over 300 metres gained a match, the statistic is mostly empty in its meaning. A player who kicks the ball fifty-metres, to an opposing player is still awarded 50 metres to his metres-gained. Compare this to a player who creates space by hand-balling backwards to a teammate who then is able to make efficient progression up the ground. They are docked metres gained, as the first possession goes backwards.
The idea behind metres gained is similar to popular yardage statistics in American football. Americans are up there with the world leaders in sports analytics. However yardage statistics are only counted on completed possessions. If Tom Brady throws a 70-yard long bomb that isn’t caught it still isn’t counted as +70 yards gained.
While, not perfect, a more balanced way to assess metres gained would be to count the distance from a player’s possession to the point where one of two things happen.
When the ball is dead (out of bounds, ball up, free kick, score), and when the ball is turned over.
Furthermore, if a player turns the ball over, the opposing teams metre count is docked from the player who turned it over. For example, after a chain of possession, Gary Ablett turns the ball over and the opposition gains 50 metres before the ball is dead. Gary Ablett is docked 50 metres.
In layman’s terms, a player’s role in a chain of effective possessions will be rewarded and turnovers will be penalised. Sounds pretty straight forward to me.
While metres gained has potential to be improved, it is not the only scope for improvement with AFL statistics and analysis.
There are other devices we can use to fully understand a player’s impact on a match. For example, a simple plus/minus (+/-) statistic could be effectively used in AFL, a +/- is simply the teams score when a player is on the field, minus the opposing teams score in that time.
It is a very simple base line to assessing a player’s influence. While +/- is definitely not perfect, it could still have relevance in Australian football.
Along with the +/- statistic, the use of an on/off statistics needs to definitely be used more in the AFL. The on/off statistic is used to compare a particular statistic for when a player is on the field, compared to when a player is on the interchange bench.
This can be done with basics such as team scores but can also be tailored to position. For example; a midfielder could have an on/off statistic for clearances, while a forward could have an on/off for marks inside 50 metres.
The unique free-flowing game-play, and lack of market compared to international sports makes Australian Football one of the more difficult sports to statistically analyse, however steps need to be taken if we want something more in depth than ‘how many A-Graders’ a team’s midfield has.
Glenn
Guest
I think Karl has validated your point, Vince. Without Karl's background as to how the data is calculated and how it is used together with other statistics, commentary teams highlighting it is pointless. People will just look at it in the same way that the NFL statistic is used, but clearly this is wrong. I agree, get rid of it from the game broadcasts.
Wayne
Roar Guru
I've always liked comparing the Hit Outs / Hit outs to Advantage. Sometimes a team was getting 60+ hitouts, but only 15 to advantage. Sure the ruckman are winning the ruck, but the team isn't doing anything with the ball
Vince Massara
Roar Rookie
Thank you for your comments, Karl! These are some great points. The more conversation we have about this stuff the better it is for the game as a whole. You all do an amazing job at Champion Data.
TomC
Roar Guru
The first sentence pretty much nails it. Although it'd be great if a few commentators could be reminded of it before they state 'X is using the ball really well - he has a disposal efficiency of 90%!' To be fair to Vince the stat is used incorrectly more than it is used correctly, and I'd be entirely in favour of more articles on The Roar debating what stats are useful. On the retention/efficiency stat I'm interested to see them used interchangeably like that. I had always thought that distance was considered when determining efficiency, so a 50m kick forward to a contest might be counted as efficient while a sideways kick to a contest was not. Is that incorrect?
Slane
Guest
Thank you Karl. I'm a big fan. Love the work you guys do at Champion Data.
Ryan Buckland
Expert
That's really interesting, as all of your forays into the Roar's comments section are Karl! Thank you. Have you considered publishing more of this kind of information? It is really useful. For example, this is the first time I've seen someone from Champion Data publicly state how metres gained is calculated, despite its widespread use across the AFL media. Also not sure if you're one for taking suggestions, but I'd like to see whether it is possible to create a +/- statistic for inside 50 entries. Would be happy to talk you through the idea; ryanbuckland at outlook dot com.
Aiden
Guest
A more interesting question that your example raises is, perhaps asking the umpires to decide best on ground is a flawed idea? Or, do Gold Coast actually perform better without Ablett? Or perhaps, like you said, a single metric can be irrelevant in isolation.
Karl Jackson (Champion Data)
Guest
Any single metric can be irrelevant when considered in isolation. Metres Gained gives context to a player's disposal count. It doesn't take into account quality, but it still adds value. Disposal count + Metres Gained + Retention/Efficiency will give you a better idea of a player's impact. Better yet, take into account where the player won the ball, how they won it, the end result and how that influenced the position of the team. That's essentially how AFL Player Ratings work. NBA is the only other sport with a comparable system (EPV). Full disclosure: I created the system. Clubs look at Metres Gained at chain level to measure the effectiveness of ball movement. eg. Defensive 50 stoppage chains: metres gained minus reply metres (what the opposition gained after a turnover). Player numbers are also used by clubs, as the raw Metres Gained number, Effective Metres Gained (effective disposals only), Assisted Metres Gained (metres gained by teammates after you get the ball to them uncontested), and Net Metres Gained (Metres Gained + Assisted Metres Gained). Metres Gained is also measured as metres gained towards goal, not just distance of kick, so you can lose metres by kicking/handballing backwards or sideways. One issue with +/- in Australian football is that the impact of a single player on an 18-person team isn't as strong as on a 5/6 person team like in NBA/NHL. Another is that many players play close to 100% of game time, meaning it is truly irrelevant for them. Example: In Gary Ablett's latest Brownlow Medal year (2013) he has the fourth-worst +/- of more than 200 players to play 20 or more games. Gold Coast had an average margin of -8.3 points per 100 minutes with Ablett on the ground (-198 from 2373 minutes), and +39.8 points per 100 minutes with him on the bench (+58 from 146 minutes). Should we take the Brownlow off him? The most comparable measure to +/- in football is at centre bounces (Essentially 4v4). Clubs have access to centre clearance differentials with/without players in the square, as well as the ability to group the three mids, or all four players to look at effectiveness.