How efficient is each team in attack and defence?
I am to answer that question throughout 2016 by introducing a very basic efficiency metric to determine each team’s attack and defence efficiencies.
Conventionally, to determine how good each team’s attack and defence is, one would look at the points they scored and the points they conceded. There’s nothing wrong with that, but it does have limitations in regards to the “pace” of the game. A team might be efficient in attack, but they might not produce a large amount of points because they are efficient in a slow or gradual manner.
This can be solved by adjusting the points produced in a game on a per possession chain basis. Unsure what a possession chain is? Basically it’s the entire period where a team has possession and can be measured to be statistically distinct. A possession chain begins with a kick-in, a stoppage clearance win, or an intercept possession from a turnover. A possession chain ends with a stoppage, a turnover, or a score on the scoreboard.
One can think of it as every time a team has a chain of possessions, they have an opportunity to score, and how many points they produce per possession is their true offensive efficiency.
Of course, this isn’t representative of the game as a whole – and teams with more chains in a game don’t necessarily play a faster style or footy or have more opportunity to score (for example many defensive games force a lot more stoppages, and that particular game with plenty of chains wouldn’t be called “fast” or “attacking”). But it’s a much better picture of each team’s attacking and defensive prowess.
One thing to factor in is the fact that some teams have more possession chains due to winning a higher proportion of clearances out of stoppages relative to their opponent. While this does contribute to a more efficient attack and defence, I’ve standardised its importance within every match. Why? Because games with high stoppages or low stoppages distort the numbers, and aren’t representative of a team’s efficiency in general. This is exemplified for a team that wins a large amount of clearances, in a game that has a lot of stoppages – it messes with the numbers. I’ve standardised this by taking the ratios for each team’s attack and defence, 95% of its efficiency is explained by its points produced per chain, and 5% is explained for their clearance win ratio that helps them generate more (or less) chains relative to their opposition.
The next adjustments to this efficiency metric will adjust for the aspect of randomness of scoring shots, as analysis has shown that scoring accuracy is over 95% random. Further adjustments might be made for strength of schedule.
If you want to learn more about this statistic, or have any feedback, feel free to contact me on Twitter.
Here are the ratings for the AFL after Round 1, 2016: