How good draft choices can make or break AFL teams
Wednesday, Dec 27, 2017, 08:46 PM | Source: The Conversation
By Jeff Borland
How good draft choices can make or break AFL teamsJeff Borland, University of Melbourne
In this series we're looking at how the economics of sports is doing away with hunches and intuition. Using data and research to evaluate players, strategies and even leagues.
In many sports, picking the right players in the draft can make or break teams. Richmond Football Club's latest premiership was built on good draft choices, according to my research with Binuk Kodituwakku.
Alex Rance was not only a key part of the 2017 premiership, but he was a tremendous 18th pick in the 2007 draft.
In the Australian Football League (AFL) each team generally chooses four or five new players through the annual draft. This is 10-15% of the entire team. Having the best players is no guarantee of winning - other factors like training and coaching also matter.
But as Melbourne Football club's recent poor performance shows, bad draft choices can weigh a team down.
Read more: How and why economics is taking over sports
As in many sporting leagues, the annual AFL draft allows teams to recruit new players who have been playing in junior competitions or in a senior competition at a lower level than the AFL.
The AFL draft is conducted in "reverse order" - the team that finished last of all 18 teams in the preceding season gets to make the first choice. This repeats until all the draft choices are used up - so the team that finished last has the 19th choice and the team that finished second-last the 20th, and so on.
This is done to improve competition within the league. Good draft choices can give a boost to struggling teams.
We created a measure to value draft choices called the Draft Performance Index. Using this measure we can compare the relative performance of teams in the draft, in the same year and across years.
It's a complex equation, but our measure essentially takes a player's performance in the five years after they were drafted and compares that with players selected later on in the same draft. It then adjusts for this comparison across previous seasons.
Our Draft Performance Index can be interpreted in a relatively straightforward way. If a draft choice equals one then the team got exactly the same value out of its pick as teams in previous seasons. You can think of this as breaking even.
If a team has a Draft Performance Index above one it means that the player chosen has done better than players picked at the same draft number in previous seasons. And if it is below one, the pick was worse than draft choices by teams in previous seasons.
Between 2005 and 2011 Richmond did extremely well, with five of their choices scoring well above one, on our measure.
Judged by value relative to the draft choice used to select these players, our measure suggests that David Astbury (taken at number 35 in the draft), was the best Richmond pick in this period. His Draft Performance Index was a staggering 3.5.
On the other end of the spectrum is Melbourne Football Club. This next chart shows its Draft Performance Index from 2005 to 2011. But this chart is organised by the draft number rather than the year.
What is immediately apparent is how low the values are for Melbourne's top picks over this period. In other words, when Melbourne was performing poorly enough to receive high draft choices, it didn't take advantage of it.
While Melbourne did better with its lower picks, those players generally aren't as important for team performance as players selected with higher picks.
The growth of professional sports has meant that player drafts are now major entertainment events. That growth has brought more revenue to teams, and with that more focus on how teams can improve - including through the draft. As our research shows, having a high pick hasn't always meant that teams select the best player.
One way teams are trying to improve their draft choices is through the use of data analytics.
Analysing the data can not only inform training and match-day strategy, but player choice too. We're still early in this data evolution, but the imperative to win combined with massive increases in computing power make it likely that we'll see huge advances in the next decade.