Sunday, April 3, 2016

AFL Power Rankings: Round 2 2016

FootyMaths alerted me this week to a new thread on Reddit which collates the various AFL Power Rankings there are around the net. It includes my rankings, along with those of FootyMaths, Matter of Stats, Fox Sports, The Roar, Sportsfan, Footy Prophet, Squiggle, ASAP, and /u/ Statisticc. (I won’t include all the hyperlinks here – you can click on the Reddit page and follow through to the individual ranking systems from there.)
 
The comments were relatively fair, in that there weren’t too many of the ‘these rankings are stupid, so-and-so obviously knows nothing about football’ variety. One thing that I did note though was that there seemed to be a general feeling among commenters that early-season results, such as Geelong beating Hawthorn, should have a higher weight.
 
 
Is this the right view to take? I’ve dealt with this topic before as early season results seem to make a lie of my and other rankings. There are reasons to think that clubs do change more in ability between seasons than within seasons. Players get half a year older, their abilities have more time to improve or deteriorate, and there are important changes in playing lists. You wouldn’t dream of, for example, having a National Basketball Association rankings system that did not account for events like LeBron James moving back to Cleveland.
 
It occurred to me that I could test this by comparing how much my rankings changed at the start of the season with the middle of the season. I didn’t compare this against the end of the season in case some clubs partially throw in the towel once their finals hopes are gone. If a new year does bring big changes, then my thinking is that my rankings should have larger changes at the start of the season as they scurry to adapt to the new season’s order. More specifically I looked at the changes in ranking points across clubs from Rounds 1-6 for each year (from 2011) compared with changes from Rounds 8-14 (because of byes, these seven rounds typically cover only six matches for each club).
 
Are there bigger changes at the start of the season? Not really (this actually surprised me a little). I won’t put all my results here – the more mathematically-minded among you can probably go back and basically verify this for yourselves if you want – but the variation in changes don’t appear to be any larger on average at season’s start. Maybe this changes if you look at a longer period than five years, but at least it’s not clear to me yet that I should be attaching any special significance to the first few rounds’ results.
 
Big changes happen within seasons too. For example, last season Fremantle fell away in the second half of the season. If a new season had started in say, July, then perhaps more may have been made of the Dockers’ drop, instead of this only being focussed on now the new season has started. Fremantle’s fall really started in mid-2015, but fewer people noticed because it was the middle of the season and they were still sitting on top of the 2015 ladder. Another example, in the opposite direction, is Sydney picking up steam in the middle of 2014.
 
There is another way of looking at this question though, which may be more pleasing to those who hold that you should attach a lower weight to last season’s results. That is, I did find some evidence that there is a stronger relationship between changes in rankings points at different points within the same season than between seasons. For example, West Coast and the Bulldogs started well in 2015, and their ranking points kept on improving throughout the season as it became clearer and clearer that they were big improvers last year.
 
Anyway, getting back to the rankings themselves, this week the big developments are Adelaide moving into the top four after thrashing Port, and Carlton returning to the bottom after getting thrashed by Sydney – Essendon actually winning a match also had something to do with that. The big improver two rounds in is the Gold Coast Suns, who have picked up ten ranking points so far, while Fremantle have lost around the same number of points. Not that you should be placing too much weight on these early season results. (Or should you … ?)
 

No comments: