With the equivalent of an AFL intermission upon us this week, I thought it time that I add something with some more content.
It is obvious that the Eagles are in a world of pain at the moment. The burning question is when will the pain end? The dilemma of the Eagles is not necessarily an unusual one, but perhaps just amplified little at the moment. The loss of their two key mid-fielders over the summer has left a gaping hole the size of a small pacific nation. Add to that a long list of injuries, and perhaps some less than well timed suspensions, and the Eagles are at a point where they would battle to win an under 14s high school game.
So we basically know what the problem is, but as an Eagles supporter there is only one question that bounces around in ones head; when will the next win come? Most will probably have heard the expression of a team being ‘due’ for a win, but what exactly does that mean? To me this is referring to an unwritten rule of thumb that a team can’t continue to lose forever. To quote a piece of popular culture: “On a long enough timeline, the survival rate for everyone drops to zero.”. This is oh so true for football teams that are winning, and although opposite often holds true for losing team, there is the odd exception like poor Fitzroy. I digress.
Instead of pondering the idea of when the Eagles might likely rise to the winners list, I have put together some numbers on how often team wins when on particular winning/losing streaks. The following graphs is based is a combination of all teams, and shows the percentage of times that a team wins when having a certain number of consecutive wins. -ve consecutive wins is consecutive losses.
The thin blue line is the actual data extracted from my records, the red line is the calculated best fit for the data, and the thick blue line is my hand predicted trend for this graph.
From this an obvious pattern emerges. As teams start to string together more games, their chances of continuing to wins increases, to a point. The opposite also seems to hold true. Once a teams starts to drop into the darkness, it is long way out. Not immediately apparent from the graph, but looking at the raw data there is a strange trend around the point after one consecutive win/loss, albeit very slight. For a team who has won only one consecutive game, they win only 47% of their next games, giving them a better chance of losing the following week. Similarly, those teams with one straight loss win 52% of their next games.
From this sort of graph, we can step back and assess where the Eagles sit in the scheme of things. The Eagles are sitting on 6 straight losses right now, which puts them in the thick of the bad side of this graph, but they do seem to be reaching a point where they should be ‘due’ for a win.
As someone who tries to sit back and look at these sort of graphs, many questions enter my mind. Why is it that at about 10 losses it that the fortunes of teams seem to start to improve? Could this be explain by players returning from injury? What is the average injury time for an AFL player? Could it be young teams that have a greater effect on this? Is 9 week the sort of time frame needed before they start to get their game together, or is there something else causing this effect? Of course keeping things like the Punter system in mind, the other burning question for me is that is this something that I give to my system and let it learn this? I have done some preliminary tests and this appears to be a learnable dataset for Punter, so this could perhaps be part of future improvements to the way Punter picks games ….
Cheers,
Hoopa
