Brave Sir Robin
Cancelled
- Jun 27, 2008
- 4,277
- 4,985
- AFL Club
- West Coast
Hey Eagles Board,
It's getting to that time of year where we know pretty well where each team stands and are starting to think about what our target should be from here. Despite losing a couple of games we shouldn't have, and looking like complete pretenders against Geelong, we are still somewhere in the discussion of contenders. As we all know, our chances hinge enormously on whether we are able to make the top 4 or not, and preferably the top 2 at that. I'm going to focus on the top 4 as our chances of making the top 2 from where we are now are extremely slim, but top 4 is of course very much within reach.
I do like playing around with predictive models and fairly recently I've started working on one for the AFL too.
Below we can see the cumulative number of times we make the top 4 based on how many premiership points we end up with. There's a bit of work to be done on the presentation (there are some odd discretisation errors that do not reflect the numerical data caused by the software I used to plot - see Adelaide, Collingwood, Essendon, Hawks, Saints) but the height of the bars shows you how many times out of 2000 each team made the top 4. Overall, we 'make it' about 400 out of those 2000 so I have us about a 1/5 chance of making finals from here - unlikely but still realistically within our grasp.
Historically, in a 22 game season 60 points is about the cutoff depending on percentage and we see that this year should be no different. For us, as we all know percentage will probably not be in our favour and so we are looking at needing more than 15 wins. That leaves us looking at 12 wins from our remaining 15 games to be any sort of a chance, or 11 and a draw. A long shot for sure, but in 2018 we managed to win 10 in a row so if we could start one of those winning streaks this week that'd sure help out.
Below is a list of games ordered by importance. In the simulated outcomes where we made the top 4 you can see the percentage of times each game was won. Anything towards the left is must win, anything towards the right is somewhat droppable. Note Adelaide appear twice as we play them twice and the home game is considered more essential to win for our top 4 chances. as you'd expect. Note also the relatively high importance of winning the Bulldogs and Melbourne matches given they are at home.
The exact values in the graph will change week to week given results, but one immediate thing we can interpret is that around 85% of our current remaining paths to the top 4 run through beating Hawthorn this week. If we lose, we're left with only 15% of our current chances.
So the main takeaway is that to make the top 4 we can only drop 3 of these, and losing any one of those games on the left hand side will be a humongous blow to our chances. Don't do it Eagles and especially do not do it this week.
The Broader Picture
Below is a prediction of the probability of finishing in each position for each team. To my eye, it looks like all teams have approximately the right probability density. The shaded area of each plot denotes the top 8. Perhaps the most groan inducing aspect of this is that our most likely result is, again smack bang in the 5-8 range. As unlikely as we are to make the top 4 from here, we're also pretty unlikely to miss the top 8. Nevertheless, there is still a cause for optimism looking at the left tail of our graph.
Further Discussion
Obviously there is some uncertainty about where our games will be played for the rest of the year and that could have a large impact on how our season plays out. For the moment I'm just assuming that all our home games will be at home. Once we know more about the fixture I can update and see how that affects things. If anybody has any results they would like to see, suggestions, criticism or permutations they would like to run put it here and I'll get around to it.
It's getting to that time of year where we know pretty well where each team stands and are starting to think about what our target should be from here. Despite losing a couple of games we shouldn't have, and looking like complete pretenders against Geelong, we are still somewhere in the discussion of contenders. As we all know, our chances hinge enormously on whether we are able to make the top 4 or not, and preferably the top 2 at that. I'm going to focus on the top 4 as our chances of making the top 2 from where we are now are extremely slim, but top 4 is of course very much within reach.
I do like playing around with predictive models and fairly recently I've started working on one for the AFL too.
Below we can see the cumulative number of times we make the top 4 based on how many premiership points we end up with. There's a bit of work to be done on the presentation (there are some odd discretisation errors that do not reflect the numerical data caused by the software I used to plot - see Adelaide, Collingwood, Essendon, Hawks, Saints) but the height of the bars shows you how many times out of 2000 each team made the top 4. Overall, we 'make it' about 400 out of those 2000 so I have us about a 1/5 chance of making finals from here - unlikely but still realistically within our grasp.
Historically, in a 22 game season 60 points is about the cutoff depending on percentage and we see that this year should be no different. For us, as we all know percentage will probably not be in our favour and so we are looking at needing more than 15 wins. That leaves us looking at 12 wins from our remaining 15 games to be any sort of a chance, or 11 and a draw. A long shot for sure, but in 2018 we managed to win 10 in a row so if we could start one of those winning streaks this week that'd sure help out.
Below is a list of games ordered by importance. In the simulated outcomes where we made the top 4 you can see the percentage of times each game was won. Anything towards the left is must win, anything towards the right is somewhat droppable. Note Adelaide appear twice as we play them twice and the home game is considered more essential to win for our top 4 chances. as you'd expect. Note also the relatively high importance of winning the Bulldogs and Melbourne matches given they are at home.
The exact values in the graph will change week to week given results, but one immediate thing we can interpret is that around 85% of our current remaining paths to the top 4 run through beating Hawthorn this week. If we lose, we're left with only 15% of our current chances.
So the main takeaway is that to make the top 4 we can only drop 3 of these, and losing any one of those games on the left hand side will be a humongous blow to our chances. Don't do it Eagles and especially do not do it this week.
The Broader Picture
Below is a prediction of the probability of finishing in each position for each team. To my eye, it looks like all teams have approximately the right probability density. The shaded area of each plot denotes the top 8. Perhaps the most groan inducing aspect of this is that our most likely result is, again smack bang in the 5-8 range. As unlikely as we are to make the top 4 from here, we're also pretty unlikely to miss the top 8. Nevertheless, there is still a cause for optimism looking at the left tail of our graph.
Further Discussion
Obviously there is some uncertainty about where our games will be played for the rest of the year and that could have a large impact on how our season plays out. For the moment I'm just assuming that all our home games will be at home. Once we know more about the fixture I can update and see how that affects things. If anybody has any results they would like to see, suggestions, criticism or permutations they would like to run put it here and I'll get around to it.
