Matter of Stats

View Original

2011 - Simulations After Round 22

Last weekend's results had relatively little bearing on the simulated final ladder positions for the teams at the top and bottom of the table, but they had significant effects on the relative chances for a handful of teams still contesting spots in the top 4 and the top 8.

(Once again, in the diagram the latest simulation results are on the left and those from last week are on the right.)

Based on the simulations (and, in the Hawks' case, on maths alone) Hawthorn are now certain of a top 4 finish and the Eagles are nearly so. Combined, the Hawks' and Eagles' probability of making the top 4 jumped by 24% this week - up 4% for the Hawks to 100%, and up 20% for the Eagles to 97% - the same amount by which the Blues' chances declined. They're now only a 3% chance of making the top 4 and no value at all at $17 on the TAB. The Eagles are also too short in this market at $1.01 and are the only other bet on offer at the TAB in the top 4 market.

St Kilda, in losing to Sydney, saw its chances of making the finals fall by 10% to 93% while Sydney's chances leapt 16% to 96%. The only other team to enhance its finals prospects was the Roos, whose chances rose by 5% to 24%. Fremantle was the only other team to register a significant change in its probability of making the finals - down 8% to just 1%.

Amongst the teams still vying for a finals berth, only Sydney represents any value at all on the TAB, and even they're priced too close to breakeven at $1.05 to make any meaningful wager a serious consideration.

The only significant value on any of the AFL Futures markets at the moment if you believe in the simulations lies with the Gold Coast for the Spoon at $3.75. Our simulations have them as 40% chances in this market, with Port Adelaide 60% chances.

To understand why the simulation results suggest value in a Gold Coast flutter, consider: 

  • The respective runs home: Gold Coast face Melbourne away and the Hawks at home; Port Adelaide face Essendon away and Melbourne at home.
  • The current ladder situation: Gold Coast are a game up on Port Adelaide but have a significantly worse percentage
  • The margin predictions of our Projection Pursuit Regression model, which are as follows

Gold Coast are predicted to lose their final two games by about 41 and 52 points. If we model the outcome of these games as Normally distributed random variables with these means and with standard deviations of 37.7 points they we'd estimate the Suns' chances in these two game as being about 14% and 8%.

Port Adelaide are predicted to lose their final two games also, but by 46 and 5 points only. Making the same assumptions as we did for the Suns' games, these predicted margins equate to 11% and 45% chances.

So, using these probabilities and ignoring draws, for Port Adelaide to move past the Gold Coast, bearing in mind Port's superior percentage, Port needs to win both games and have the Suns win 1 or none, or Port need to win just one game and have the Suns lose both. Assuming that the results of all four games are independent gives us the following:

  • Port wins both and Gold Coast lose both = 11% x 45% x 86% x 92%, which is about 4%.
  • Port wins both and Gold Coast lose one = 11% x 45% x (14% x 92% + 86% x 8%), which is about 1%
  • Port wins one and Gold Coast lose both = 86% x 92% x (11% x 55% + 89% x 45%), which is about 36%.

In total then, Port are about 40% chances of relegating the Suns into the Spoon position. The credibility of this result rests almost entirely on your assessment of Port's chances at home against the Dees in Round 22.

Frankly, Port Adelaide have been diabolical in the last few rounds so you'd be a brave punter to bet with them for anything other than the toss - and even then you'd want odds. Still, at $3.75 that's almost a 50% margin if they're truly a 40% chance of doing what's necessary to climb off the bottom rung.