Simulating the Finals Before Week 3

/

Again this week I'm providing a downloadable Excel spreadsheet so that you can make your own projections of the remainder of the season. The cells to change are the ones coloured yellow, which you should alter to reflect your opinion about the relevant probability for the listed game and winner.

The spreadsheet as downloaded includes the probabilities from the most recent update of the model I've been using to project the Finals this year. These probabilities, as you can see, suggest that Sydney are about 51% chances for the Flag, Hawthorn 37% chances, Port Adelaide 7.5% chances, and the Roos 4% rank outsiders.

Based on those probabilities and current TAB prices the only Flag market wager with a positive expectation is one on Sydney at $2.25.

Simulating the Finals Before Week 1

/

Any statistical modeller looking to project the results for the 2014 season must address the fact that teams finishing higher on the competition ladder have won about three-quarters of the Finals played since 2000, a substantially higher proportion than home teams generally have won over that same period.

My own view is that this history represents a credible Bayesian Prior and, accordingly, the binary logit model that I've built to estimate the probability of victory of the team finishing higher on the competition ladder during the Finals series reflects this through the large value of its intercept term:

Pr(Victory by higher finishing team) = logit(6.81 + 0.0249 x MARS Rating of Higher Finishing Team - 0.0308 x MARS Rating of Lower Finishing Team)

One consequence of recognising the importance of ladder position is that the Swans' chances for the Flag are significantly bolstered, as evidenced by the following matrix, which details the team-versus-team probabilities that I've used for my Finals simulations, based on the binary logit described above.

These team-versus-team probabilities, used in 1,000,000 simulations, produce the following summary outcomes for the Finals series, especially for the last two weeks:

Based on these results, Sydney's Flag chances, currently priced at $2.75 on the TAB, represent  the only value. A wager on them for the Flag carries a 30% positive expectation.

Looking next at modelled GF Quinella results we find that the most likely GF has the Swans pitted against the Hawks, an outcome eventuating in almost 37% of simulations. Barring that outcome, a Sydney v Fremantle GF is next most likely with a simulated probability of about 18%.

After that, the third-most likely GF sees the Hawks face the Cats. It turned up in about 1 simulated GF in 6, which makes its $9 TAB price tag seem quite profitable. It is, in fact, the only Grand Final Quinella with a significantly positive expectation.

A Swans v Hawks GF priced at $2.75 also carries a positive expectation, though the size of the edge is so small that it's not one I'd be willing to wager on.

(Two GF matchups are, by the way, impossible as a consequence of the Finals System that the AFL follows. These are the matchups of this weeks' Elimination Finals: Port Adelaide v Richmond and Kangaroos v Essendon.)

Simulating the Finalists for 2013 : Post Round 22

/
What a strange and intriguing time it is to be following - and trying to predict - AFL football. As I write this, just three days from the start of the final home-and-away series of games, and just one week more distant from the Finals, there's still no complete resolution of the Essendon debacle and no definitive statement about whether they will or will not play in those Finals.
Read More