There's not a lot of value in the TAB Sportsbet market for the AFL Finals right now, save for some generous pricing in relation to the Hawks in the Flag and Make-the-GF markets.
At least that's what you'd come to believe if you thought that MARS' team Ratings were accurate and that a binary logit model using result and Ratings data for every game since Round 1 of 1999 was the best way to extract meaning from those Ratings.
That binary logit is this:
- ln(Prob(Home team wins)/(1-Prob(Home team wins)) = 1.874 + 0.0340 x Home Team MARS Rating - 0.0355 x Away Team MARS Rating
This equation provides the following team-versus-team probabilities:
Reviewing the current head-to-head prices on TAB Sportsbet for the games in Week 1 of the Finals, the probabilities shown here for those same games appear reasonable in that none is more than about 7% points different from the probabilities implicit in those prices.
If, for each week of the Finals, we assume that Home Ground Advantage accrues to the team that finished higher on the ladder as at the end of the home-and-away season, we arrive at the following assessments of the 8 remaining teams' chances based on 100,000 simulations of the 2013 Finals series.
The Hawks are clear favourites according to these simulations, finishing 1st in 42.5% of them. They are though only equal-favourites with the Cats on the TAB markets, and their current price of $2.75 is assessed as representing a 17% edge, making them the only recommended wager in the Flag market.
In addition, the Hawks are assessed by the simulations as almost 65% chances to make the GF, which means that their current TAB Sportsbet price of $2.75 carries a positive expectation. No other team team represents value in the Flag or make-the-GF markets.
It's interesting to note how desperately unlikely it is, according to the MAFL simulations, that either the Blues or Port will capture the Flag. The Blues are about 100/1 shots of even making the main dance according to the simulations, and Port are about 250/1 propositions.
Lastly, let's take a look at GF pairings, where we find that value in the TAB Sportsbet market resides only in the price for a Hawks v Pies GF.
That's also true of the eight next-most likely pairings.
It's only when we arrive at the Hawks v Pies pairing, which is assessed as about a 1.8% probability but which is priced at $67, that we stumble across a pairing with any apparent value in the TAB market. It carries about an 18% edge.
None of the remaining 16 possible GF pairings represent value either, the least likely of which is a Port Adelaide v Carlton matchup, which eventuated in only 4 of the 100,000 simulations I undertook. That results therefore represents 14.6 bits of surprisal, meaning that it's slightly less likely than tossing 14 coins and having them all land Tails.
Richmond v Port Adelaide, and Fremantle v Port Adelaide Grand Finals are assessed as being only slightly more likely, representing 11.3 and 12.4 bits of surprisal respectively.
To me, contemplating the difference between an outcome representing 11.3 and one representing 14.6 bits of surprisal is like imagining how -11.3 degrees Celsius feels relative to -14.6 degrees Celsius: it's the quintessential distinction without a difference. (To my readers in Canada, whose temperature sensitivity is, doubtless, far better than mine, I apologise for what might seem to you to be an inappropriate parallel.)
The last aspect of these results that I'll note is that Fremantle, despite finishing 3rd on the ladder, does not appear in any of the three most-likely GF pairings. That's mostly because they finished the home-and-away season Ranked only 4th on MARS Ratings, behind the Hawks, Cats and Swans. The Dockers' inferiority relative to the Swans is supported by the TAB's current prices, which has the Swans as slightly more likely to win the Flag and to play in the Grand Final.