# 2011 - Simulating the Finals - Part II

A strong case could be made for the position that last week's loss by the Pies shouldn't have led to as large a rerating on MARS as it did, but however you look at it, it did bring them back to the field at least somewhat in Flag considerations.

With new MARS Ratings come fresh head-to-head probability assessments, and here they are:

One set of comparisons we can immediately make is the probabilities in this matrix for the four matchups this weekend, with the probabilities implicit in the current TAB head-to-head markets for these games.

• For the Cats v Hawks game we have prices of \$1.68 / \$2.15, which yields a probability range estimate for the Cats of (53-60%)
• For the Pies v Eagles game we have prices of \$1.18 / \$4.50, which yields a probability range estimate for the Pies of (78-85%)
• For the Saints v Swans game we have prices of \$1.33 / \$3.15, which yields a probability range estimate for the Saints of (68-75%)
• For the Blues v Dons game we have prices of \$1.33 / \$3.15, which yields a probability range estimate for the Blues of (68-75%)

(It's been a while since I've used this, more complex probability range calculation. It provides the logical range in which the bookmaker's assessment of a team's victory probability must lie given the two team prices, and is formed by calculating (1 / Own Price) and (1 - 1 / Opponent's Price). The derivation of this is somewhere in a previous newsletter, but I can't find exactly which one right now.)

The probabilities in the matrix above are similar to these ranges, but above the range for the Cats (67%), the Pies (89%) and the Blues (89%), and below the range for the Saints (64%).

Using these new MARS Ratings to simulate the Finals series produces the following outcome.

On the left are the results for the current set of 10,000 simulations, and on the right are those for the simulations we ran last week prior to the round.

The big winners - rightly so I'd suggest - are the Cats, whose Flag chances virtually doubled on the strength of last week's outing. At \$4.25 for the Flag with the TAB, these simulation results suggest that the Cats represent outstanding value. No other of the finalists offers a positive expectation in the Flag market. In fact, none of them come even close.

In the Grand Final quinella market, the chances of a Pies v Cats Granny rose to 40% with last week's results, so at \$3 on the TAB this proposition also represents excellent value. A Geelong v Hawthorn GF also became more likely, though \$6 is not quite enough to make it an attractive insurance wager.

One other interesting feature of the latest set of simulations is the change in some teams' conditional Grand Final probabilities. Collingwood, who last week were considered 82% chances to lift the Cup if they made the GF are now rated just under 70% chances, while the Cats have gone from about 2/1 hopes to even money propositions. Bear in mind that these conditional probabilities are weighted probabilities across all possible opponents - we're not saying that the Cats would be even money chances if they faced the Pies in the Big One.

Comment