2018 : Simulating the Finals Series After Round 22

/

Time to update our views on how, according to MoSHBODS, this year's finals might play out.

Firstly, let's look at the relationship between teams' ladder finishes and ultimate finals performance, applying the methodology described here to the 50,000 final ladders produced by our original simulations for this week.

Note that, here too we are no longer perturbing the team ratings since we're so close to the Finals. This will have the effect, relative to previous weeks, of favouring stronger teams because the absence of additional variation makes upsets less likely. The optimal perturbation of team ratings for use in simulations - both for home and away, and for finals simulations - is something I plan to revisit in the off-season. I think the solution will be to add variability in increasing amounts the more distant is the simulated game.

And, speaking of favouring stronger teams, we see that the latest simulations have Richmond as about 47% chances for the Flag. Had we used the same methodology as we did last week, the estimate would have been 34%. In all likelihood, the most accurate estimate lies somewhere between those two figures, and is likely closer to 47% than to 34%.

In the new simulations, West Coast look pretty good for a Preliminary Final berth, as do Collingwood. Hawthorn are about 40% chances of progressing that far, and Sydney about 33%.

WEEK OF ELIMINATION IN FINALS

In this next chart we look at teams' chances for various Finals finishes, ignoring their home and away ladder positions (ie we focus solely on the heights of the bars in the previous chart). The numbers shown inside a cell are the percentage of simulation replicates (multiplied by 100) in which the specified team went out in the specified week of the Finals (or achieved the specified result).

(Note that 0 entries reflect estimated probabilities greater than 0% but under 0.05%).

Here, firstly, is the view when we don't perturb the ratings.

Richmond, as noted, are assessed as about 47% chances for the Flag and about 67% chances of playing in the Grand Final. Collingwood are about 19% chances, West Coast 10% chances, Melbourne 8%, and Hawthorn 6%.

For comparative purposes, here's the view if we perturb the ratings as per the previous weeks' methodology. In terms of Flag chances, West Coast, Hawthorn and Sydney are the main beneficiaries, largely at Richmond's expense.

GRAND FINAL PAIRINGS 

In this final chart we look at all of the Grand Final pairings that occurred in at least one of the simulation replicates. The numbers shown inside a cell are the percentage of simulation replicates (multiplied by 100) in which the team named in the row defeated the team named in the column in the Grand Final.

(Note that 0 entries reflect estimated probabilities greater than 0% but under 0.05%).

We see that the Grand Final pairing of Richmond v Collingwood is most common, and appears in about 20% of replicates. Next is Richmond v West Coast, which appeared in roughly 15% of replicates, and then Richmond v Melbourne (9%).

The most-likely Grand Final not involving Richmond is a West Coast v Collingwood one, which occurred in about 8.5% of replicates.