Matter of Stats

View Original

2018 - Round 27 : Well Here I Am Again ...

The 2018 AFL Grand Final pits the team that finished 2nd in the home-and-away season against the team that finished 3rd for the fourth time since 2000.

In the previous three encounters, the team from 2nd has not fared at all well, losing all three times - not that, to reiterate, that fact is of any predictive value.

More generally, across that 18 year period, teams from 2nd on the ladder have a 6 and 5 record in Grand Finals, while those from 3rd have a 5 and 3 record, which, while superior, isn’t statistically significantly better at any significance level you wouldn’t be embarrassed about choosing.

In short, based on history (which is irrelevant) it’s either a certainty for the Pies, or pretty much a coin toss.

Let’s see if we can get a more useful assessment of relative chances by polling the MoS forecasters.

TIPS AND PREDICTIONS

Amongst the Head-to-Head Tipsters it’s only Home Sweet Home (who’ve chosen, as MoS law dictates, the designated home team, which for this game is the Eagles), and Consult The Ladder (who’ve chosen the team that finished higher at the end of the home and away season, which is also the Eagles), who have gone contrarian this week. The remainder have all opted for the favourites in the shape of the Pies.

That locks in the final places for the top 5 Head-to-Head Tipsters, leaving ENS_Linear in first, Bookie Knows Best in 2nd, and MoSHBODS_Marg and the RSMP twins tied for 3rd.

For the Margin Predictors, it’s a unanimous vote of confidence in the Pies, though to varying degrees, with RSMP_Simple forecasting the narrowest Pies win, and MoSSBODS_Marg forecasting the largest. The gap between them is only 10 points.

Overall, the mean absolute deviation (MAD) across the 10 Predictors is just 3 points per Predictor, which is quite low by historical standards.

In comparison to the current leader, Bookie_Hcap, a best result for MoSSBODS_Marg would see it move from about 82 points behind to about 78 points behind, while a best set of results for MoSHBODS_Marg would see it move from about 216 to about 214 points behind.

More relevantly, MoSSBODS_Marg will retain 4th spot ahead of ENS_Linear provided Collingwood win by 4 points or more. So, at best, MosSSBODS_Marg will finish 4th, and, at worst 5th, which is an acceptable if unextraordinary finish. MoSHBODS_Marg’s finish in 9th is, to be honest, a little more concerning. In the off-season, I’ll be assessing whether that finish can reasonably be attributed to chance, or whether it requires a more specific intervention.

Turning lastly to the Head-to-Head Probability Predictors, we find MoSSBODS_Prob assigning the lowest victory probability estimate to West Coast and MoSHBODS_Prob assigning the highest victory probability.

The differences in probability estimates, however, are small enough that they’ll not disturb the existing ordering on the Head-to-Head Probability Leaderboard.

WAGERS

Investors have no wagers at all this week on the head-to-head or line markets, an outcome that’s not at all unfamiliar to MoS visitors from seasons past.

MoSHBODS, to make a head-to-head wager, would’ve needed to see a $2.38 or better price for the Eagles, or a $1.88 or better price for the Pies. Unfortunately, neither of those demands were met.

Similarly, for MoSSBODS to make a line wager would have required a +14.5 handicap for the Eagles or a -1.5 handicap for the Pies. Here too, we’re a fairly long way off, with the two bookmakers, at best, asking the Pies to give 6.5 points start, and the Eagles to enjoy only 7.5 points start.

(Please click on the image at right to access a larger version.)

To finish, here are MoSHBODS' and MoSSBODS' opinions about likely team scoring.

Commentary about these forecasts will come in the week's Overs/Unders blog.