2016 Round 26 : A Weekend of Favourites

/

Last weekend, both the Head-to-Head and Line Funds passed entirely on proceedings, leaving the Over/Under Fund to provide the round's only wagering action. This weekend, the opposite is true, and Investors find themselves with two head-to-head bets, two line bets, and no over/under bets at all.

Both of the home teams are favourites this week, which means the head-to-head bets are priced under $2 and the line bets involve us giving the away team start.

Those starts though are not much smaller than MoSSBODS' forecast home team victory margins, so the two line bets are appropriately small in size, the two of them together accounting for less than 1% of the original Line Fund. It's the head-to-head bets, therefore, that will most shape the round's profits and losses.

The ready reckoner provides detail about what those profits and losses might be, and reveals that the Geelong v Sydney game carries a 2% swing between the best and worst outcome, while the GWS v Western Bulldogs game carries only a 1.7% swing.

Best case, the Overall Portfolio will rise by 1.4% this weekend; worst case, it'll fall by 2.3%. A loss by either favourite will make a profit for the weekend impossible.

(I have, by the way, removed the "Draw" result from the ready reckoner this week, as I should have last week, such result being impossible in a Final where payouts are determined after extra time should it be required.)

TIPS AND PREDICTIONS

MoSSBODS_Marg has locked in top spot on the final Head-to-Head Tipsters ladder this week by opting for both of the favourites, as it ensures it'll finish the round two tips clear with only the Grand Final to be played.

In fact, this week only CTL has opted to select other than the two favourites, it choosing the Minor Premiers (the Swans) over the Cats in the Friday game.

MoSSBODS_Marg has been slightly braver as a Margin Predictor, its forecast of a 13 point Cats win equal-largest of all the Predictors. In that game, RSMP_Simple has forecast the smallest victory margin of just 2 points.

In the other game, C_Marg has forecast the largest win for the Giants (32 points), while Bookie_LPSO and Bookie_Hcap have forecast the smallest victory margins (15 points).

Across the two games the absolute difference between Bookie_LPSO's and RSMP_Simple's forecasts is 8 points, which is about 3 points less than the lead currently enjoyed by Bookie_LPSO over RSMP_Simple at the top of the Leaderboard. A couple of favourable results for RSMP_Simple, therefore, could make for an extremely tight finish to the end of the season.

C_Marg, meantime, has the chance to climb three places into 5th if results go its way (ie a very small Cats win or a Cats loss, paired with a very large Giants win), lifting it above Bookie_Hcap, ENS_Greedy and Bookie_9, and putting itself within a couple of points of ENS_Linear as well.

Amongst the Head-to-Head Probability Predictors, C_Prob (53%) and MoSSBODS_Prob (65%) define the range of probability estimates for the Cats against the Swans, and Bookie_OE (66%) and C_Prob (81%) define it for the Giants against the Dogs.

MoSSBODS DETAILS

This weekend sees both away teams incurring MoSSBODS' 3 Scoring Shot Travel Penalty, though in neither case does it change the team that MoSSBODS would otherwise have predicted to win.

It does though, of course, increase the forecast margin of those victories - to 13 points in the case of Geelong and 18 in the case of the Giants.

I've been thinking a lot about the appropriateness of imposing the Travel Penalty in Finals. A case could be mounted that teams playing Finals interstate might alter their travel arrangements to reduce or eliminate the inconvenience suffered during the home and away season - but then, travel is travel.

So, what does the empirical evidence suggest?

In the table at right I've summarised MoSSBODS' Performance in terms of mean absolute error (MAE) and accuracy across 6 eras of V/AFL history and treating separately home and away season games and Finals. As well, for those eras where interstate Finals have been played, the table also separately records MoSSBODS' performance in games where the Travel Penalty was applied ("With TP") and those where it wasn't ("No TP").

That table shows MoSSBODS has been more precise with its margin forecasts and more accurate in predicting winners in Finals where the Travel Penalty has been imposed than in home and away games from the same era. Also, in the modern era, it's been about as precise and more accurate in Finals with the Travel Penalty than in those without, while in the previous era it was both more precise and more accurate in Finals where the Travel Penalty was imposed. 

On that basis, I feel a little more comfortable continuing to impose the Travel Penalty in Finals, but this is something I intend to review in the off-season.

Anyway, on to a comparison of MoSSBODS' projections for this week with those of the TAB Bookmaker.

We see that the level of concordance between MoSSBODS and the TAB Bookmaker is fairly high this week. 

Both expect the Giants v Dogs game to produce more points than the Cats v Swans game (though in MoSSBODS' case, only just), and both expect team scores to lie in a range from about the mid 70s to about 90 points. MoSSBODS expects both games to produce fewer points than does the TAB Bookmaker, but not sufficiently fewer to trigger over/under wagers.

Lastly, an update on the comparative performances of MoSSBODS and the TAB Bookmaker after last week's solid MoSSBODS showing that saw it out-forecast the Bookmaker on Away team scores, totals and margins:

Home Team Scores

  • MoSSBODS: 19.7 MAE; +3 Average Error (Actual - Predicted)
  • TAB: 17.9 MAE; +2 Average Error

Away Team Scores

  • MoSSBODS: 20.2 MAE; +2 Average Error
  • TAB: 20.3 MAE; -2 Average Error

Aggregate Scores

  • MoSSBODS: 26.6 MAE; +6 Average Error
  • TAB: 25.4 MAE; -2 Average Error

Game Margins

  • MoSSBODS: 29.9 MAE; +3 Average Error
  • TAB: 29.4 MAE; +6 Average Error