2016 Round 21 : Overs/Unders Update

/

MoSSBODS is this week predicting lower totals than the TAB Bookmaker in eight of the nine contests, sufficiently lower in three games to warrant wagers by the Overs/Unders Fund.

Those three unders wagers take the total proportion of the original Overall Portfolio Fund at risk to 9.7%, the highest proportion for a single round since Round 17 and the fifth-highest proportion for the entire season.

They also include a bet on the Tigers v Cats game, which means that only two contests have been left wager-free by all three Funds.

MoSSBODS COMPARISON TO TAB

In the three games where the Overs/Unders Fund has ventured a bet, MoSSBODS' opinion about the total differs from the TAB Bookmaker's (as it must) by over 2 goals. In four other games it differs by between about 8 and 10 points.

Only in the Hawks v Roos clash is MoSSBODS predicting a larger total, and there the difference is practically zero. Across all nine games, MoSSBODS' predicted average total is 170 points; the TAB Bookmaker's is 179 points.

On a team level, the story is also one of generally lower scoring expectations by MoSSBODS than by the TAB. MoSSBODS, for example, has Gold Coast scoring 85 points to the TAB's 105.5, and Sydney 88 to the TAB's 100. Looking across the entire round, six of the nine home teams and eight of the nine away teams are expected to score fewer points by MoSSBODS than by the TAB, As well, the TAB has six teams scoring 100 points or more this week, MoSSBODS only two. Essendon is the notable exception to this trend, MoSSBODS expecting them to score 94 points and the TAB only 83.

Notwithstanding these differences, MoSSBODS and the TAB agree about which games are most likely to be the two top-scoring (in order, Lions v Blues, and Dons v Suns), and which are likely to be the three low-scoring (Dogs v Pies, Tigers v Cats, and Saints v Swans).

COMPARATIVE PERFORMANCE

The switch of methodology for MoSSBODS proved beneficial again last week, MoSSBODS producing smaller mean absolute errors (MAEs) than the TAB on home team scores, away team scores, and on aggregate scores. It was only on game margins - predictions for which are unaffected by the change in methodology - that the TAB prevailed.

Overall, the comparative performances of MoSSBODS and the TAB are now:

Home Team Scores

  • MoSSBODS: 19.7 MAE; +5 Average Error (Actual - Predicted)
  • TAB: 17.8 MAE; +1 Average Error

Away Team Scores

  • MoSSBODS: 20.2 MAE; +1 Average Error
  • TAB: 20.4 MAE; -5 Average Error

Aggregate Scores

  • MoSSBODS: 26.9 MAE; +6 Average Error
  • TAB: 25.3 MAE; -3 Average Error

Game Margins

  • MoSSBODS: 29.7 MAE; +5 Average Error
  • TAB: 29.0 MAE; +6 Average Error