Matter of Stats

View Original

2019 - Round 27 (Week 4 of the Finals) - Tigers Burning Brighter

The bookmakers have the Tigers as about 3-goal favourites over the Giants, and there’s not a great deal of disagreement coming from any of the forecasters about that assessment.

TIPS AND PREDICTIONS

The Head-to-Head Tipsters are unanimous in their support for the Tigers, thereby ensuring that Bookie Knows Best will finish on top of the Leaderboard by one tip from RSMP_Simple and ENS_Linear (and MoSHPlay_Marg in the unlikely event that it chooses GWS and they win).

Bookie Knows Best’s final tally of 136 or 137 from 207 compares with its tally of 145.5 from last season, which was good enough only for 2nd place. ENS_Linear topped the table in 2018 on 147.5 tips. It’s fair to say then that this year it has been harder to tip winners than last year.

Amongst the Margin Predictors, all but C_Marg have forecast margins in a narrow range between 16 and 22 points. As a result, RSMP_Weighted is guaranteed 1st place on that Leaderboard, RSMP_Simple is guaranteed 2nd, and Bookie_LPSO is guaranteed 3rd. MoSHPlay_Marg could finish as high as 5th, but only if it forecasts a Tigers win of more than 32 points or of less than 10 points, and the Tigers oblige.

RSMP_Weighted’s final mean absolute error (MAE) of around 26.5 points per game compares with its 26.7 points per game result from 2018, which was good enough only for 6th. Bookie_Hcap took 1st place with 26.2 points per game in 2018.

The all-Predictor average looks like coming in at around 26.9 points per game this season, which is very similar to last year’s 26.8 points per game.

Turning, finally, to the Head-to-Head Probability Predictors, we find probability estimates spanning a range of 12% points from 59% (C_Prob) to 71% (Bookie_RE, Bookie_LPSO, and MoSSBODS_Prob). The differences are not large enough to alter the final rankings of the top 3 Predictors, which will wind up being MoSSBODS_Prob, MoSHBODS_Prob, and Bookie_OE.

MoSSBODS_Prob’s final average of around 0.1158 bits per game compares with its 0.1831 bits per game last season, which was good enough only for 4th place. Bookie_OE finished 1st in 2018 on 0.1889 bits per game. This year, the best bookie-based Predictor will finish with an average around 0.1059 bits per game, which suggests that probability forecasting has also been more difficult this year compared to last.

WAGERS

Neither MoSSBODS nor MoSHBODS could find value in any of the markets this week so, for the first time this season, Investors will finish the round in exactly the same position as they started.

In fact, even had MoSHBODS been allowed to wager on the away team Giants, it would have declined to do this as well, since its minimum price would have been around $3.45.

With no wagering this week, that means we know that:

  • The Head-to-Head Fund will finish up 29.1c from a 14.8% ROT on a 1.97 turn. It made 75 bets with an average size of about 2.6% of the Fund, and collected on 40 of them.

  • The Line Fund will finish up 34.3c from a 17.9% ROT on a 1.92 turn. It made 92 bets with an average size of about 2.1% of the Fund, and collected on 53 of them.

  • The Overs/Unders Fund will finish down 5.4c from a -3.5% ROT on a 1.56 turn. It made 78 bets with an average size of 2% of the Fund, and collected on 40 of them. That’s slightly better than the chance level of 50%, but not sufficiently better to eke out a profit at prices in the $1.88 to $1.90 range.

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

Commentary about these forecasts will come in the week's Overs/Unders blog and MoSHPlay’s various forecasts will be posted after the relevant team line-ups are made public.