Of course just the one game this week, which sees the MoS twins in near agreement but forecasting very differently from the bookmakers.

Those bookmakers are forecasting a Total two or three points higher than the all-season-to-date average, while the MoS twins are about 10 points higher still.

WAGERS

That large difference fuelled another MoSSBODS Totals bet for this week.

At the time of placing that bet, the forecast was “medium chance of showers”. It’s now “partly cloudy” and “slight chance of a shower”, but the bookmakers’ Totals lines are unchanged, which means either that they are still expecting rain, or never were.

The estimated overlay on the Cats game is about 11 points.

PREVIOUS WEEK

Last week saw MoSSBODS take mean absolute error (MAE) honours for Away Team Scores and Game Totals, MoSHBODS do the same for Home Team Scores, and the TAB and Sportsbet share honours for Game Margins.

None of which saw the TAB lose the season-long lead for any of the four metrics.

Those leads are now as follows:

Game Margins: the TAB 9 points ahead of Sportsbet, 163 points ahead of MoSHBODS, and 292 points ahead of MoSSBODS
Home Team Scores: TAB 20 points ahead of MoSHBODS, 26 points ahead of Sportsbet, and 127 points ahead of MoSSBODS
Away Team Scores: the TAB 41 points ahead of Sportsbet, 102 points ahead of MoSHBODS, and 145 points ahead of MoSSBODS
Game Totals: 13 points ahead of MoSHBODS, TAB 26 points ahead of Sportsbet, and 71 points ahead of MoSSBODS

The last row of the Errors section of that table is interesting in that it reveals:

Home Team margins were about 0.5 to 0.9 points lower than expected, according to the MoS twins, but 0.9 to 1 point higher than expected according to the bookmakers
For MoSSBODS this was because Home Teams scored 0.5 points per game less than expected and Away Teams scored 0.4 points per game more than expected
For MoSHBODS this was because Home Teams scored 0.2 points per game more than expected and Away Teams scored 0.8 points per game more than expected
For the TAB this was because Home Teams scored 0.3 points per game less than expected and Away Teams scored 1.2 points per game less than expected
For Sportsbet this was because Home Teams scored 0.2 points per game less than expected and Away Teams scored 1.1 points per game less than expected
This meant that MoSSBODS, on average, overestimated Totals by 0.1 points per game, MoSHBODS underestimated Totals by 1 point per game, the TAB overestimated Totals by 1.4 points per game, and Sportsbet overestimated Totals by 1.5 points per game

In all cases, being within about 1.5 points of the true averages suggests that there was little bias in the underlying forecasting process.

Wagering and Game-by-Game Performance

There was one bet last weekend, a successful overs bet with Sportsbet, which left MoSSBODS with a 19 and 19 season-long record against the TAB, and an 8 and 6 season-long record against Sportsbet.

Overall, MoSSBODS was on the right side of the TAB and Sportsbet on both games, while MoSHBODS was on the right side of both in only one game.

That took MoSSBODS to 109 from 215 (51%) against the TAB and Sportsbet, and MoSHBODS to 121 from 215 (56%) against the TAB and 120 from 215 (56%) against Sportsbet.

It’s interesting to look at the performance of MoSSBODS and MoSHBODS based on the size of the estimated overlay against the TAB or Sportsbet.

At the risk of horrid cherrypicking, it turns out that MoSHBODS does best when the estimated overlay is under 6 points. In those games it is 82 from 137 (60%) against the TAB, and 81 from 135 (60%) against Sportsbet.

MoSSBODS does better for estimated overlays between 6 and 12 points where it is 30 from 57 (53%) against the TAB, and 31 from 55 (56%) against Sportsbet.

The round-by-round performance is shown in the following table.