2013 : Round 3 - Results

/

Another successful SuperMargin wager, a 2 from 3 performance from the Head-to-Head Fund, and a 3 from 5 record for the Line Fund was enough this week to add 1.3c back to the price of the Recommended Portfolio, leaving it now down by just 1.2c on the season.

(Note that the Portfolio price shown for the past two weeks has been slightly in error as I'd been adding the ROIs for the two SuperMargin Predictors, Combo_NN2 and Bookie_9, rather than averaging them. Practically, that means our loss in Round 1 and our profit in Round 2 were both smaller than I reported.)

The result could easily have been ever better for Investors as Freo failed by just a kick on Friday night to deliver us another SuperMargin victory as they lost by 4 points to the Dons. Still, having gone over 30 games in the back end of season 2012 without a SuperMargin success, I'm very satisfied with Combo_NN2's 3 from 10 record in 2013.

TIPSTERS AND PREDICTORS

The new Margin Predictors continue to perform well, so much so that three of them - the two RSMP Predictors and Bookie_LPSO - occupy positions 2nd through 4th, behind Combo_NN2 whose predictive accuracy is what's been partly fuelling the profit-making proclivities of the Margin Fund.

Combo_NN2 also leads the Head-to-Head Tipping race along with the strangest of bedfellows in the two Easily Impressed Heuristic Tipsters. They're all on 20 from 27 (74%).

Bookie - RE, the Predictor whose Home team probability assessments are made based on a particular assumption about how overround is embedded in TAB head-to-head prices (see this blog for details), leads all-comers in the Probability Predictor race. In second place is Bookie - OE, whose Home team probability assessments are made based on a different assumption about how overround is embedded in TAB head-to-head prices - specifically, that overround is embedded equally in each team's price.

The Line Fund algorithm's Probability Assessments have yielded a very slightly negative Probability Score so far this season.

MARGIN PREDICTORS AND SUPERMARGIN PERFORMANCES

Combo_NN2 has been having quite a year in the SuperMargin market. In fact, as the chart below shows, it has selected the correct bucket in 6 of the 27 contests so far this season, 3 of the 10 times when it has selected the Home team to win (and so the Margin Fund has wagered), and 3 of the 17 times when it has selected the Away team to win.

That overall performance is as good as any other Margin Predictor in MAFL.

The other Predictors to have made 6 correct selections are: 

  • BA (Bookie Actual), which selects its bucket on the basis of the points start from the line market
  • BL, which is based on the Bookie-LPSO Margin Predictor
  • BRS, which is based on the Really Simple Margin Predictor using Simple Averaging
  • BRW, which is based on the Really Simple Margin Predictor using Weighted Averages
  • B3, which is based on the Bookie_3 Margin Predictor
  • PP3, which is based on the ProPred_3 Margin Predictor
  • W7, which is based on the WinPred_7 Margin Predictor 

Amongst these Predictors, PP3 has the best ROI because it has selected winning buckets paying $7.50 on four of six occasions. Combo_NN2 has the next-highest ROI having secured prices above $7 on its winning wagers less often.

So far it's been an unusually good year for predicting the right bucket in the SuperMargin market. Long may it continue.