2012 Finals Week 2 Results: Better But Still Not Super

It's legitimate now, I think, to feel aggrieved by our run of outs on SuperMargin betting. Our two wagers this weekend missed by just 1 and 4 points, meaning that the current sequence of 36 losses in a row now includes 6 losses by less than a goal and 2 more of only single digit proportions. Across the entire season we've now missed 14 SuperMargin wagers by less than a goal and 21 by fewer than 10 points. Small wonder then that the Margin Fund is down by 80c on the season, making it responsible for 16c of the 19c that Overall Portfolios have lost this year.

Had the Margin Fund decided to embark on this losing spree far earlier in the season, the stop-loss rule that I'd set up for determination at the end of Round 12 would have triggered and we'd all be about 6c better off right now. Still, as my Dad was keen on saying when he thought you were uselessly speculating, "your aunty could have been your uncle".

On the positive side, at least both home team favourites did eventually prevail - though each had scares - these wins allowing the Head-to-Head Fund to reap a profit of a little less than 10c on the round. In aggregate then Overall Portfolios rose by just under 2c to leave them at 81c for the season.

Two wins by the two favourites by about the pre-game bookmaker-predicted margins made for another weekend of strong performances by the Head-to-Head Tipsters and Margin Predictors, but to no alteration in the ranking of any of them.

The Margin Predictor results are particularly striking, with even the worst Predictor, Win_7, recording an average MAPE across the two games of just 11.6 points. Best performed for the week was Bookie_9 with an absolute average error of just 3.34 points per game.

Such stunningly precise margin predictions meant that no Predictor was in error by more than one bucket in SuperMargin wagering on either game. Indeed only three Predictors failed to be right in either game: the TAB Bookmaker himself (BA), Combo_7, and the Margin Fund's guru, Combo_NN2.

Three of the Head-to-Head based predictors landed both SuperMargin wagers, continuing the amazing season that they've unfurled, which has seen all four of them select the correct bucket in over 30 games and three of them predict correctly in 36 games or more - half as many again as the TAB Bookmaker himself has correctly predicted.

The non-TAB Probability Predictors were all rewarded for the boldness of their predictions this week, in particular WinPred, which had one of its best scoring rounds this season despite the round being only two games long. This performance has cemented WinPred's 2nd position on the Ladder, widening the gap between it and the H2H Predictors.

With its relatively cautious probability predictions of near 50% for both games this week, the Line Fund algorithm was incapable of recording a probability score very different from zero. In the end it went the wrong way on the Pies v Eagles clash, having the Pies as only 48% chances of covering the spread, which they eventually did, causing the algorithm to record a slightly negative probability score for the round.