2012 Round 21 Results : Our Two Cents Worth

It can only be politely described as misplaced optimism when you spend Sunday afternoon and evening hoping that a team vying for the Spoon can contrive to lose by only 5 or 6 goals to a team vying for the Flag and, in so doing, turn a small wagering loss for the round into a moderate profit.

Well, misplaced that optimism was, so Investors suffered a loss of just over 2c on the round, leaving us down 23c on a season moving rapidly towards the Finals. It's going to take something extraordinary to drag us into profit over the remaining 6 weeks but I've not yet given up hope entirely. I'm buoyed partly by the recognition that the Line Fund is 15 and 11 across the last five rounds, partly by the fact that the Head-to-Head Fund correctly predicted the Lions' upset-win this week (though didn't wager enough on it), and partly by the knowledge that we've been within one bucket of collecting on the SuperMargin market in 19 contests so far this year while actually collecting in only 10. All that, of course, could turn out to be just another example of misplaced optimism ...

Meantime the Head-to-Head Tipsters have racked up another solid round, averaging 6.6 from 9 on the back of wins by 7 of the favourites. Many of the Tipsters had selected those 7 favourites and wound up with scores of 7 for the round, though none did better than this. The round's worst result was, once again, Home Sweet Home's, this week with a 4 from 9. H2H_Unadj_10 was one of the Tipsters picking 7 and so still leads the pack, now on 141 from 180 (78%), and remains a single tip ahead of eight other Tipsters on 140.

It's evidence of how predictable this season's been that, Home Sweet Home aside, the worst Head-to-Head Tipster, Easily Impressed I, which most weeks looks only at the previous week's results to make its picks, is tipping at better than 63%.

Only two Margin Predictors had sub-30 MAPEs for the round, with Bookie_3's 29.10 points per game the better of the two, allowing it to shore up its number 1 ranking on the MAFL Leaderboard where it now holds a season-long 29.26 points per game record. Combo_7, whose MAPE for this week was 31.55 points per game, remains in 2nd as the only other Predictor with a sub-30 season-long performance. In third is Bookie_9 with a 30.07 season-long MAPE having returned a 31.47 points per game result for the round. The all-Predictor average for the round was 31.17 points per game.

Collectively, the Margin Predictors averaged 5.2 correct line picks this week, which means that they've predicted at better-than-chance levels in 5 of the last 7 rounds.

Combo_7 had the best results on SuperMargin markets, selecting the correct bucket in two games and missing by a single bucket in one more. Bookie_3's results were also impressive, though less lucrative, as it selected the correct bucket in just one game but was off by a single bucket in three more. (Combo_NN_2, by the way, also selected the correct bucket in one game but this was unfortunately in a game where it was predicting an Away team victory and so did not result in a wager and subsequent collect for Investors).

Bookie_3 has selected the correct bucket in 21 games this season, the same number as ProPred_7 and Combo_7. This is a considerable distance behind the 27 of H2H_Adj_7, the 30s of H2H_Unadj_3 and H2H_Unadj_10, and the 31 of H2H_Adj_3. Had we wagered on H2H_Adj_3's SuperMargin predictions in every game this season we'd now be enjoying a 31% ROI. It's been right a remarkable 1 game in 6.

All four Head-to-Head Probability Predictors recorded positive probability scores this week, with the TAB Bookmaker, unusually, worst amongst them. Best was WinPred, which moved itself into 2nd place with its second successive best-on-ground performance.

The Line Fund algorithm also earned a positive probability score for the round, the third time it's done so now in the past four weeks.