2012 Finals Week 4 Results: No Collect, Not Much Surprise

When the Hawks goalled (one 'l' or two? It's so hard to know for these verbed nouns) near the end of the first term to finish it leading by 19 points, our SuperMargin wager looked promising, at risk mostly from a rampaging Hawks dominating the second term and establishing a lead that would put our preferred 30 to 39 point victory margin out of range. But then the Swans did as the Swans usually do, finding a way to shut down the opposition and work their way back into the contest. So effective were they, in fact, that they eventually established a 27-point lead just over 8 minutes into the third term.

And then it was the Hawks turn to surge, giving Investors hope once more when they goalled to establish a 12-point lead 11 minutes into the final term, before a run of 19 unanswered points from the Swans over the course of the next eleven and a half minutes finally put paid (well unpaid really) to the week's sole wager.

The season finishes then with the Head-to-Head Fund having made 33 wagers for 18 collects. Across those bets it turned the Fund just over 1.5 times which, if we assume an average 3 to 5% vig on each dollar invested, suggests an "expected" loss of between about 4.5 and 7.5% for the Fund assuming it's about as good at the TAB Bookmaker at assessing team probabilities in the games on which it wagers. The Fund actually finished down by 6.3%, which is pretty much in the middle of that expected range. 

Next let's review the Line Fund, which made 95 bets for 50 wins, turning the Fund 4.75 times in the process. With an average vig on Line Bets of about 5.25%, the expected loss for this Fund was about 25%. Compared to that benchmark, the Fund's break even result has some merit - but this Fund, nor any other Fund, was not created merely to lose less than it "should".

That leaves only the Margin Fund's performance to review. It made 114 wagers for just 10 collects (and a frustrating 23 near-collects) this seasonwhile it turned the Fund 2.85 times, losing 85% as it did so. Estimating an average vig for wagers in this market is difficult as the total overround is generally vast - 40% or more - but is likely to have been spread more thinly on the wagers that this Fund generally preferred, which were those shortest priced. If we assume a range of 20 to 40% for the actual vig, then the Fund's expected loss was somewhere between 57% and 114%. That also puts this Fund's actual loss at about the middle of the expected range.

After we apply the relevant Fund weightings to these performances we come up with figures for the Overall Portfolio of 242 bets for 78 collects, a turn of 3.4 times, and an actual loss of 19.4% compared with a range of expected losses of between 25 and 38%. In short, we lost - far more heavily than I'd hoped, somewhat less heavily than the TAB expected we would, and far less heavily than we might have given the activity levels of the underlying Funds. There's the "glass half empty", "glass half full", and "glass amazingly still here" views of this year's MAFL wagering performance.

Moving next to tipping, none of the Head-to-Head Tipsters picked the upset, so the final record for the season's best Tipster, H2H_Unadj_10, was 162.5 and 44.5 (78.5%). Amongst the Tipsters still active across the Finals series, all but Home Sweet Home finished with season-long accuracy levels of 73% or higher. That is a truly remarkable result.

Bookie_3 was, yet again, the best Margin Predictor for the season, though it was only one of seven Predictors ending 2012 with sub-30 MAPEs. That too is truly remarkable, and means that not only were this season's winners generally predictable, but so too were their margins of victory. One dramatic piece of evidence in support of this is the fact that three of those top seven Margin Predictors were able to predict the correct game margin to within a single goal in almost one game in five.

To continue the chain of remarkable outcomes, all but three of the Margin Predictors finished the season with line betting records that exceeded chance, eight of them by so much that following their predictions would have made for especially lucrative line wagering. Predicting line outcomes with 56% accuracy across an entire season of 207 games is positively oracular.

The TAB Bookmaker finished highest amongst the Probability Predictors, its average score of 0.260 per game just a little below its 0.278 per game score from last season. WinPred finished highest of the remaining Head-to-Head Probability Predictors, improving significantly on its fifth-placed finish of 2011.

Another Probability Predictor to improve this year was the Line Fund algorithm, which followed last year's average score of -0.050 per game with an average this year of just -0.027.

Wagering Results by Team

It's only been a few weeks since I looked at the Fund's wagering performance with each team, both when wagered on and when wagered against, so here I'll just, for completeness' sake, provide the final tables for your review.

If we weight these results by team across Funds based on how the Funds are weighted in the Overall Portfolio and apply a 50:50 weighting to the results for each team when they're wagered on versus versus when they're wagered against, we obtain the following table.

It's interesting that the largest returns this season came from games in which teams from outside the top 8 were involved. Amongst the finalists, only the Eagles, Cats and Roos made positive contributions to the Overall Portfolio.

On the other hand, the second- and fourth-largest losses are attributable to two other teams from the top 8, the Pies and the Hawks, which between them might be said to be responsible for over 13c of the 19c we lost this year.

One thing I am pleased about is the fact that we wound up making money on the new entrants, the Giants, which suggests that my decision to MARS Rate them below 1,000 at the start of the season was the right thing to do.

Some other things I'd note from this table are the stark differences in the returns when wagering for versus when wagering against the Roos (+7.6% vs -6.6%), the Swans (+4.8% vs -10.4%), the Saints (-11.8% vs +1.8%), and the Hawks (+1.8% vs 12.4%).

Thank you to everyone who's dropped by MAFLOnline this season. I've not yet decided if I'll be returning next season but, if I do, I hope to see you here again.