2013 : Round 22 - Wagers & Tips

The Head-to-Head Fund has yet another wager on a longshot this week, but that bet is at least somewhat balanced by a more sizeable wager on the short-priced favourite Lions and by a more modest wager on the underdog Roos.

It's a measure of the Roos' unrewarded talent this season that they, despite sitting 11th on the ladder, take on the likely minor premiers priced at just $3.20.

(Regular readers will probably notice the change of format this week in all of the regular tables from this blog. During the past week I've had to rebuild most of these from scratch because, as I suspected and wrote about last weekend, my sparklines add-in for Excel has proven stubbornly resistant to repair. The new formats are chartless - some might even say charmless too - yet data-rich. I miss my boxplot and stripe charts too, but their reappearance will have to wait until next season.) 

As well as those three Head-to-Head bets, Investors have a stake in four line betting contests this weekend, two on favourites and two on underdogs, and can also look forward to monitoring the highly-variable fates of 14 SuperMargin wagers, including yet another wager on a draw. (We'll get one yet ...)

So, for the first time since Round 12, in which only six games were played, we've a stake in every contest.

Yet again this week the upside is well-spread and often attractive while the downside tends to be dispersed and diminutive. In four games we've a maximum upside of over 2c, and in two more we stand to win over 1c should results go our way. Only one game has a downside of over 2c, however, while four more have downsides of between 1c and 2c.

GWS holds the key to Investors' best and worst possible results. A GWS win over the Tigers represents the most lucrative outcome of the weekend by some distance and would add almost 13c to the Recommended Portfolio, while a GWS loss by 11 goals or more represents the largest potential loss, putting just over 2c at risk.


Since I was rebuilding the table showing the head-to-head predictions I took the opportunity to include the predictions not just of the official Head-to-Head Tipsters but also the implicit head-to-head predictions of the Margin Predictors. I include all these Tipsters' performances in the weekly MAFL Leaderboard anyway, so it made sense to make their predictions more obvious.

This does make for a slightly longer table though ...

The table also now includes the running tally of each predictor to help you assess the credibility of any tipster's contrarian opinions, and on which basis the table is sorted. The Disagreement figure, which appears as the final column, is the same metric as before - the probability that a randomly selected tipster will disagree with the chosen tipster in a randomly selected game - but is now diluted by the inclusion of better-performed and less-dissenting tipsters. Note how small, for example, the Disagreement figures are for the best-performed tipsters.

We now have 27 tipsters and they all broadly agree about the outcome of every game except two: the Geelong v Sydney and St Kilda v Gold Coast games where there are six or seven dissenters, though none amongst the tipsters in 1st and 2nd positions on the ladder.

In the next of the reconstructed tables we see that the Margin Predictors are unanimous in their opinions about every game except the Saints v Suns clash where seven have opted for a win by the Saints (who, by the way, are the favourites). And, once again this week there's also broad agreement about the likely size of these victories, with the standard deviation about the all-Predictor average margin climbing into double digits in just two games.

The Head-to-Head Probability Predictors are also of one mind in every game except the Saints v Suns clash where the directly Bookmaker-derived Predictors are siding with the Saints and the remaining Predictors are siding with the Suns. That level of disagreement has resulted in this game having the largest range of probability estimates (20%), though the range is almost as large at 14-18% in five more contests. WinPred is the week's boldest Predictor, and ProPred the most conformist.

Four teams are rated better than 60% chances on the line market according to the Line Fund algorithm, and three more are rated almost as highly at 59%. That should make for large positive or negative probability scores for this algorithm this week.