It's like the Heritage Fund of a few years back has returned this week, a little older and wiser, chastened perhaps by its GFC experience, but still willing to have a nibble at a rank outsider. 

At least that's how it felt for me as I placed a 0.33% wager on Melbourne at $11 head-to-head. Even if the miracle occurs and the CEO-less Dees, fresh from a 148-point last-round loss to the Dons, topple the West Coast, it'll only be worth 1.7c for Recommended Portfolios. But, somehow that's enough.

Investors have two other Head-to-Head, five Line, and six SuperMargin wagers this week.

I feel compelled to admit that I'm beginning to appreciate the Margin Fund's approach to wagering this season, which sees it opting for the relative safety of twin buckets in many games while confidently plumping for a single bucket twice over in other games. Maybe I've been inordinately affected by the Fund's potentially transitory profitably though.

In any case, here's what that menagerie of wagers represents in terms of risks and rewards:

A Melbourne victory promises the greatest return to Investors then, though a Richmond win by 30-39 points or a Port Adelaide or Geelong win of optimal proportions seem to be more likely pathways to otherwise significant returns. Alternatively, a Richmond draw, loss or victory by 19 points or less, or a Port Adelaide loss by 16 points or more are the worst possible results from an Investor's viewpoint, each representing a loss of about 0.9c from the Recommended Portfolio. A Melbourne loss by 60 points or more, a GWS loss by 28 points or more, or a Gold Coast loss by 16 points or more are almost equally poor outcomes for Investors.

TIPSTERS AND PREDICTORS

Amongst the Head-to-Head Tipsters there's almost twice the disagreement this week as we saw last week, primarily in relation to the Gold Coast v Brisbane, Collingwood v Hawthorn, and Port Adelaide v Adelaide games. 

As usual, it's Home Sweet Home that is the Tipster most different from the All-Predictor average for the Head-to-Head Predictors, though BKB, ProPred and WinPred are also unusually contrarian this week. In three games, the Head-to-Head predictions are predominantly for the Underdogs: the Dons over Freo, the Suns over the Lions, and the Crows over Port. In one other game - the Pies v Hawks contest - the Underdogs have solid Head-to-Head Predictor support.

In total, across all games, the Average Disagreement rate for Head-to-Head Predictors is this week, at 25%, almost double that of last week.

The Mean Absolute Difference of the Margin Predictors from the All-Predictor average is also lower this week, down from 5.9 points per Predictor per game to 5.0 points. Most different are the predictions of Combo_NN1 at 8.7 points per game, while least different are the predictions of Combo_7 at just 2.9 points per game. In general, the standard deviations of predicted victory margins are quite small this week, in all but one game coming in at single digits and in five games coming in at less than 7 points. Combo_NN1 is the contrarian Predictor in three games, the Really Simple, Head-to-Head and Combo_NN2 Predictors in one other game.

There's also more dispersion across the Probability Predictors, the average absolute percentage point difference in their probability assessments this week coming in at 5.2% per Predictor per game as compared to 4.2% last week. ProPred, with a mean difference of 6.7% points is this week's largest contrarian ahead of both Head-to-Head Probability Predictors on about 6.5%. The most conforming Predictor is the Risk-Equalising Bookmaker Probability Predictor with an average difference of just 3.9% points per game.