2014 - Round 17 : The 13.5% Goal

/

I can't remember a game in the history of MatterOfStats (or of MAFL, as MoS was formerly known) where the range of potential Recommended Portfolio outcomes from a single game spanned a 13.5c range, let alone one where Investors could swing from one end of that range to the other on the basis of just one opportune or inopportune goal.

That is, however, what Friday night's game presents us with, so it's inevitable that this game will heavily colour Investors' ultimate assessment of the weekend's proceedings.

In that game we've one of our three Head-to-Head wagers, though we might have had two more wagers of this type had we bet earlier in the week when Richmond and Essendon were at shorter odds. The lengthening of those odds was, apparently, in both cases enough to convince the Head-to-Head Fund that its previous assessment of the value in those two teams in the Head-to-Head market was erroneous. Go figure.

Still, it's not as if we're lacking for action. The Line Fund, despite a disappointing 1 from 6 performance last week, has made a record seven wagers, thereby putting over a third of the Line Fund at risk on five teams receiving start and only two being made to offer it.

The Margin Fund has been relatively restrained, making just six wagers of its own, though it's inflicted three Chasms of Death including one with an extended, 20-point valley floor.

To make the obvious point, the result in the Crows v Hawks game will be season-defining for the Portfolio, but there are six other games that also have best- and worst-case outcomes that span over 5c of Portfolio value. It's only at the back-end of the Round, when our fiscal fate will already have been largely determined, that we'll get to enjoy some games that mean nothing at all to Portfolio (the Dons v Pies game) or nothing much (the Fremantle v GWS game).

Across the nine games, the best and worst scenarios span almost a 50c range, from a high of +24.3c to a low of -23.7c.

TIPS AND PREDICTIONS

It is, of course, the Head-to-Head Fund algorithm that has precipitated our large wager on the Crows this week, and its enthusiasm has been transmitted to the H2H Margin Predictors. They have, variously, predicted that the Crows will win by between 12 and 22 points and have been joined in that opinion by the Win and ProPred pairs of Predictors.

In only two other games are there contrarian Tipsters: the Dogs v Suns game where the two ProPred Tipsters are predicting very narrow Dogs wins, and the Dons v Pies game where Combo_NN_1 is alone in tipping a Dons upset victory.

Combo_NN_2 is Extreme Predictor in three contests this week and is also owner of the largest Mean Absolute Deviation (MAD) from the all-Tipster average. Bookie_3, also Extreme Predictor in three contests, has the second-highest MAD, while RSMP_Simple finds itself in the unaccustomed position of being Extreme Predictor in one contest and registering the third-highest MAD. C_Marg, for a change, lies outside the MAD medals in fourth.

Turning next to the Head-to-Head Tipsters we find considerable disagreement there too about the Crows v Hawks game, with 16 of the 28 Tipsters opting for the upset Crows victory. In the Dees v Cats, and in the Swans v Blues games the only dissenter is Home Sweet Home, while in the Dogs v Suns, Lions v Eagles, Tigers v Port, and Dons v Pies games there are relatively minor levels of underdog support, a majority of which comes from the less well-performed of the Tipsters.

Home Sweet Home is in its usual place as the Tipster with the greatest Disagreement Index, though this week it's only narrowly ahead of Easily Impressed I and just a little further ahead of Short Term Memory I.

The Head-to-Head Probability Predictors are also split in the Crows v Hawks game, 4-4 in their case and with a range of probability predictions spanning a whopping 43%. In seven of the other contests the Predictors are unanimously behind the bookmaker favourites. Only in the Dogs v Suns game is there any disagreement at all, and that's down to ProPred's assessment of the Dogs as 51% chances.

C_Prob, establishing new extremes of extremism, is Extreme Predictor in five contests this week and, consequently, has easily the round's highest MAD. WinPred, Extreme Predictor in four contests, has the second-highest MAD, while ProPred, Extreme in two, lies third on MAD.

The Line Fund algorithm is a little less sharp than it has been in recent weeks but still offers six assessments of about 60% or more the sharpest of which is its 82% assessment of the Pies' chances.

ChiPS PREDICTIONS

ChiPS continues to be almost entirely guided in its margin predictions by the relative Ratings of the competing teams, with the correlation between the Predicted Margins and Ratings Difference this week standing at +0.98.

Only one team is deemed worthy of a wager, Melbourne at $6.75, on whom ChiPS has lavished a 1.2% wager.