MAFL 2010 : Round 20 Results

Meh.

It's weekends like this one that make me remember a conversation that took place between two of my then team-members outside my office one morning some years ago when things were busy and nerves were generally a little frayed.

One team member, arriving at the desk of the other and clearly unhappy about something that had come to his attention in the last 12 milliseconds or so - the approximate time I imagine it had taken him to traverse the distance between his desk and his colleague's - felt obliged to 'share' his unhappiness with the desk's owner who patiently listened for about a minute and then delivered this gem, part parental, part 'get a grip': "Calm down. Nobody's died; nobody's been hurt".

And so it was this weekend for MAFL Investors, though a few might well feel that their chances of a profitable season did die at least a little over the course of proceedings.

All told, the Recommended Portfolio shed 7% to be down by just under 10% on the season, and MIN#002's Portfolio dropped 11.5% to be down by almost 25% on the season. MIN#017's Portfolio fell least of all, by just 0.1%, but is still down by almost 50% on the season and must surely now have chances for attaining profitability roughly equal to the chances of West Coast avoiding the Spoon.

Whilst five results contributed to the loss suffered by the Recommended Portfolio, I hold Freo and the Lions most culpable for the outcome. Together, their losses meant the difference between an increase of a couple of percent and the Fund's eventual 7% loss, and both teams had ample opportunity to win the games that they ultimately lost.

Anyway, here's the latest Dashboard:

That's now two weeks of losses in a row for the Heuristic-Based Fund, so another loss in Round 21 would see Short-Term Memory cede control of the Fund to another heuristic for the final week of the home-and-away season. At this point Easily Impressed I would be that heuristic as it has the best season-long level-stake home team only wagering performance.

Enough talk of wagering and loss, let's move on to MAFL Tipping.

Easily Impressed II scored just 4 from 8 this week, but that was enough to for it to hold outright 1st on the MAFL Tipping table. It's now on 105 from 160 (66%), one tip ahead of Consult The Ladder (which recorded 7 from 8, the week's best result), ELO, Follow The Streak, Silhouette, and Short-Term Memory II. BKB scored 6 from 8 this week but still finds itself four tips off the pace.

On average, the MAFL Tippers scored 4.69 correct tips in Round 20, which is about 0.2 tips better than their collective performance in Round 19.

Only four teams changed positions on the MARS Rating table. Geelong grabbed 1st spot from Collingwood, despite both teams winning handsomely in their respective encounters. Geelong, however, defeated a 1,040+ rated team, while the Pies trounced a team rated around 977, so the similarly-sized victory margins were quite rightly differentially rewarded in terms of MARS Rating points. Both teams are now rated the highest they've been all season.

Sydney and Freo were the other teams to swap MARS positions, with Sydney's victory over Freo being enough to propel them ahead of Freo and into 7th spot on the MARS ladder.

Season-long Mean Absolute Prediction Errors (MAPEs) took another beating this weekend, so much so that, for the first time this year, all Margin Tippers have MAPEs above 30. BKB retains the lead on the MAPE measure, however, and ELO is now 2nd having moved ahead of LAMP by dint of going backwards more slowly.

Median APEs were for the most part unaffected by the weekend's result, the only exception being LAMP's Median APE, which rose by half a point. Nonetheless, LAMP still leads on this metric, with HAMP and ELO still sharing 2nd.

HELP recorded its worst line-betting result of the season, correctly predicting just 1 from 8 results - a performance which, I can hear my Dad proclaim, "You couldn't do if you tried". (Of course if you could achieve this result intentionally, you'd be far better flipping your predictions and bagging 7 from 8, so the point here is probably moot, but never mind.)

To complete the tale of woe, SSM fared little better on line tipping, scoring just 2 from 8, and even our only notionally wagering model, the Bookie Probability Model, would have lost money this weekend, despite landing two of its three wagers. SSM did though manage 6 from 8 on head-to-head tipping.