How Good Are MAFL Predictions?

How Good Are MoS Predictions?

All MoS predictions are attempts to foretell an aspect of an upcoming game of football: which team will win and, for some predictors, by how much or how much more likely are they to win.

While it's tempting to compare the predictive abilities of MoS Tipsters and Predictors against what might be achieved by a naive and clueless prognosticator making predictions at random or on the basis of some very simple rules, the fairest comparison, I think, is to the predictions, albeit sometimes implicit, of a bookmaker. For this purpose I use the TAB Sportsbet bookmaker.

Across the seasons, no single MoS Predictor or Tipster has consistently outperformed the TAB Sportsbet bookmaker, though a few have done so for portions of seasons or even for entire seasons. In 2012, for example, a number of the Head-to-Head Tipsters predicted more game outcomes correctly than did the TAB bookmaker; some Margin Predictors recorded average absolute prediction errors little different from those implied by the bookmaker's prices and spreads, and many correctly predicted the line market winners for more than 55% of games; but the bookmaker's implicit probability assessments were clearly superior to any similar assessments made by MoS algorithms.

For 2011 it's probably fairer to say that the TAB bookmaker demonstrated greater superiority in head-to-head tipping, margin and probability prediction, but some MoS Tipsters and Predictors still managed to stay close - though it was mostly different Tipsters and Predictors that performed well in 2011 compared to those that performed well in 2012.

In terms of the wagering performance of MoS Funds the headline result is that they've not, collectively, turned a profit. It is true to say, however, that they've done better than the TAB Sportsbet bookmaker would have expected in that they've lost at a rate less than the average vig embedded in the prices they've faced. As well, successive Funds wagering in the Line market have shown consistently profitable tendencies - well, at least until the horror 2015 season about which we do not speak.

On balance I think a fair adjudicator would give a points decision to the TAB bookmaker so far, but the bout's not yet over. As I type this, with the 2016 season just a few weeks away, I have genuine enthusiasm about the prospects for a return to profitability. But then, of course, I would ...

(Turns out that 2016 enthusiasm was well-placed - ed.)

Return to MoS Primer | Statistical Analysis, Modelling and Prediction