2013 : Round 4 - Results

Like the tortoise in one of Zeno's paradoxes, Investors once again halved the distance between themselves and their Achilles: the profitability of the Recommended Portfolio. That Portfolio is now priced at 99.5c.

The Head-to-Head Fund is the sole reason they're in the red this year. It dropped another 0.7c this weekend with a 1 from 3 performance, the only of the three Funds to register a net loss over the nine games. It's now down 6.5c on the season, which with its 30% weighting makes it responsible for knocking almost 2c off the value of the Recommended Portfolio.

Meantime the Line Fund has finally dragged itself into net profitability. It snagged 2 from 3 this week, adding 1c to its value and pushing it to a net positive 0.1c for the season having called the team correctly in 9 of the 17 contests on which it proferred a wager-worthy opinion.

Comfortably the star of the show again this week though - as well as for the season so far - was Combo_NN2, whose two more wins on SuperMargin betting added almost 11c to its notional price. Unfortunately, Bookie_9, who I added this year as an insurance for the notoriously-underperforming Combo_NN2 of the latter part of last season, lost all six of its SuperMargin wagers, which sliced 7.5c from its notional price. The net result of these two contrasting performances was an increase in the value of the Margin Fund by just 1.6c. It's now up by almost 14c on the season, which even with its paltry 10% weighting in the Recommended Portfolio, represents a +1.4c contribution to the value of that Portfolio.


Favourites prevailed in all but two of the week's contests, which ensured relatively strong performances by most of the Head-to-Head Tipsters. Best amongst them were the WinPred pair, who both tipped the upset Cats win (though not the Blues') and so bagged eight from nine. 

Combo_NN2 scored seven from nine and still leads All-Tipsters on the Leaderboard by a single correct prediction.

Combo_NN2 also heads the Margin Predictors having returned a creditable mean absolute prediction error (MAPE) of just over 26 points per game this week, the third-best score of the round, behind only Bookie_3's score of 25.09 and Win_7's score of 25.90.

The two RSMP-based Predictors still remain well-placed in terms of season-MAPE, occupying 3rd and 5th spots between them.They also have the two best line betting performances amongst all the Margin Predictors, RSMP-Weighted having selected correctly over two-thirds of the time, and RSMP-Simple having done so almost 60% of the time (a performance matched only by Combo_NN2).

The Bookmaker-derived Probability Predictors continue to dominate that part of the Leaderboard, the Risk-Equalising variant leading all-comers at this stage. The Head-to-Head-based Predictors are performing less well, one unfortunate symptom of which is the loss-making status of the Head-to-Head Fund.

Finally, note that the Line Fund algorithm is showing a net positive average probability score for the season to-date, a position I can't recall it holding at any stage of any previous season. Its success this week came from landing three of the four confident predictions in which it attached about a 65% probability to a particular line betting outcome.

Strong performances as measured by the probability score metric are generally associated with profitable Kelly-staking outcomes and, indeed, had we been Kelly-staking in every game on the basis of the Line Fund algorithm's probability assessments we'd have enjoyed an ROI of +17.3% so far this season. Hindsight, of course, is always a wonderful thing.