2013 : Round 1B - Results

At this point in the season all but the Margin Fund wager smaller amounts than they'll wager later in the year. This is because statistical analysis of historical data suggests that the various algorithms tend to perform less well in these earlier rounds; Round 1 of 2013 has provided ample additional evidence for such hypotheses.

The Head-to-Head Fund won only one of its five wagers, landing the Suns' upset win over the Saints but losing on its other bets on the Blues, Giants, Dees and Hawks. That leaves it down 5.5c on the season. The Line Fund finished 2 and 4, dropping 2.8c in aggregate as a result, though it might consider itself a little unlucky to not finish 3 and 3, losing its Roos wager by just 2 points.

In these early rounds of the season our Margin Fund is meant to take up the slack as this is the time that, historically, it's been most prescient. In Round 1 of 2013 the Fund's guiding Margin Predictor, Combo_NN2, did indeed select the correct bucket in two of the nine contests but, unfortunately, its accuracy came in games where it selected an away team victor and so was completely ignored by the Margin Fund. As a consequence, the Fund's 0 and 4 record for the round leaves it down by 5c.

All told Investors have sustained not much damage, however, and find themselves down by just a little less than 4c.

HEAD-TO-HEAD, MARGIN AND PROBABILITY PREDICTION

Only two favourites were successful this week, so Combo_NN2's 4 from 9 record proved sufficient for it to outright lead all Head-to-Head Predictors at the end of Round 1. Its Margin Predictions were also relatively good, leaving it 2nd on the Margin Predictor ladder behind only Combo_NN1, which turned in an impressive MAPE of 33.37 points per game for the round.

Right now it's nice to have four Margin Predictors ahead of any of the more-directly TAB Bookmaker related Predictors in Bookie_3, Bookie_9 and Bookie_LPSO. It is the Bookmaker-derived Probability Predictors, however, that lead the way on their Leaderboard however, filling as they do all podium positions.

The Line Fund algorithm had a fairly subdued start to the season, assigning a probability of greater than 50% to just 2 of the 9 line-betting victors. Fortunately for Investors, both of the successful predictions resulted in wagers - less fortunately, so too did four of the unsuccessful predictions.