This blog is a day delayed - though not, I'm pleased to say, as a consequence of any dissatisfaction by me with the wagering outcome - because I've been away on business in NZ and refused to pay the usurious rate that the hotel there was demanding of its inmates for internet access. Even the AUD/NZD conversion rate couldn't make that transaction worthwhile financially.
In short, this week Investors made a 1.4c profit on the back of a single successful Line wager and a couple of astute-in-hindsight SuperMargin bets. That's left the Recommended Portfolio down by just 2.3c on the season.
TIPS AND PREDICTIONS
Seven Head-to-Head Tipsters picked the card this week and a few more bagged five, which helped lift the all-Tipster average to 4.8 from 6 for the Round and which represents the highest average accuracy for a single round this season. Low score was 3 from 6, recorded by a handful of Tipsters, simultaneously justifying and prolonging their bottom-of-the-table positions.
Win_7 fared best amongst the Margin Predictors with a Mean Absolute Prediction Error (MAPE) of 20.4 points per game, some 7 points per game better than the last-placed C_Marg and also marginally better than Combo_7 and Bookie_3 who sit above Win_7, though both only narrowly, on the MoS Leaderboard.
Whilst C_Marg remains resolutely and forlornly in last place amongst the Margin Predictors, C_Prob is doing its bit to uphold the Chi legacy, and this week leap-frogged Bookie_LPSO to grab 3rd place amongst the Probability Predictors. Sometimes there's a reward for contrarianism.
The ProPred and WinPred Predictors were the only ones to select the correct bucket more than once this weekend, all four of them choosing correctly in two of the six games. For none of them though was this sufficient to drag them into profitability, a state still enjoyed solely by Combo_NN1 and even then only if we take heed of its predictions when it opts for a Home team win or draw.
In other probability scoring news, the Line Fund algorithm turned in yet another positive probability score this week, extending its streak of above-zero scores to 4 rounds. The only depressing point about this fact is that is means season-long Kelly wagering on this algorithm's outputs would have provided a 25% ROI had wagering been confined to Home team and a 45% ROI had it been confined to Away teams. As always, of course, regret is understandable is retrospect and hindsight is 20:20.