2015 - Round 15 Results: Better

Only the most speculative of the wagers failed to provide a return this week, the three others combining to lift the value of the Head-to-Head Fund by a little over 6c and, in turn, lift the value of the Overall Portfolio by almost 2.5c. That Portfolio is still down for the season, and considerably so, but the gap to profitability has at least been hammered below 20% again. 

Ultimately, any thoughts of eventual profitability require that the Line Fund become active and start picking winners. We await developments.

TIPS AND PREDICTIONS

In a week that threatened to be problematic for the Head-to-Head Tipsters, with a number of games expected to be close-fought affairs, an all-Tipster average of 6.8 from 9 seems very acceptable.

So closely matched were the teams in two of the games, in fact, that the TAB Bookmaker installed equal-favourites in those games, in which circumstance BKB is assigned the tip of Consult The Ladder. In both cases that was a losing prediction and meant that BKB scored only six from nine for the round, dropping it from a share of the lead into outright fourth.

View fullsize

Bookie_9 and Combo_7, who shared joint-leadership coming into the round with BKB and Bookie_3, fared much better than BKB, their eight from nine scores leaving them as lone joint-leaders now one tip ahead of Bookie_3 and two ahead of BKB.

Other Tipsters recording the round high eight from nine score were the three Win family and three ProPred family Tipsters, H2H_Adjusted_7, and the still distantly-last Home Sweet Home.

Margin prediction also proved relatively easy this week, the all-Predictor Mean Absolute Error (MAE) of 22.8 points per game per Predictor the second-lowest for the season. Best was ENS_Greedy's 19.7 MAE, while six other Predictors recorded sub-22 MAEs, including all five of the top Predictors on the Leaderboard.

That Leaderboard remains congested at the top with the first six Predictors separated by less than four goals and the top three by only about half a goal. C_Marg, the third of the top three, lies third despite having the extraordinary record of being the Predictor least often tipping within three goals of the final result, a feat it has managed in only 34% of all games. Its overall MAE performance has been bolstered by its relatively rarely being a very long way from the final margin. 

View fullsize

C_Marg also has the best line-betting performance of all MoS' Margin Predictors having selected the correct team over 60% of the time. Seven other Predictors, the four H2H family, Bookie_3, Combo_NN2, and ProPred_7, have slightly poorer rates of accuracy but profitable line-betting records nonetheless.

The Head-to-Head Probability Predictors also turned in a solid set of performances, their +0.30 average probability score the highest since Round 10 and the third-highest of the season. Best performed was C_Prob, its +0.37 average allowing it to close the gap to the three directly bookmaker-based Predictors above it. Those three remained in the same order at the end of the round - Bookie-OE, Bookie-RE, Bookie-LPSO - despite Bookie-LPSO recording the best probability score and Bookie-OE the poorest probability score amongst them.

The Line Fund algorithm recorded its worst probability score since Round 7, so Investors should be pleased that the Fund opted not to wager this week.