2014 - Round 13 Results : Four in a Row

Thanks to another impressive performance by the Line Fund and a 1 from 1 result for the Head-to-Head Fund, Investors enjoyed their best round of the season this weekend and their fourth successive week of profitability. The Head-to-Head Fund rose by 3.6c on the back of its lone successful wager on the Crows, while the Line Fund spiked up by 9c after selecting the correct team three times out of four.

Only the Margin Fund registered a loss, landing just one of its 12 wagers to drop by 4.5c, but it was near enough in two other contests for me to give it a pass on the round.

All up, the Recommended Portfolio grew by 4.7c during the weekend to end it 6.9c in profit for the season, which is the highest it's been since Game 7 of Round 2 and the highest it's been at the end of any round all season.

What's especially pleasing is the recent form of the Line Fund, which has a 16 and 6 record over the last six rounds and during which it's risen in value by 30c.

Ah, if only we didn't have the Margin Fund's 41c loss for the season dragging us down (he said, cherry-picking his data adeptly enough to represent Australia in the discipline).


This week, the round went largely to script through to the end of Game 4, with all four of the pre-game favourites eventually, if only narrowly in the case of the Swans and the Eagles, collecting the 4 points. After that though, the underdogs prevailed, winning four of the remaining five contests to plunge BKB into a 5 from 9 performance for the round and resulting in an all-Tipster average on MatterOfStats of just 5.02 tips. 

Bookie_9 recorded six correct predictions as did Combo_7, which meant that these two Tipsters remained in 1st and 2nd places respectively. Bookie_3, who'd tipped a draw in one of the games this week, registered a score of 5.5, which saw it finish the round level with Win_3, Win_7 and WinPred, all on 76 from 108 (70%).

The H2H group of Tipsters all recorded the round-high score of 7 while as a group, BKB aside, the Heuristic Tipsters tipped poorly and now find themselves between 13 and 22 tips adrift of the Bookie_9, and at least 7 tips adrift of the Tipsters directly above them, which include that H2H cohort.

For a time this weekend it seemed that sub-20 Mean Absolute Prediction Errors (MAPEs) were a possibility for the round, but the later results, including the Cats' 96-point thrashing of the Saints, make MAPEs as low as that ultimately unattainable.

In the end the best MAPE that any Predictor could muster was H2H_Unadj_10's score of 23.0 points per game, and the average MAPE came in at 26.2 points per game. That performance was good enough to elevate H2H_Unadj_10 into 6th place on the MatterOfStats Leaderboard, its highest ranking since Round 5. Further up the table, the top 3 positions were left unchanged by the weeks' results. The more directly Bookie-based Predictors continue to lurk.

Seven Margin Predictors now have line betting records sufficient for them to have generated a profit had they secured $1.90 for all their implicit line market selections this season: Win_7, Win_3, Combo_7, H2H_Unadj_10, H2H_Adj_3, H2H_Adj_7 and H2H_Unadj_3.

Three of those Predictors - H2H_Unadj_10, H2H_Adj_3 and H2H_Unadj_3 - selected the correct SuperMargin bucket in four (yes, four) of the week's contest, while Combo_7 guessed correctly in three, and H2H_Adj_7 in two. As a result the four H2H Predictors are all now in profit on SuperMargin betting for the season as a whole, and Combo_7 is at breakeven.

Turning our attention to Head-to-Head Probability prediction we find that the upset results have had a significant impact on the middle placings on the MatterOfStats Leaderboard, where WinPred has now slipped into 3rd after recording the round's best score, relegating Bookie-LPSO to 4th and C_Prob to 5th.

The Line Fund algorithm registered another positive probability score for the week, extending its run to two weeks and preserving its season-long positive probability score. What's been especially unusual about the performance of this algorithm this season has been its ability to predict away team line market victories better than home team line market victories. Had we Kelly-staked the algorithm's predictions in every game this season, our ROI on home team line victories would be 20% and on away team line victories 38%. 

Hindsight is again though, of course, a fine thing indeed.