2015 - Round 14 Results: Not Yet

After a round in which the football was very much secondary to the tragedy that proceeded it, it seems almost inappropriate to fuss about its results. But fuss about them I will in a moment and at my usual length - because that's what we do around here. Before that though a couple of links to particular memorable pieces that caught my attention over the weekend.

The first is by Andrew Gigacz, published on AustralianFootball.com, and the second by Titus O'Reily from his self-titled TitusOReily.com website. They're both wonderful pieces in their own, distinctive ways and I recommend you spend time in their company. They will linger.

To matters football then, where I'm unhappy and insufficiently surprised to report yet another week's unprofitable wagering. It all started well, especially for the Line Fund, its two-from-two record after five of the round's eight games enough to see Investors in black ink during a round for the first time since Round 4. This despite the fact that the Head-to-Head Fund had, earlier, landed only one of its two wagers before racking its cue and registering a small loss for the week. On Saturday night, the Dogs' narrow inability to cover their 12.5 point spread dragged Investors nearer an overall loss, and the Dons' capitulation to the Saints on Sunday then finished the job.

In the end the Head-to-Head Fund finished down by 1.6c, the Line Fund down by 1c, and the Overall Portfolio down by 1.2c. That Portfolio is now worth 21.9% less than it was at the start of the season.

TIPS AND PREDICTIONS 

Just two favourites lost this week, the Roos and the Dons, which established 6 from 8 as the weekend's benchmark result for the Head-to-Head Tipsters. Twenty six of them achieved that benchmark, two (Easily Impressed I and Home Sweet Home) fell a tip short, and two more (Short Term Memory I and II) went a tip over, the mathematical symmetry of these collective results producing an all-Tipster average of exactly 6. That 75% accuracy rate is the fourth-highest for a single round this season.

With the top 23 Tipsters all recording the benchmark score, there was little movement on the Tipster Leaderboard for the Head-to-Head Tipsters. That Leaderboard remains headed by four Tipsters, now all on 81 from 116, which is about a 70% accuracy rate.

View fullsize

There was, though, some movement amongst the Margin Predictors, Bookie_9 doing just enough relative to its upstairs neighbours to climb over Combo_7 and the two RSMP Predictors into 2nd. It's now less than a single point behind C_Marg, whose Mean Absolute Error (MAE) of 39.5 was the worst of any of the Predictors this week. Such is its nature that C_Marg, I fear, is always going to be like this.

The best MAE was H2H_Unadj_3's 30.2 points per game, it and its three siblings now with Line Betting records marginally profitable for the season as a whole. They join C_Marg and Combo_NN2 in having that status.

View fullsize

The all-Predictor average of 34.2 points per game per Predictor this week was the seventh-highest average for the season. It's hard to turn in a sub-30 MAE when one favourite contrives to lose by 55 points and another does the same but by twice the margin.

Neither of the losing favourites was short-priced though, which allowed Head-to-Head Probability Predictors to record another set of good scores - only slightly worse, in fact, than last weeks' set. The four best Predictors recorded the week's four best scores, but in reverse order, with C_Prob registering the round's highest score and Bookie-OE registering the fourth-highest score. That had no effect on their laddering ordering, however, which still sees Bookie-OE heading all-comers.

The Line Fund algorithm, meantime, recorded a near-zero probability score, an indication, Investors will hope, of improved levels of calibration and discernment.