Every so often I need a round like Round 15 to maintain my belief - or preserve my delusion, you decide which - in the efficacy of a statistical approach to wagering.

All three Funds made a profit across the round, the Head-to-Head Fund landing its sole wager on the Saints to rise by just over a cent, the Line Fund landing five of its seven wagers to climb by 12.5c, and the Margin Fund bagging its final wager of the weekend on Freo by 30 to 39 points - having also almost landed an earlier wager on the Swans - to finish up by 7.5c. Combined, this litany of non-errors was worth over 8c to the Overall Portfolio.

So, after fifteen rounds, the Head-to-Head Fund is down by less than 1c, the Line Fund is down by only about 4c, and the Margin Fund is 2.5c in profit, which means that the Overall Portfolio is down by less than 2c and the odds on a profitable season have shortened considerably. It's probably not yet, however, on the top line of that particular market.

This profitable weekend of wagering came as seven of the nine favourites snagged the competition points, the losses by the Pies and the Dons dragging the all-Tipster average down to 6.4 from 9 for the week. Various tipsters recorded the week's best score of 7, while Home Sweet Home, Shadow and Short Term Memory II managed just 4, the week's worst performance.

H2H_Unadj_10 still leads out all Tipsters, now on 98 from 126 (78%), remaining one tip ahead of a peloton of a dozen other Tipsters, BKB not yet amongst them.

The two upset losses, especially the Saints' 71-point victory over the Dons, coupled with the Hawks' thumping win over the Giants, all served to elevate the all-Predictor average MAPE for the week. At 39.31 points per game, this average was the 2nd-highest recorded for a single round this season.

Best MAPE for the week belonged to Combo_7 at 35.19 points per game, a performance that cemented its 1st place on the ladder and left Bookie_9 and Bookie_3 in 2nd and 3rd places respectively with MAPEs for the week 2 to 4 points per game higher. The week's worst performance went to Combo_NN2, who recorded a 44.17 MAPE, its worst single-round performance this season, surpassing its 43.13 MAPE from Round 14.

Despite the high MAPEs, the Margin Predictors correctly predicted an average 5.2 of the 9 line results, and all but two of them selected the correct SuperMargin bucket in at least one game. Bookie_9, in fact, picked the right bucket in 3 games and was out by a single bucket in another, while Combo_NN2 was bucket-perfect twice and off-by-one in one other game.

Across the entire season, the various H2H Margin Predictors have selected the correct bucket in between about 17.5% to 19% of games; Combo_7 and Combo_NN2 in about 13.5% of games; and Bookie_9, the TAB Bookmaker, and ProPred_7 in about 12.7% of games. W3 and W7 have both been within one bucket in one-third or more of all games, but been bucket-perfect in only 6% or 7%.

All Head-to-Head Probability Predictors recorded positive probability scores for the round, best among them H2H_Adj, which has nudged its way into 2nd behind the TAB Bookmaker as a result.

The Line Fund algorithm also recorded a relatively rare positive probability score - only the fifth time it has done so this season.