2016 - Round 6 Results - Line Lead

A small profit this week, all of it thanks to the Line Fund, whose 3 from 4 performance added another 5c to that Fund's value. The Overs/Unders Fund fared worst, dropping a little over 2c on a 2 from 5 performance, while the Head-to-Head Fund shed about 1c after just 1 successful wager from 3 attempts.

All up that meant a 1.3c increase in the value of the Overall Portfolio, which now stands up 23.8c on the season.

I mentioned in passing in the earlier blog where I revealed the weekend's wagers that I'd had a lesson in the need to check carefully one's inputs while wagering for the week. It's always a hurried time once the full set of line markets are posted as I'm keen to determine and place the Head-to-Head's and Line Fund's bets as quickly as possible before the market has a chance to move.

This week, in my haste, I failed to update a portion of a spreadsheet, which meant that the wagering opportunities were assessed for the current round of matches using last week's prices. That made a number of propositions look very attractive - suspiciously so in hindsight - and I soon found myself placing a number of significant wagers, nine in all, on five different teams. I'd say it was about a minute after I'd placed those bets that I suspected what I'd done. Existential dread has nothing on that feeling.

As a result of my carelessness, I ended up with unwarranted head-to-head wagers on Adelaide, Geelong, Carlton and West Coast, unwarranted line wagers on Adelaide, Geelong and Carlton as well, plus larger-than-warranted line wagers on Richmond and West Coast. Fortunately, West Coast's last quarter push was enough to land the head-to-head and line wagers on them and left me with only an $8 loss for the whole nightmare.

Lesson, I hope, learned.


Only six favourites won this week, though the levels of dissension amongst the Head-to-Head Tipsters was sufficiently high that we saw an all-Tipster average of 6.6 correct tips from 9. Five Tipsters managed scores of 7, while the other four scored 6.

That left MoSSBODS_Marg two tips clear at the top of the Leaderboard on 41 from 54 (76%).

Some very large and unexpected victory margins pushed mean absolute errors (MAEs) higher this week, resulting in an all-Predictor average of 36.0 points per game per Predictor. Best was ENS_Greedy_MoSSBODS' 32.4 points per game, ahead of MoSSBODS_Marg's 33.1 points per game, a result good enough to leave it atop the Leaderboard as one of only two Predictors - C_Marg being the other - with a season MAE of under 30 points per game.

All but two of the Margin Predictors, the RSMP pair, now have season-long better-than-chance Line market records, Bookie_LPSO the standout having correctly selected 63% of the time.

Best amongst the Head-to-Head Probability Predictors was C_Marg, though its performance was not quite good enough to lever MoSSBODS_Prob from the head of that Leaderboard. Somewhat amazingly, MoSSBODS- and ChiPS-based Tipsters and Predictors now occupy 1st and 2nd positions in all three MoS competitions, ahead of the bookmaker-based Tipsters and Predictors in every case.

For now I think the evidence supports the belief that it is possible to forecast and wager using statistical models that include no bookmaker data as inputs. For now, for now ...