2011 Round 9 Results: A Small Setback

Richmond and Hawthorn brightened what was an otherwise fairly dull weekend for Investors, though a typo in my Ready Reckoner for the Port Adelaide v Fremantle game meant that the total loss was only just over 3% for the round and not 5% as suggested by the Ready Reckoner.

The Head-to-Head Fund landed 2 of 4 wagers, jumping 4% in the process to move back into profitability for the first time since Round 1. Meantime, the Line Fund managed just 3 from 8, shedding a tick over 10% to leave it down just over 13% for the season and leaving Portfolios down 5.7% for the season.

Whilst this week's performance wasn't one for the scrapbook, I'm not entirely unhappy with the results for this and the preceding two rounds taken as a whole. This week, Melbourne were within a kick of winning on Line Betting and the Pies, having spotted the Crows an absurd start quite late in the game, came within 2 goals of covering the 52.5 points spread regardless. As well, the Head-to-Head Fund - along with, it would seem, a significant number of other now bona fide "smart money" wagerers, most of them with the decided advantage of sentience - immediately spotted the value in the Tigers.

What's more, Investors are still only a good week's wagering away from profitability.

With three upsets, head-to-head tipping proved difficult this week, as evidenced by the all-tipster average of 4.6 from 8, the third-lowest round average for the season to date. The week's best performance, 6 from 8, was compiled by Bookie_9 just ahead of a slew of other tipsters that bagged 5. Worst round of the week went to Easily Impressed I, which registered just 2 from 8, consigning it to equal last on the MAFL head-to-head tipster ladder, some 11 tips behind the leader.

Atop that ladder, now leading by two clear tips, is Bookie_9 on 49.5 from 69 (72%). Combo_NN_1 is next on 47.5 (69%), and then come Combo_NN_2 and three of the H2H family all on 46.5 (67%).

Margin prediction was looking frightfully easy until the Eagles' blowout win over the Dogs on Sunday evening blew out most Margin Predictors' MAPEs for the round to the mid 30s. Bookie_9 returned the best result, an MAPE per game of 33.69 points. Next best was ProPred_3 on 34.48, fractionally ahead of Win_7 on 34.49. The round's worst MAPE performance was recorded by Combo_NN_2 with 39.70 points per game. At 35.40 points per game, the average MAPE across all Margin Predictors for Round 9 was the third-highest of the season.

Bookie_3 still has the best season-long MAPE, now at 27.29 points per game, up 1.2 points per game on its MAPE as at the end of Round 8. Combo_7 remains 2nd, though it's narrowed the gap to Bookie_3 to just 0.9 points per game. Bookie_9 still holds 3rd and is now only 0.24 points per game adrift of Combo_7, while two more Predictors, Combo_NN_2 and Combo_NN_1, also have sub-30 season-long MAPEs. (For quite some time during the current round all Margin Predictors has sub-30 MAPEs.)

Despite the generally large MAPEs returned by the Margin Predictors this round, line betting on the basis of their margin predictions would have been a profitable endeavour in most cases, as a majority of Predictors tipped 5 from 8 and the all-Predictor average was 4.4 from 8, the third-best result for the season. Bookie_3 and ProPred_7 continue to have line-betting records for the entire season that are significantly better than chance.

On probability scoring H2H, in both its unadjusted and adjusted forms, still heads the TAB Sportsbet bookmaker, an achievement that remains astonishing to me. (Consistent with this, Kelly-staking on the basis of H2H Adjusted's probability assessments - both for Home and for Away teams - would have produced an 11.7% ROI so far this season. The MAFL Head-to-Head Fund's ROI is currently just under 1%.) All four head-to-head probability tipsters returned positive probability scores across the eight games.

The Line Fund, however, returned a net negative probability score for the round, its 4th-worst per game result for the season. Not a round then, it turns out, to be wagering on all eight of its opinions.