2013 : Round 21 - Wagers & Tips (Updated)

Right now must be a challenging time to be a bookmaker. Just how do you price a team whose coach, assistant coach, head doctor and football manager are the subject of charges by the AFL Commission, none of which will be resolved for a couple of weeks? 

Well, if you're the TAB Sportsbet bookmaker, you don't.

As I type this, all markets for the Essendon v Roos clash on Saturday are suspended, but that matters little to Investors since they wouldn't have wagers on these games - based, at least, on my best guess as to what the prices would be - even if they could.

What they do have, instead, is two head-to-head bets, both on underdogs, one at $2.60 and another at $17, four line bets, three on favourites and one on an underdog, and 10 SuperMargin wagers, all on favourites.

(For the purposes of determining wagers and of coming up with tips and predictions, I've assumed that the Dons v Roos game is priced head-to-head at $2.20/$1.65, which is the last price I have for the TAB before the market's suspension, and that the handicap is +7.5 points. Should the market suspension be lifted before the game, I'll reassess our interest in a wager and recalculate tips and predictions based on the new prices.)

Once again this week, the Head-to-Head Fund's wager on a home team longshot has meant that one of its wagers represent the round's by-far most significant upside: an upset Melbourne win over Freo would add almost 10c to the Recommended Portfolio.

Other games with moderately significant upside are the Port Adelaide v Gold Coast and Sydney v St Kilda clashes, where favourable results would nudge the Recommended Portfolio up by almost 2c.

Downside lurks in the Power v Suns, Lions v Giants, Dees v Dockers and Swans v Saints games, where, at most, about 2c is at risk.

That leaves just two games where, as it stands, Investors have no interest: the Dons v Roos and the Dogs v Crows matchups. To be honest, I'm very happy to have nothing at stake on the Dons v Roos clash.

TIPSTERS AND PREDICTORS

In four of this week's contests there are at least five dissenting voices amongst the Head-to-Head Tipsters, loudest in the Port Adelaide v Gold Coast and Western Bulldogs v Adelaide games where there are six Tipsters raising their hands for the underdogs, but almost as loud in the Dons v Roos and Eagles v Cats games where one less than half a banker's dozen are united in their minority support. 

The levels of disagreement are so high, in fact, that the Disagreement Index stands at 30%, which means that if you chose a Tipster and game at random then there's a 30% chance that another Tipster would have chosen the other team in the game in question. 

Amongst the Margin Predictors there are far higher levels of agreement, with Predictors selecting different winners in only the Dons v Roos game where Combo_NN1 is alone in foreseeing a Dons victory. (By the way if the TAB posts a market for this game I will be recalculating the Margin predictions, as well as the Head-to-Head tips and probability forecasts, and rerunning the Line Fund algorithm and then reposting this chart.)

There's similarity even in the predicted margins, evidenced by the single-digit nature of most of the standard deviations of the margin predictions.

The Head-to-Head Probability Predictors carry a lone dissenting voice too, ProPred in the Eagles v Cats game where it finds difficulty in separating the teams; in every other contest, all Predictors are behind the favourites. And, like the Margin Predictors, their average level of dissent, here measured as the average mean absolute difference (MAD) in the probability estimates, is at a very low 4.6%.

Finally, the Line Fund algorithm rates three teams as better than 60% chances on line betting: the Roos (64%), the Lions (65%), and the Crows (65%).

I'll update this post later should the TAB venture an opinion about the Dons v Roos game.

UPDATE (THURSDAY PM)

The TAB Sportsbet Bookmaker finally relented sometime this afternoon or this evening, posting markets for the Dons v Roos game that were more generous in relation to Dons pricing than I was expecting, but not generous enough to entice any of the Funds to wager on the embattled probable finalists.

Right now I'm having some technical difficulties with the Excel add-in that generates the mini-charts for the Tipsters & Predictors section of the blog, so I'll not be posting a new image here until I can fix the problem, but the key changes are that:

• None of the Head-to-Predictors has changed its prediction
• Most of the Margin Predictors are now predicting slightly larger losses for the Dons (and Combo_NN1 is still resolutely behind the Dons by about the same margin as it was predicting previously)
• The directly Bookmaker-related Head-to-Head Probability Predictors have lopped about 12% points off the Dons' victory chances, while the remaining Head-to-Head Probability Predictors are largely unchanged
• The Line Fund algorithm now rates the Dons as 52% chances to win on line betting, which is almost, but not quite enough to result in a wager