Wednesday, July 21, 2010

MAFL 2010 : Round 17

First, some good news.

Back in Round 14, Investors in the Recommended Portfolio had a line bet on the Saints, who played Melbourne at Dockland. In the Bet Summary for the week I had the Saints giving the Dees 35.5 points start, which meant that their eventual 100-65 victory was a half-point short of covering the spread, a fact that I seem to recall lamenting at some length, certainly in the conversations that I had in the following week, and probably in the blog too (which I'm not of a mind to review at the moment).

Well every couple of weeks I reconcile the TAB's view of our finances with the view that I maintain via a spreadsheet and in performing this reconciliation I discovered that the Saints were actually giving 32.5 rather than 35.5 points start. So, our line bet was a winner and I can report that this skyrockets the Recommended Portfolio price up by 1c. It now stands at 91.3c.

Recommended Portfolio holders would do well to bear in mind that they're a little wealthier than they thought they were when they started reading this blog, as they review this week's assortment of wagers.

In total, there are 27 of them representing over one-quarter of Total Initial Funds. New Heritage owns seven of them, which together represent about 60% of that Fund and which have been wagered at a weighted average price of $1.29. This relatively low weighted average is due partly to the fact that the two largest New Heritage wagers are at prices of $1.10 or lower.

Including this week's 60% contribution, every dollar in the Heritage Fund has now been wagered 5.6 times. Unfortunately, so far it's only come back a little over 5 times.

The Heuristic-Based Fund has made 5 bets totalling 25% of the Fund with a weighted average price of $1.35. The average price is boosted by its largest wager, which is on the Eagles at $1.90 who are narrow underdogs in their game with the Blues.

ELO-Line also has 5 wagers representing 25% of the Fund. Four of these wagers are on favourites giving start, and one is on an underdog receiving start.

Prudence has made 4 selections totalling 21% of the Fund and, true to its name, has secured a weighted average price of $1.17, due largely to the fact that its 2 largest wagers are, like the New Heritage Fund, on teams priced at $1.10 or below.

Shadow is the next most active Fund with 3 wagers totalling 15% of the Fund. These are at a weighted average price of $1.18, only fractionally higher than Prudence's and this only because Shadow's largest wager is on the Roos at $1.40.

That leaves only the Hope Fund, which has ventured just 3 wagers for 11% of the Fund this week though these have been placed at a weighted average price of $2, the highest of all the Funds due largely to a wager on the Dees at $2.25.

These Hope bets are also the wagers for MIN#002, and those of the New Heritage Fund are the wagers of MIN#017.

Here's the detail:


Key matches this weekend for those with the Recommended Portfolio are the St Kilda v Hawthorn game (which offers a round-high 1.9% upside), the Geelong v Lions game (which carries a round-high 6.0% downside and a near round-high 6.8% swing), and the Collingwood v Richmond game (which carries a round-high 6.9% swing).

MIN#002 cares most about the Roos v Essendon game as it represents for him both the largest possible gain (+2.0%) and the largest swing between potential win and potential loss (-7.0%).

MIN#017 faces five games for which the potential swing is 14% or more. The St Kilda v Hawthorn game offers the greatest upside of just under 6%, the Cats v Lions game threatens the largest loss at just over 13%, and the swing between the best and worst possible outcomes is greatest for the Roos v Dons clash where the difference is almost 15% (though this is only a tad higher than the swing for the Dogs v Freo game).

So the week's Ready Reckoner, as Russian Roulette, is startlingly unbalanced with respect to upside and downside risk for all Investors:


What better way to emphasise for Recommended Portfolio holders the disparity between potential profit and possible loss for most of this week's games than to construct a visual representation (aka a chart) of the Ready Reckoner?



(Games where there are two bars of equal length are those in which Investors have either no head-to-head bet or no line bet. As such Investors' return is unchanged by the result on whichever market they have no stake in.)

MAFL Tipsters are unanimous or near to it for five of this week's games and almost 50-50 split on the remaining three contests.

Here are the details:

  • St Kilda are 8-7 favourites over Hawthorn. All five Margin Tippers predict a Saints victory, by margins ranging from 7.5 to 20 points, while Easily Impressed I and II, the leading MAFL tipsters, both tip a Hawks win.
  • Collingwood are unanimous favourites over the Tigers, with the Margin Tippers predicting Pies wins by between 14 and 58 points.
  • Geelong are unanimous favourites over the Lions, making this the second week in a row that the MAFL Tipsters have been united in discounting the Lions' chances. The Margin Tippers predict victory margins in the 22 to 52 point range.
  • The Kangaroos are 13-2 favourites over Essendon. Once again the Margin Tippers are unanimous in their support, though for this game their predicted victory margins span a narrow range of 15 to 23 points. Easily Impressed I and II are split on this game's outcome.
  • Carlton are 9-6 favourites over the Eagles. This is the first game for which the Margin Tippers disagree about the likely result, with three predicting a Carlton win and two an Eagles win. HAMP tips West Coast by 1 point, making this its Game of the Round, which it also is for BKB. Easily Impressed I and II are split on this game too.
  • The Dogs are 10-5 favourites over Fremantle, though all five Margin Tippers are on the Dogs with predicted margins ranging from 15 to 28 points. Easily Impressed I and II both predict wins by the Dogs.
  • Sydney are 13-2 favourites over Melbourne. Four of the five Margin Tippers predict Sydney victories, including ELO, which tips a 13 point Swans win, and Chi, who tips a 4 point Swans victory, making this both tipsters' Game of the Round. LAMP predicts a 1 point Melbourne win and also has this as its Game of the Round. Easily Impressed I and II both predict Sydney wins.
  • Adelaide are 14-1 favourites over Port, with the Margin Tippers again unanimous in their support. Predicted margins range from 10 to 24 points. Easily Impressed I and II both predict Adelaide victories.
Here's a chart that might help you decide whether or not to take heed of any of that. It shows the cumulative tipping performance of the MAFL Tipsters as at the end of Rounds 11 and 16.



You'll recall that I'm field-testing two new algorithms for next season: a Bookie Probability Model (BPM) that will bet head-to-head on home teams only, and a Super Smart Model (SSM) that will bet on the line market.

BPM this week recommends only one bet: 1.2% on Adelaide, which it sees as value at $3.75. (Actually it also recommends a second bet on St Kilda, but the recommended size is 0.02%, which means that you'd need a bank of $5,000 in order for this recommendation to equate to the minimum allowed $1 wager. I think it's fair to deem a 0.02% recommended bet as somewhat below the level of 'who cares'.)

Here's what SSM makes of the round:



SSM agrees with only one of ELO's five bets, that on the Dogs, and disagrees with those on the Saints, Pies, Cats and Port. SSM would also have wagered on the Roos, the Dees and the Eagles, all of which selections would meet ELO's home-team-only wagering constraint. Well with that level of disagreement we certainly won't be running both ELO-Line and SSM next season.

HELP has come up with a more sensible set of probability assessments this week and has also gone with a balanced 4 home and 4 away team split:


2 comments:

Mitch said...

"visual representation (aka a chart)..."

Hahaha have you been reading my QH proposals? I've been using the euphemism 'data visualisation', which is even better than yours :)

JN asked me what on earth it was and I said 'aka charting'!

TC said...

Strictly speaking, I reckon a table qualifies as an instance of 'data visualisation' ... actually, come to that, a lone number probably does too.