Matter of Stats

View Original

The MoS Tipsters, Predictors and Funds for 2018

Last year was a very good one here on MoS, with one or other of MoSSBODS' or MoSHBODS' algorithms out-tipping the bookmakers on head-to-head, margin and probability forecasting.

Given that, it would seem silly to change much at all - so I won't be.

On the Tipster and Predictor front, ENS_Linear has again been recalibrated this year using only the data for seasons 2008 to 2017 and the same three base learners:

Each learner regresses home team margin on home team bookmaker probability (calculated assuming equal overround on both teams), home team handicap in the line market, home and away team prices, MARS ratings, team venue experience values, team form (measured by recent MARS rating changes), and the game's interstate status.

The MoSHBODS Team Rating System will again provide a Head-to-Head Tipster, as well as a Margin and a Probability Predictor. Its margin prediction will continue to be the difference between its home and away team score predictions plus the 2-point bias adjustment, and its head-to-head tip will be the home team if this adjusted margin is positive, and the away team otherwise.

Its probability prediction will be derived from its score predictions for each team, both converted to expected scoring shot predictions and then used in a simulation drawing on a version of this team scoring model from 2014. For that simulation we use:

For the Bivariate Negative-Binomial Scoring Shot Distribution

  • Mean Scoring Shots = Expected Values according to MoSHBODS
  • Size Parameters = 163 for the Home team and 96 for the Away team
  • Correlation between Home and Away Scoring Shots = -0.24

For the two univariate Beta-Binomial Scoring Shot Conversion Distributions

  • Number of Scoring Shots = result drawn using the Bivariate Negative-Binomial
  • Mean Conversion Rate = 53.7% for the Home team and 53.2% for the Away team
  • Theta Parameters = 262 for the Home team and 109 for the Away team

The MoSSBODS Team Rating System will be used analogously to provide Head-to-Head, Margin and Probability forecasts. (Note that neither MoSHBODS nor MoSSBODS have been reoptimised for 2018, except to the extent that the Venue Performance Values will have altered on the basis of all of the games from 2017.)

At the end of last season I was fairly convinced that I'd over-optimised ChiPS in the lead up to the season, but a closer review of the results suggests that this probably wasn't the case. I base this on the fact that the results were not poorest for those team and venue combinations where I'd applied no regularisation to the HGA values, allowing them to drift as high as +27 for Geelong at Docklands, and as low as -24 for Collingwood at Docklands.

So, I've used the same unconstrained approach in reoptimising ChiPS for the 2018 season, this year based on the results from 2008 to 2017. ChiPS still has a number of sizeable HGA values, but I'm not as concerned about that as I was a few months back. Hindsight will, as ever, reveal the wisdom or otherwise of that opinion.

In total, there will again be nine Head-to-Head Tipsters, 10 Margin Predictors, and six Head-to-Head Probability Predictors, as listed in the table below.

(You can find out a little more about some of the forecasters in this blog from 2016.

As well, if you're new here - and welcome if you are - this post from 2015 will give you some idea about how it all works here on MoS, although many of the statistical models discussed there are no longer in use.) 

That just leaves one question to answer about 2018: how will the MoS Funds operate?

The short answer is that, once again, there'll be a Head-to-Head, a Line and an Over/Under Fund, but the weights will be different in 2018. The details follow.

HEAD-TO-HEAD FUND (WEIGHTING 35%)

Last year, head-to-head wagering was informed for the first time by MoSHBODS which meant that, in the fine MoS (small 'c') conservative tradition, only a 20% weight was attached to the Head-to-Head Fund. That Fund was highly profitable overall, with only four rounds producing a loss after Round 5, at which point the Fund was down by 1.2%. The next six rounds - from Round 6 to Round 11 - added over 42c to the Fund price, and the last 16 rounds another 21c, to leave the Head-to-Head Fund finally up by about 62% on the season with a 21% ROI on a 2.9 turn. Of the 114 bets, 57 were collects.

Those results would seem to justify some upweighting of the Fund and it will, accordingly, carry a 35% weight for season 2018.

The Fund will otherwise operate exactly as it did last year, including being free to wager on home and on away teams, and employing a fractional Kelly betting staking strategy where each bet is a proportion of initial funds equal to:

1/7 * (Price x Estimated Probability - 1)/(Price-1)

provided that Price x Estimated Probability > 105%. In other words, we'll again require a minimum 5% overlay before we'll wager.

If last season is a good, rough guide, I expect that the Fund will make about 105-125 bets across the season (55% to 70% of them on home teams), with an average size of about 2.5%, which gives a turn of about 2.9. A reasonable target for this Fund would be a 5% ROI, which would provide about a 15% return on funds.

As always, previous performance does not guarantee future returns (and we all know about the difficult second season after an above-average rookie year).

LINE FUND (WEIGHTING 35%)

Last year, line wagering was again informed by MoSSBODS, and saw profits in all but seven rounds. The season finished with the Line Fund producing about a 41% profit from a 24% ROI on a 1.7 turn. Of the 97 bets placed, 58 (60%) were winners.

The Line Fund will rely on MoSSBODS again, and it too will be allowed to wager on home and on away teams. As per last year, it will place a line bet on a team proportional to the size of its measured 'advantage', only when the handicap set by the TAB bookmaker is assessed by MoSSBODS as being in error by more than 8 points. 

Bet size, as a percentage of initial Funds, will again be determined as:

Bet Size = (Assessed Margin + Handicap Offered - 8)/300
provided Assessed Margin + Handicap Offered > 8

So, for example, if the Home team is giving 12.5 start and MoSSBODS assesses them as winning by 33 points, the bet would be (33 - 8 - 12.5)/300 = 4.2% of initial Funds.

Instead, if the Away team is receiving 34.5 start and MoSSBODS assesses them as losing by only 12 points, the bet would be (34.5 - 8 - 12)/300 = 4.8% of initial Funds.

The Line Fund will carry the same weight as the Head-to-Head Fund this year, 35%, which is 5% points less than it did last year.

Using the last few seasons as a rough guide, I would expect the Fund to wager on about 50% of the games (a little over half the time on home teams) with an average bet size of about 2% of the Fund, on which basis the expected turn would be about 2. An ROI of 10% would seem to be a reasonable target and would provide about a 20% return on funds.

Caveats apply.

OVERS/UNDERS FUND (30%)

The Over/Under Fund was the least impressive of the MoS Funds last year, making losses in 13 rounds, and profits in only 12. Quantitatively, the wins did outweigh the losses however, and the Fund recorded an 8% profit from a 4.4% ROI on a 1.9 turn. It made 95 wagers, collecting for 53 (56%) of them.

In 2018, the Fund will continue to use MoSSBODS for its wagering advice and will again require a 6 point minimum overlay in both directions before wagering.

Because of its 2017 performance, the Fund will have a slightly lesser weighting of 30% in 2018.

We should expect this Fund to wager in about 50% to 60% of games, or around 100 to 120 games across the season. Roughly 50% to 60% of those wagers are likely to be unders bets. Each wager will be 2% of initial Funds, so we can expect a turn of around 2 to 2.4. History suggests a target ROI of about 4% is reasonable, which would yield about an 8% to 9% return on funds.

All three Funds will again wager in whichever of the TAB or Centrebet markets are offering the most attractive prices (and are accepting my bets).

 

On the team rating front, I'll definitely be writing about MoSSBODS and MoSHBODS each week, though I've yet to come to a landing on whether or not to include ChiPS and MARS ratings in those write-ups. Sometimes they feel a little superfluous and I wonder if my attachment to them is merely an artefact of history.

And, finally, I plan also to provide a weekly Team Dashboard and to blog occasionally about any interesting statistical analyses that I get the time to conduct.

So, pretty much like last year. Hope to see you around.