Predicting Head-to-Head Market Prices

/
In earlier blogs I've claimed that there's not much additional information in bookie prices that's useful for predicting victory margins than what can be derived from a statistical analysis of recent results and an understanding of game venues.
Read More

The Relationship Between Head-to-Head Price and Points Start

/
I've found yet another MAFL-related use for the Eureqa tool, this time to determine the precise relationship between a team's head-to-head price and the start it's giving or receiving on line betting. A simple plot of the history of a team's head-to-head price (or the probability that can be inferred from it) versus its start on line betting makes it obvious that there's a relationship between the two and that it's a non-linear one, but in the past I've been constrained by my own (lack of) ingenuity and persistence in generating sufficient possibilities to find its exact nature.
Read More

What Do Bookies Know That We Don't?

/
Bookies, I think MAFL has comprehensively shown, know a lot about football, but just how much more do they know than what you or I might glean from a careful review of each team's recent results and some other fairly basic knowledge about the venues at which games are played?
Read More

In-Running Wagering: What's the Best Strategy?

/
With services such as Betfair now offering in-running wagering opportunities, the ability to accurately assess a team's chances of victory at any given point in a game is now of considerable commercial value. Imagine, for example, that your team, who are at home, lead by 18 points at the first change. Would a wager on them at $1.40 be advised?
Read More

Looking At Team Performance Quarter-By-Quarter

/
AFL Football - as the cliche goes - is a game of four quarters. The benefit of this arrangement is that AFL scores provide twice as much information about the ebb and flow of each contest as the scores of any other form of football in this country. With the quarter-by-quarter information alone we can perform some interesting analyses for every team.
Read More

Modelling AFL Team Scoring

/
Today's blog is the first in a series that will look at statistically modelling the scoring behaviour of teams in the AFL. If you're profoundly reductionist about it, you can think about a team's footy score as being the product of the number of scoring shots it creates and the proportion of those scoring shots that it converts into goals.
Read More

Goalkicking Accuracy Across The Seasons

/
Last weekend's goal-kicking was strikingly poor, as I commented in the previous blog, and this led me to wonder about the trends in kicking accuracy across football history. Just about every sport I can think of has seen significant improvements in the techniques of those playing and this has generally led to improved performance. If that applies to football then we could reasonably expect to see higher levels of accuracy across time.
Read More

Scoring Shots: Not Just Another Statistic

/
For a while now I've harboured a suspicion that teams that trail at a quarter's end but that have had more scoring shots than their opponent have a better chance of winning than teams that trail by a similar amount but that have had fewer scoring shots than their opponent. Suspicions that are amenable to trial by data have a Constitutional right to their day in court, so let me take you through the evidence.
Read More

Why April's Conceivably Better Than March

/
It's an unlikely scenario I know, but if the players on the AFL Seniors lists ever got to talking about shared birthdays I'd wager they'd find themselves perplexed. As chestnuts go, the Birthday problem is about as hoary as they come. It's about the probability that two randomly selected people share a birthday and its longevity is due to the amazement most people express on discovering that you need just 23 randomly selected people to make it more likely than not that two or more of them will share a birthday. I'll venture that few if any of the 634 players on the current Seniors lists know that but, even if any of them did, they'd probably still be startled by what I'll call the AFL Birthday phenomenon.
Read More

The Other AFL Draft

/
Drafting is a tactic well-known to cyclists and motor-racers and confers an advantage on the drafter by reducing the effort that he or she (or his or her vehicle) needs to expend in order to move. There's a similar concept in round-robin sports where it's called the carry-over effect, which relates to the effect on a team's performance in a particular game that's due to the team its current opponent played in the previous round. Often, for example, it's considered advantageous to play teams the week after they've faced a difficult opponent, the rationale being that they'll have been 'softened up', demoralised and generally made to feel blah by the previous match.
Read More

Using a Ladder to See the Future

/

The main role of the competition ladder is to provide a summary of the past. In this blog we'll be assessing what they can tell us about the future. Specifically, we'll be looking at what can be inferred about the make up of the finals by reviewing the competition ladder at different points of the season.

I'll be restricting my analysis to the seasons 1997-2009 (which sounds a bit like a special category for Einstein Factor, I know) as these seasons all had a final 8, twenty-two rounds and were contested by the same 16 teams - not that this last feature is particularly important.

Let's start by asking the question: for each season and on average how many of the teams in the top 8 at a given point in the season go on to play in the finals?

2010 - In Top 8.png

The first row of the table shows how many of the teams that were in the top 8 after the 1st round - that is, of the teams that won their first match of the season - went on to play in September. A chance result would be 4, and in 7 of the 13 seasons the actual number was higher than this. On average, just under 4.5 of the teams that were in the top 8 after 1 round went on to play in the finals.

This average number of teams from the current Top 8 making the final Top 8 grows steadily as we move through the rounds of the first half of the season, crossing 5 after Round 2, and 6 after Round 7. In other words, historically, three-quarters of the finalists have been determined after less than one-third of the season. The 7th team to play in the finals is generally not determined until Round 15, and even after 20 rounds there have still been changes in the finalists in 5 of the 13 seasons.

Last year is notable for the fact that the composition of the final 8 was revealed - not that we knew - at the end of Round 12 and this roster of teams changed only briefly, for Rounds 18 and 19, before solidifying for the rest of the season.

Next we ask a different question: if your team's in ladder position X after Y rounds where, on average, can you expect it to finish.

2010 - Ave Finish.png

Regression to the mean is on abundant display in this table with teams in higher ladder positions tending to fall and those in lower positions tending to rise. That aside, one of the interesting features about this table for me is the extent to which teams in 1st at any given point do so much better than teams in 2nd at the same point. After Round 4, for example, the difference is 2.6 ladder positions.

Another phenomenon that caught my eye was the tendency for teams in 8th position to climb the ladder while those in 9th tend to fall, contrary to the overall tendency for regression to the mean already noted.

One final feature that I'll point out is what I'll call the Discouragement Effect (but might, more cynically and possibly accurately, have called it the Priority Pick Effect), which seems to afflict teams that are in last place after Round 5. On average, these teams climb only 2 places during the remainder of the season.

Averages, of course, can be misleading, so rather than looking at the average finishing ladder position, let's look at the proportion of times that a team in ladder position X after Y rounds goes on to make the final 8.

2010 - Percent Finish in 8.png

One immediately striking result from this table is the fact that the team that led the competition after 1 round - which will be the team that won with the largest ratio of points for to points against - went on to make the finals in 12 of the 13 seasons.

You can use this table to determine when a team is a lock or is no chance to make the final 8. For example, no team has made the final 8 from last place at the end of Round 5. Also, two teams as lowly ranked as 12th after 13 rounds have gone on to play in the finals, and one team that was ranked 12th after 17 rounds still made the September cut.

If your team is in 1st or 2nd place after 10 rounds you have history on your side for them making the top 8 and if they're higher than 4th after 16 rounds you can sport a similarly warm inner glow.

Lastly, if your aspirations for your team are for a top 4 finish here's the same table but with the percentages in terms of making the Top 4 not the Top 8.

2010 - Percent Finish in 4.png

Perhaps the most interesting fact to extract from this table is how unstable the Top 4 is. For example, even as late as the end of Round 21 only 62% of the teams in 4th spot have finished in the Top 4. In 2 of the 13 seasons a Top 4 spot has been grabbed by a team in 6th or 7th at the end of the penultimate round.