MAFL 2012 : Team Dashboard for Round 13

/

Right now, the teams that rank 1st to 4th in terms of points conceded per game also fill the first four places on the competition ladder, while those that rank 1st to 4th in terms of points scored per game currently lie 6th, 8th, 11th and 4th respectively. It seems then as though success in the competition is more related to defence than it is to offence.

Sydney, Essendon and West Coast's defensive superiority has come from denying opposing teams' opportunities to score: they have the three best records in terms of scoring shots conceded per game of all teams in the competition.

Collingwood, however, has not denied teams scoring shots - it has only the 12th best performance in the league on this measure - but has presented them with opportunities that have been relatively hard to convert. Collingwood's opponents have converted only 46.7% of their scoring opportunities, a rate fully 2 percentage points lower than for any other team in the competition.

Fremantle and St Kilda are two other teams that - like Sydney, Essendon and West Coast - have been relatively good at denying their opponents scoring opportunities. That's why they occupy positions 4 and 6 in terms of scoring shots conceded per game. What's hurt them though is that their opponents have been able or allowed to convert those opportunities at relatively high rates: about 55% of the scoring opportunities they've presented to opponents have been converted into goals. 

MAFL 2012 : Team Dashboard for Round 12

/

The latest Team Dashboard follows.

This week I'll draw your attention to the columns at the right of the Scoring Shot Data section where I use the Win Production Function that I created last year, which allows me to estimate what a team's winning percentage in a season should be based solely on its offensive and defensive goal and behind scoring statistics during that season.

If we compare the ranking that we obtain from each team's expected winning percentage with current ladder positions we find a difference of two places or more for ten teams:

  • Adelaide and Fremantle are 4 places higher on the competition ladder than their scoring statistics suggest
  • Collingwood, Geelong and West Coast are 2 places higher
  • Hawthorn is 5 places lower (they've won only two-thirds of their games but their scoring statistics suggest they should have an 80% winning rate)
  • Carlton, Gold Coast, St Kilda and Sydney are 2 places lower

MAFL 2012 : Team Dashboard for Round 10

/

The latest Team Dashboard follows.

Looking only at teams' Quarter-by-Quarter performances, one way of classifying them is as follows:

Teams with one relative standout quarter

  • Adelaide (weakest in 3rd terms)
  • Lions (strongest in 4th terms)
  • Gold Coast (strongest in 2nd terms)
  • Richmond (weakest in 2nd terms)
  • St Kilda (strongest in 3rd terms)
  • Sydney (weakest in 4th terms)
  • Western Bulldogs (strongest in 2nd terms)

Teams with two relative standout quarters

  • Carlton (strongest in 4th terms, weakest in 2nd)
  • Essendon (strongest in 2nd terms, weakest in 4th)
  • Geelong (strongest in 2nd terms, weakest in 3rd)
  • Hawthorn (strongest in 2nd terms, weakest in 3rd)
  • Kangaroos (strongest in 2nd terms, weakest in 4th)

Teams with three relative standout quarters

  • Collingwood (strongest in 1st and 2nd terms, weakest in 4th)
  • Fremantle (strongest in 3rd terms, weakest in 1st and 2nd)
  • Port Adelaide (strongest in 3rd terms, weakest in 2nd and 4th)

Teams weak in every quarter

  • GWS and Melbourne

Teams strong in every quarter

  • West Coast

 

MAFL 2012 : Team Dashboard for Round 8

/

The latest Team Dashboard follows in which you can find out that: 

  • Adelaide has the 2nd best record of all teams in Quarters 1 and 4, but is 13th in Quarter 3s.
  • Brisbane, in aggregate, has outscored its opponents in only Quarter 3s.
  • Carlton ranks 4th and 5th in Quarters 1 and 4, but ranks only 11th in Quarters 2 and 3.
  • Collingwood's quarter-by-quarter record declines as the game progresses. It's 3rd based on Quarter 1 performances, 5th based on Quarter 2, 9th based on Quarter 3, and 15th on Quarter 4.
  • Essendon is ranked 2nd on Quarter 2 performances.
  • Fremantle is ranked last on Quarter 1 performances and just 14th on Quarter 2 performances, but improves in the 2nd half to be ranked 2nd in Quarter 3s and 3rd in Quarter 4s.
  • Geelong is ranked 6th on Quarter 4 performances.
  • Gold Coast is ranked 7th on Quarter 2 performances.
  • GWS, in aggregate, has been outscored by almost 2:1 in every quarter, almost 3:1 in 1st terms.
  • Hawthorn is ranked 14th on Quarter 3 performances and 1st on Quarter 2s.
  • Kangaroos, in aggregate, have outscored their opponents in every quarter.
  • Melbourne have yet to win a final term, having drawn only one of eight.
  • Port Adelaide are ranked 3rd on Quarter 3 performances.
  • Richmond are ranked 10th on Quarter 1 performances and 17th on Quarter 2s, but 5th on Quarter 3s and 4th on Quarter 4s.
  • Sydney are ranked 1st on Quarter 3 performances but only 13th on final terms.
  • West Coast are ranked only 7th on Quarter 3 performances.
  • Western Bulldogs, in aggregate, have outscored their opponents in only Quarter 2s. 

MAFL 2012 : Team Dashboard for Round 6

/

The latest Team Dashboard is below.

I know we have some new visitors to MAFL this year, so I thought I should provide an explanation of the rightmost columns of the Scoring Shot Data section. These columns are based on some analysis I undertook to create a "win production function" for AFL, which is an equation that relates a team's scoring shot data to the number of games we'd expect they'd have won given that data. You can think of it as a way of measuring how well or how poorly a team has been "rewarded" for its scoring performance.

As it stands now, 3 teams (the Roos, Tigers and Saints) have won at least one game fewer than we'd expect given their current scoring statistics, and 1 team (Collingwood) has won one game more than we'd expect. Three other teams - Adelaide, Fremantle and West Coast - have also won more games than their statistics would suggest, though the benefit is not yet equal to a full game.

Similarly, the Gold Coast, Hawthorn, Melbourne and Port Adelaide have won fewer games than their statistics would entitle them to, though the difference is also not yet equal to a full game.

MAFL 2012 : Team Dashboard for Round 5

/

The latest Team Dashboard follows. Here's a factoid for each team:

  • Adelaide : 3rd highest scoring shots created per game and 5th lowest scoring shots allowed, but last in terms of both own and opponents' conversion of those scoring shots
  • Brisbane Lions : Sub 100 percentage for every quarter except the 3rd
  • Carlton : lowest Opponent conversion rate in the league
  • Collingwood : still yet to win a final term
  • Essendon : percentage over 100 in every quarter
  • Fremantle : 14th in 1st and 2nd quarters, 4th in 3rd and 4th quarters
  • Geelong : 3rd best conversion rate in the league
  • Gold Coast : yet to win a 1st quarter
  • GWS : yet to win a 3rd quarter
  • Hawthorn : yet to lose a 2nd quarter
  • Kangaroos : 1st in scoring shot production across the league
  • Melbourne : 2nd lowest scoring shot production and 3rd highest scoring shots conceded
  • Port Adelaide : 3rd best in 3rd quarters
  • Richmond : yet to win a 2nd term
  • St Kilda : lowest scoring shots conceded per game
  • Sydney : highest scoring shot conversion rate and greatest own vs opponent scoring shot conversion rate differential
  • West Coast : best 1st and 3rd quarter records in the league (and 2nd best 2nd quarter performance)
  • Western Bulldogs : 3rd worst 1st quarter performance in the league

 

MAFL 2012 : Team Dashboard for Round 4

/

It's still a little early to be discerning and commenting on any significant team differences, so I'll again content myself with a few parenthetic observations on this week's Team Dashboard: 

  • West Coast are the first team to have scored more than 500 points for the season
  • Melbourne and GWS are the first teams to have conceded 500 points
  • St Kilda's scoring statistics so far this season are, according to the Win Production Function, good enough to have earned them 1.3 wins more than they've achieved
  • The Roos' scoring statistics are good enough to have earned them an extra 1.1 wins
  • Collingwood have not won a final term so far this season
  • Geelong have not won a 3rd term 

MAFL 2012 : Team Dashboard for Round 3

/

The Suns won their first two quarters for the season this week, though this wasn't quite enough to secure them their first competition points at the same time. That moves them equal now with GWS and the Dogs in having won just 2 of the 12 quarters in which they've taken part.

Conversely, the Eagles have won all of their 1st, 2nd and 3rd quarters, and lost only a solitary 4th term - as it happens, to GWS last weekend, when the game was already long since decided.

Carlton, Hawthorn and Freo, at this early stage, are emerging as the competition's slow starters and big finishers, while the Swans are exhibiting the opposite behaviour. Only the Roos, the Saints and the Eagles have demonstrated an ability to perform strongly in all four quarters.

MAFL 2011 : Team Dashboard for Round 24

/

The final MAFL Team Dashboard for season 2011 follows.

Here are a few final observations: 

  • The Roos, despite not making the Finals, have nonetheless finished the home-and-away season with a percentage above 100.
  • St Kilda and Sydney kicked only 275 and 267 goals respectively across their 22 matches, fewer than, for example, Melbourne, which finished 13th
  • Essendon conceded 326 in its 22 games, more goals than the 15th-placed Lions conceded
  • Port Adelaide kicked an amazingly low 188 behinds all season, 36 fewer than any other team
  • Adelaide, though finishing 14th on the ladder, were 5th best on Opponent Conversion Rate. They were also 9th in 1st terms and 8th in final terms
  • Brisbane, 15th on the ladder, finished 7th on Own Conversion Rate. They also won 2.5 games fewer than their scoring statistics suggest they should have. As well, they finished 11th on 1st terms and 10th on 3rd terms
  • Carlton ran 5th on the ladder but 14th on Own Conversion Rate. So good were their scoring statistics generally, they won 2 games fewer than the Win Production Function predicts
  • Collingwood finished as minor premiers but were 5th on Own Conversion Rate
  • Essendon came 8th but ranked 14th on Opponent Conversion Rate and 12th on 1st terms
  • Fremantle were 11th on the ladder but 15th on Own Conversion Rate. They won over 2 games more than their scoring statistics suggest they should have. They also finished 6th on 3rd term performances
  • Geelong finished 2nd but were 11th on Opponent Conversion Rate and 8th on 1st terms
  • Melbourne ranked 13th on the ladder but last on final terms
  • Port Adelaide finished second-last but were, astonishingly, second on Own Conversion Rate
  • Richmond came 12th but were 8th on Own Scoring Shots per Game and 16th on 1st terms
  • St Kilda wound up 6th but were 13th on Own Scoring Shots per Game and 2nd on Opponent Conversion Rate. They were also 1st on 1st terms, 10th on 2nd terms, and 12th on 3rd terms
  • Sydney were 7th on the ladder but 16th on Own Conversion Rate and 1st on Opponent Conversion Rate
  • West Coast finished 4th despite the twin handicaps of finishing 13th on Own Conversion Rate and 16th on Opponent Conversion Rate. They also finished 14th on 3rd terms
  • The Western Bulldogs came 10th but were 1st on Own Conversion Rate and 5th on 3rd term performances

 Finally, I'd note that: 

  • Collingwood won 82% of 2nd terms and more than doubled its opponents' scores in 2nd and final terms
  • The Gold Coast won less than 20% of its 1st, 2nd an 3rd terms. Only mercy, perhaps, allowed it to win 27% of final terms
  • Port Adelaide won only 18% of 2nd terms
  • Richmond won only 20% of 1st terms
  • St Kilda won 77% of 1st terms and recorded a percentage of 150, then won only 41% of 2nd and of 3rd terms, and 55% of final terms, recording a percentage of just 85 in 2nd terms in doing so
  • Sydney recorded a percentage of just 84 in 3rd terms
  • West Coast recorded a percentage of just 87 in 3rd terms

 

MAFL 2011 : Team Dashboard for Round 23

/

Ranking the teams on the basis of what the Win Production Function suggests their winning rate should be given their scoring shot performance across the season, only one change would be made to the composition of the current top 8 on the ladder:  the Dons would drop to ninth and the Roos would grab 8th.

Further, within the top 8 there'd be only one change of order with the Blues taking 4th at the expense of West Coast; the five remaining teams would be ordered exactly as per their position on the competition ladder.

Further down the ladder the shuffling would be only a little more vigorous. Apart from Melbourne, who'd drop from 12th to 15th on the strength of their scoring shot performance, and the Dogs, who'd climb from 13th to 10th, no team would change position by more than one spot.

So, on the basis of just two pieces of data about each team, scoring shot differential and conversion differential, we could replicate the current competition ladder ordering to within one spot for every team bar two of them, neither of them in finals contention. Seems the Win Production Function does a good job of distilling the key aspects of team performance for season 2011, as it has done for many seasons in the past.

Anyway, here's the latest Team Dashboard.

MAFL 2011 : Team Dashboard for Round 22

/
In a post for the Statistical Analysis journal in 2009 I investigated whether it was better for a team to lead or to trail narrowly at the end of a quarter. This analysis was motivated by a journal article investigating the phenomenon for NCAA basketball, which found that teams in this competition that trailed narrowly at half-time went on to win at a rate statistically significantly greater than 50%. I found only scant evidence for this in VFL/AFL football.
Read More

MAFL 2011 : Team Dashboard for Round 20

/

From this week's MAFL Team Dashboard I note that there are two teams sitting below Fremantle that have a superior percentage - the Roos and the Dogs. Further evidence for the fortuity of the Dockers' position resides in the fact that, based on their relative scoring shot production and relative conversion rate for those scoring shots, Fremantle have won almost 1.5 games more than they "should" have.

Other teams whose scoring performance is most misaligned, according to the Win Production Function, with their winning rate are: 

  • Melbourne, who've won just under 2 games more than their scoring statistics would predict
  • The Suns (also about 2 games more)
  • Adelaide and West Coast (about a game more)
  • Brisbane (almost 2 games less)
  • Carlton and the Roos (each over a game less)
  • Sydney and the Dogs (each about a game less) 

 

MAFL 2011 : Team Dashboard for Round 19

/

The latest (clickable as always) MAFL Team Dashboard follows.

This week I calculated the rank correlations between teams' ladder positions and their ranking on some of the metrics in the Scoring Shot and Performance by Quarter sections. Here are the results: 

  • Rank correlation between ladder position and Scoring Shots For: +0.86
  • Rank correlation between ladder position and Scoring Shots Against: +0.97
  • Rank correlation between ladder position and Scoring Shots Difference: +0.98
  • Rank correlation between ladder position and Own Conversion Rate: -0.09
  • Rank correlation between ladder position and Opponents' Conversion Rate: +0.20
  • Rank correlation between ladder position and Conversion Rate Difference: +0.12
  • Rank correlation between ladder position and ranking on Q1 performances: +0.74
  • Rank correlation between ladder position and ranking on Q2 performances: +0.89
  • Rank correlation between ladder position and ranking on Q3 performances: +0.57
  • Rank correlation between ladder position and ranking on Q4 performances: +0.85

What I make from all of this is that, so far this year, the better-performed teams have differentiated themselves by forcing their opponents to have more difficult scoring shots (and hence poorer conversion rates) and by being especially effective in 2nd and 4th quarters. It's interesting to note that seven of the teams currently sitting in finals positions on the ladder are amongst the top 8 teams ranked on 4th quarter performances.