Matter of Stats

View Original

Team-by-Team and Venue-by-Venue Conversion Rate History

Today's post continues the recent theme of entries here in the Statistical Analysis part of the site, taking yet another look at Scoring Shot Conversion rates but this time on a team-by-team and venue-by-venue basis.

Before I dive into the outputs (which, I'll warn you now, are nothing more than summary tables - there'll be no stunning visualisations in this blog), a few notes on the data:

  • The data used in this blog is all sourced from the big lists section of the afltables site, which provides a complete chronological history of every VFL/AFL game ever played.
  • It covers the period from Round 1 of 1897 to the end of the 1st Week of Finals in 2015.
  • Team data is aggregated on the basis of the team name as it was known at the time and as recorded in the data. So, for example, there's data for both North Melbourne and the Kangaroos, shown separately. Similarly, teams that have merged or otherwise changed name have a separate entry for each name (for example, Footscray and the Western Bulldogs).
  • Finals are defined as any game in the base data where the round number does not start with an "R". That means, for example, that the Sectional Finals from the early years are treated as non-Finals since they carry a round number in the data.
  • In Finals, the afltables site treats the winning team as the Home team (in that it records their score data first) and the losing team as the Away team (in that it records their score data second).

VENUE-BY-VENUE

The first table shows the venue-by-venue scoring statistics across the entire history of the VFL/AFL and is ordered based on the number of games played in the regular season at the venue.

Some highlights in the table are the following:

  • The MCG is, by far, the ground at which the most regular season and Finals have been played. In we include Finals, the MCG has been used at least twice as often as any other venue.
  • In the regular season, Home teams tend to convert at a slightly higher rate than Away teams at almost every venue. The exceptions amongst grounds on which at least 100 games have been played are Docklands, Lake Oval, Western Oval, Glenferrie Oval, Football Park and East Melbourne Oval.
  • There are only 15 grounds that have been used at least 30 times on which the Conversion rate is better than 50%.
  • The highest Conversion rate of all has been recorded at Docklands where it's been 55.4%. That's over 1% point higher than the next-best ground, York Park, where the rate has been 53.3%.
  • In Finals, across all venues, winning teams on average have a 6.3% point better Conversion rate. At the MCG, where over 70% of Finals have been played, the gap is +6.3%, and at Docklands, where just over 10% of Finals have been played, the gap is +5.7%.

I wondered if some of the variability in Conversion rates across venues might be explained by the dimensions of their playing surfaces, which is why I included the length (L) and width (W) data. The correlation of Conversion with either dimension is small, however, and the dimension data hard to source and probably unreliable anyway. (If anyone has better, more definitive data, I'd be happy to update this table with it and recalculate the correlations.)

One definite source of variability in the venue-by-venue Conversion rates is the portion of VFL/AFL history for which each ground was used. Average Conversion rates have changed quite substantially over the years, as shown in this post from 2010, wallowing in the 40% to 45% range for much of the first few decades of the sport, rising to about 50% by the time of the Second World War, tapering back to about 45% into the late 1960s, before rising gradually to more-common current levels of around 52% to 55%. So, any ground that was used, say, mostly before 1940 will tend to have depressed Conversion rates, for example.

Given that, a more useful comparison might be to look only at what I'll call the Modern Era of AFL, which I'll deem to have started in 1980, after which average Conversion rates have been consistently above 50% for each season.

The table for just this period appears below, and reveals that:

  • In the regular season Home teams still tend to have higher Conversion rates than Away teams, although this is not now the case for the MCG, Princes Park and Junction Oval in this shorter slice of history.
  • Princes Park is now the venue at which average Conversion rates have been highest, with teams kicking at over 55% at this venue, which is over 2% points higher than at Adelaide Oval where the average is next-highest.
  • Docklands now has only the 8th-highest Conversion rate, it being one of only two grounds used at least 100 times in the period where the Conversion rate was lower in the Modern Era than in the pre-Modern Era (the other ground is Subiaco where the decline is much smaller).
  • The average Conversion rate is above 50% at every venue used more than 100 times during the period.
  • Winning Finalists have enjoyed a 6% point higher Conversion rate than losing Finalists during this period.

TEAM-BY-TEAM

Next, and lastly, we'll look at the same data on a team-by-team basis, recording separately each team's scoring data when playing at Home versus playing Away during the regular season, and as winners versus losers in the Finals.

Again we'll initially review the entire span of VFL/AFL history where we find that:

  • The team with the all-time highest Conversion rate is the Western Bulldogs (54.9%), they over 0.5% points ahead of the Kangaroos (54.3%) when playing under those names.
  • Worst Conversion rate is University's 42.6%, which is over 3% points worse than South Melbourne's 46.0%.
  • Best Conversion rate at Home belongs to the Kangaroos (55.3%), theirs just slightly better than the Bulldogs' 55.1%. 
  • The worst Home Conversion rate is University's (42.1%), ahead of South Melbourne's (46.8%) and Fitzroy's (47.1%).
  • Best Conversion rate when playing Away is the Bulldogs' 54.8%, ahead of three other teams all on 54.3%, the Kangaroos, Port Adelaide and Sydney.
  • Only 9 of 24 teams have a lower Conversion rate when playing at Home compared to when playing Away, the largest deficit being the Brisbane Lions' (1.3%).
  • Amongst those teams with at least 30 games Finals experience, the best Conversion rate is just 51.9%, this belonging to Sydney. Only four other teams with at least the minimum level of Finals experience have a Conversion rate above 50% in Finals.
  • Of the 14 teams with at least 30 games of Finals experience only four have a higher Conversion rate in Finals than in the regular season (Melbourne 49.8% v 48.8%; North Melbourne 50.5% v 49.8%; Richmond 50.9% v 49.1%; St Kilda 49.4% v 48.6%).

This analysis too is coloured by the different portions of VFL/AFL history in which the various team names were active so I'll again consider only the Modern Era in an attempt to control, to some extent, for the variability of Conversion rates across time.

In the Modern Era, we find that:

  • The Western Bulldogs still have the best overall Conversion rate and the Kangaroos still lie 2nd, their entire histories being completely subsumed within the Modern Era. Essendon now, however, slip into 3rd where they'd been 14th in the previous all-time analysis.
  • South Melbourne have the worst Conversion rate during this period (49.4%), though the Gold Coast (50.3%) and GWS (51.2%) are only slightly superior (and have played twice as many games or more).
  • At Home, it's still the Roos with the best Conversion rate of 55.3%, though Essendon now joins the Western Bulldogs in 2nd with 55.1%. South Melbourne is worst (50%), ahead of the Gold Coast (50.6%) and GWS (50.9%).
  • Playing Away, the Dogs still have the highest Conversion rate (54.8%), but Hawthorn lies 2nd (54.3%) ahead of North Melbourne (53.8%).
  • Only 8 of the 23 teams have a lower Conversion rate at Home than Away, North Melbourne now having the largest deficit of 1.4% points, just ahead of the Lions' deficit of 1.3% points. The Kangaroos have the largest surplus, converting at a rate 2% points higher at Home than when playing Away. Clearly something happens to the Roos when they play as the Roos rather than as North Melbourne.
  • Amongst the seven teams with at least 30 games Finals experience, all have a Conversion rate better than 50% but only one, Essendon, has a better Conversion rate in Finals than in regular season contests.

I'll finish just by noting that the results here suggesting teams' Conversion rates in Finals are lower than their Conversion rates in the regular season are entirely consistent with the analysis in the previous post and in the one prior to that also.