The Allianz Leagues are under way, providing GAA fans with competitive games in their own county grounds. Home advantage is a rarity in the Championship but how much of a difference does it make in league play? Emmet Ryan and Ciaran Ruane decided to find out.
This article is in direct response to a piece written by Colm Keys for the Irish Independent on 4 February 2011 regarding home advantage in the GAA’s Allianz Football League.
We take issue with the method Mr Keys used to gauge home advantage and the figures he used for analysis. Essentially, in his efforts to work out how much individual teams benefitted from playing at home his focus was heavily on one side of the debate: What a team does in its own park. In order to truly assess the impact of home performance, how a team performs away from its own ground must also be taken into account.
Mr Keys did succeed in presenting the overall figures for home and away wins and losses but not on how it affected on a county-by-county breakdown. Some individual examples of teams travelling better were cited in the piece but the box in the original print version solely addressed home performances.
Our approach
We decided to go further than the three-year period of 2008-10 analysed by Colm Keys, expanding to a five-year period of 2006-10 for a better overall sample. We have however also broken down the data for the 2008-10 period specifically as well so we can analyse that in relation to the original article.
Furthermore we decided to look at how home advantage affected scoring over the five-year period in question. High scoring sports such as American football and basketball both account for the difference playing at home makes towards the difference to a total outcome. In both of those cases, bookmakers allow for a 3.0 point advantage in favour of the home team (despite the different scoring systems). We found the Allianz Football League to be rather different to both of these examples.
All of our data can be found online in this Google spreadsheet. It is ordered in terms of teams’ win percentage for home games versus away games, taking draws as being worth half a win as they are worth one point compared to two for a win. This also mirrors the format used by Simon Kuper and Stefan Syzmanski in their analysis of national team performance in the book Soccernomics.
Analysing the data
In the five-year period from 2006-10 there were 583 games played in the Allianz Football League, excluding finals played in neutral venues. Of these 326 (56%) were won by the home team compared to 212 by away teams (36%) and there were 45 draws (8 %). This aligns well with the original analysis from the Irish Independent. The next question we sought to answer however was how this affected the individual counties:
Of the 33 teams playing in the league during that period, all 33 had a better record at home than away. Some of these were not significantly better however and we will get to that.
The difference in terms of individual counties was staggering. Offaly proved the team with the broadest difference. The Faithful had a win percentage of 79% at home and 24% away. This produced a difference of 55 percentage points between home and away form.
Meath also proved far stronger at home than away, with a 79 win percentage in home matches compared to 26 percent in away games, for a difference of 53 percentage points. Meath and Offaly switch places in the 2008-10 period covered by the Irish Independent. Meath topped the table with a 67 percentage-point difference in win percentage while Offaly was second with a 65 percentage-point difference in win percentage.
Fermanagh and Monaghan were the counties least affected by where their games were played. Over 5 years, both teams’ win-percentage was less than half a percentage point better at home than away. This difference is so small as to render it statistically insignificant given the small sample size for individual teams.
In the case of Monaghan, further research into how Ulster teams in particular fared at Clones would be warranted as the venue is used extensively through the Ulster Championship. Mayo replace Roscommon when we look solely at 2008-10. Indeed in this period both Mayo and Fermanagh performed worse at home than away, the only counties to do so. Mayo suffered an 11 percentage-point decrease in win percentage at home, the worst record amongst all counties, whereas Fermanagh’s decline was six percentage points.
The advantage is small on the scoreboard
While every county proved superior to some degree at home in terms of wins and losses, the difference on the scoreboard proved much smaller than we expected. On average, home teams scored 14.5 points during the five-year period, compared to 13.2 points per game by away teams (converting goals into their value of 3 points). This meant that based on these figures, a game between two identical teams where one had home advantage would yield only a 1.3-point advantage on average on the scoreboard for the home team.
Again there were substantial differences in the average margins. Leitrim had the broadest difference in performance, typically winning by 4.6 points at home while losing by 2.4 points away, making them 7 points better off in relation to their opponents when playing at home rather than away. Limerick and perennial whipping boys London were joint-second with an average improvement of 5.8 points relative to their opponents when at home rather than away.
Despite being 19th in terms of the difference in winning percentage, Dublin were surprisingly showed the fifth biggest differential. During 2006-10, Dublin typically won home games by an average of 4.5 points but lost away games by an average of 0.7 points, leaving a margin of 5.2 points.
Only five of the 33 participating counties saw their margin improve when they played away from home. The biggest margin of these five was Fermanagh who lost by 3.3 points on average at home and lost by 2.4 points on average away, making them 0.9 points worse off in relation to opponents when playing at home.
Why we care
Reporting of statistics is something we take rather seriously. Data can tell us a lot but repeatedly there are examples in the media, whether intentional or accidental, of bad science (thanks Ben Goldacre) being reported as fact with mathematics used as a cover. Readers deserve better. We take no umbrage with Colm Keys, indeed we wouldn’t have thought of doing our own more detailed analysis without his initial report and we are grateful to him for that. Instead our aim is to promote higher standards.
We welcome any questions and criticisms of our research because if there are improvements to be made, we want to make them. As indicated with our comments on Monaghan’s record in Clones, there is definitely room for further research and we hope our efforts can encourage others to got out drill down through the data.
Wordpress only allows one author so just for clarity’s sake, Ciaran Ruane also contributed to this article.
Update: Like we said, accuracy matters so we have made some updates to round the figures where the sample size is too small to be statistically significant with the one exception where we were highlighting a figure as being insignficant.
Emmet Ryan and Ciaran Ruane write about sport at Action81