By John Considine
Two days ago, the Irish Times raised the issue of the GAA distributing statistics on their games. The piece by Andy McGeady highlighted the range of people currently engaged in collecting such statistics and the possibility for some action by the governing body. It could be inferred that the market for GAA statistics needed greater regulation. While I would have some sympathy with this view, there are many issues that would need to be addressed.
McGeady pointed out that there is a commercial market for such material, e.g. GaelicStats.com. There are also other sites from which interested parties can gather data, e.g. dontfoul.wordpress.com. In addition, many media outlets provide some basic statistics on the games they cover. Therefore, one might wonder why there is the need for the GAA to collect (or purchase) and distribute such statistics. Who would benefit from such a change? Not Rob Carroll of GaelicStats.com (unless the GAA purchased his services to collect the statistics). The media might get the statistics cheaper than having to purchase them. The GAA and the wider public might benefit from a greater evidenced based discussion of the games. However, the case for greater GAA intervention is far from settled.
For the moment, let us go with the McGeady suggestion. It is possible to read the McGeady piece as suggesting the GAA coordinating the collection and distribution of statistics from county teams. Alternatively, the GAA could outsource the collection of data to a professional firm like Opta or GaelicStats. Or, it could use its army of nationwide volunteers to collect the data. If the GAA desired then it could establish a GAAStats wiki. In this case they could have those who use the stats collection apps on their phones to submit their data to a central system. The central system would facilitate some cross checking and verification of the data. It may not have the internal consistency of an Opta system but it could benefit from the wisdom of crowds. Such a system is a possibility. However, it would require some planning and regulation.
The biggest issue would involve deciding on the exact definition of the data to be collected. Let me illustrate some of the issues using data on the 2015 Munster Senior Hurling Final between Tipperary and Waterford. I'm going to use the stats provided by the broadcaster of the live event (RTE) and two newspapers (Irish Examiner and Irish Independent). Below are the scoring statistics. The numbers in blue font are "correct" but many of those in black font are open to interpretation.
McGeady pointed out that there is a commercial market for such material, e.g. GaelicStats.com. There are also other sites from which interested parties can gather data, e.g. dontfoul.wordpress.com. In addition, many media outlets provide some basic statistics on the games they cover. Therefore, one might wonder why there is the need for the GAA to collect (or purchase) and distribute such statistics. Who would benefit from such a change? Not Rob Carroll of GaelicStats.com (unless the GAA purchased his services to collect the statistics). The media might get the statistics cheaper than having to purchase them. The GAA and the wider public might benefit from a greater evidenced based discussion of the games. However, the case for greater GAA intervention is far from settled.
For the moment, let us go with the McGeady suggestion. It is possible to read the McGeady piece as suggesting the GAA coordinating the collection and distribution of statistics from county teams. Alternatively, the GAA could outsource the collection of data to a professional firm like Opta or GaelicStats. Or, it could use its army of nationwide volunteers to collect the data. If the GAA desired then it could establish a GAAStats wiki. In this case they could have those who use the stats collection apps on their phones to submit their data to a central system. The central system would facilitate some cross checking and verification of the data. It may not have the internal consistency of an Opta system but it could benefit from the wisdom of crowds. Such a system is a possibility. However, it would require some planning and regulation.
The biggest issue would involve deciding on the exact definition of the data to be collected. Let me illustrate some of the issues using data on the 2015 Munster Senior Hurling Final between Tipperary and Waterford. I'm going to use the stats provided by the broadcaster of the live event (RTE) and two newspapers (Irish Examiner and Irish Independent). Below are the scoring statistics. The numbers in blue font are "correct" but many of those in black font are open to interpretation.
The problem comes from the definition of a scoring chance. Are balls that trickle wide over the end line defined as a scoring chance? Are all balls that that the goalkeeper collects from an opposing player defined as a scoring chance? What happens where a player has possession in the middle of the field, 50m from goal, is unchallenged, and decides to play the ball into a team mate 10m from goal? It is clear that there would need to be some agreement on what constitutes a scoring chance. Defining this is a bigger coordination problem for a wiki but it remains a problem for Opta and others. The commercial operations would also need to answer these questions and train the coders appropriately.
The numbers in red font (above) are a problem that is easier to explain and clarify. It seems that 65s are classified as frees. The denominator is harder to explain. Waterford committed 11 fouls. What frees are not included in the denominator? Ones that are outside the scoring range? Again there would need to be some common understanding on these issues.
Both teams hit one free wide. They also struck one wide each from a side line puck - something not included in the statistics above.
The difficulty with definition carries over into what we might call the "possession" statistics. Here are the statistics from RTE, the Irish Independent, and the Irish Examiner. Not all sources agree on the stats to be collected and there is some disagreement on the exact numbers.
The numbers in red font (above) are a problem that is easier to explain and clarify. It seems that 65s are classified as frees. The denominator is harder to explain. Waterford committed 11 fouls. What frees are not included in the denominator? Ones that are outside the scoring range? Again there would need to be some common understanding on these issues.
Both teams hit one free wide. They also struck one wide each from a side line puck - something not included in the statistics above.
The difficulty with definition carries over into what we might call the "possession" statistics. Here are the statistics from RTE, the Irish Independent, and the Irish Examiner. Not all sources agree on the stats to be collected and there is some disagreement on the exact numbers.
The stats on discipline also show a difference of one in the frees Tipperary conceded. A possible explanation is the 13m conceded by Tipperary on a Waterford free during the second half. It is interesting to note that Waterford players attracted a yellow card for over 50% of their fouls. It is also interesting that they conceded 5 frees to Tipperary's one foul in the second half.
What other statistics should be included and how would they be defined? These are all questions that would need to be considered. What should we make of the fact that there were 17 side line pucks (with 13 of those going to one team)? Or is it important to know that both teams got two 65s? Do we need to know the number of hand passes in a game (59 with one team performing 37)?
What statistics would we need to examine the role of the "sweeper"? It is hard to know. However, some numbers might suggest that there is greater variation to the "sweeper" role than is acknowledged. Padraic Maher (Tipperary) collected the ball on three occasions where it was battled or blocked by the Tipperary full-back. By contrast, Tadhg de Burca (Waterford) got three hand passes from his full-back. There was also a difference in what each sweeper did with the ball. Maher accounted for 33% of all Tipperary hand passes on the day (and he also delivered a few short stick passes). By contrast, De Burca accounted for about 5% of Waterford's hand passes.
McGeady's suggestion is a good one but it will require a bit of thought. Maybe the GAA should facilitate this discussion.
What statistics would we need to examine the role of the "sweeper"? It is hard to know. However, some numbers might suggest that there is greater variation to the "sweeper" role than is acknowledged. Padraic Maher (Tipperary) collected the ball on three occasions where it was battled or blocked by the Tipperary full-back. By contrast, Tadhg de Burca (Waterford) got three hand passes from his full-back. There was also a difference in what each sweeper did with the ball. Maher accounted for 33% of all Tipperary hand passes on the day (and he also delivered a few short stick passes). By contrast, De Burca accounted for about 5% of Waterford's hand passes.
McGeady's suggestion is a good one but it will require a bit of thought. Maybe the GAA should facilitate this discussion.
RSS Feed