Monday, October 06, 2014

Looking Past the GAA's Black Card Propaganda

The science of statistics, for all the black arts associated with it, has one golden rule. It is this: correlation does not imply causation.

Because A and B happened in sequence does not mean that A caused B, or that B is the result of A. If it starts raining on the day you leave home without your coat, that does not mean that your coatlessness caused the shower. It is much more likely the rain was caused by the meeting of weather fronts of different temperatures than your own childlike optimism.

In much the same way, the black card statistics trumpeted so loudly by the GAA at the end of last week should be met with a certain skepticism. All statistics should be met with skepticism of course, but ones make causal claims as – shall we say, ambitious? – as these are very difficult to take.

The press release on the GAA’s own website boldly claimed that “With the introduction of the black card, the average number of points per game in the 2014 championship is roughly 9.5% higher than in 2013; the number of points scored has increased by just shy of five points per game since 2010.”

When you read “2010,” you may have heard a loud whirring sound. That is the sound of spinning. 2010 has nothing on God’s green earth to do with the black card. The only reason its tagged on there is because five sounds like a lot.

The black card has not been popular, not least as so few people know what exactly a black card offense is. But rather than admit they got it wrong, the GAA finds itself like the Black Knight in Monty Python and the Holy Grail, denying the bleeding obvious. They are fooling nobody who takes a minute to think about it.

But if the black card isn’t the reason there were more scores in Gaelic football in 2014 than there were in 2013, what is? It could be any one of a number of reasons.

Gavin Cummiskey in the Irish Times theorises that “Dublin’s record -breaking summer of scoring must be factored into the increase.” Bless. This blog has taken Cummiskey at his word, and has indeed factored Dublin’s record-breaking summer into the increase.

If you remove the games that Dublin won from the 2014 scoring averages, the average total score per game drops from 34.92 to 34.4. The earth really didn’t move because of Dublin.

So what could it be? Are there any patterns deeper in the data?

Here are two tables. The first shows the average total points per game since the qualifiers were introduced in 2001, broken down by year and competition (the four provincial championships, the qualifiers and the All-Ireland series that starts in August).

And here are the average margins, broken down the same way.

The numbers are colour-coded, from green for the highest totals or margins, into white for average, down to red for the lowest. There is no sharp correlation between margin and totals per game, but there is certainly a case to be made for further investigation into the idea that the current inequality of the Championship is a greater factor in more points being scored.

Are more points being scored because more hidings are being handed out than heretofore? Look at Connacht. Is it a co-incidence that the highest scoring totals coincide with the current Mayo dominance?

Look at Ulster. Ulster is consistently lowest in totals and lowest in margin of victory. But isn’t a low margin of victory a good thing? Doesn’t it mean the games were competitive? Doesn’t everybody know that Ulster is easily the most competitive province? So why are the GAA squawking about point totals as measures of football excellence?

The statistics for the All-Ireland series sit badly with the theory that the greater scoring is significant of teams getting hammered rather than beautiful football being encouraged by the introduction of the black card – the high totals are not matched by high margins, as they are for some years in the other competitions.

But there are mitigating factors here. Firstly, the last eight teams in the country are the best teams in the country. This has been established in both theory and practice over the past five or more years. Naturally, then, these teams score more than teams that aren’t as good.

It was also in the All-Ireland series that the referees’ reluctance to issue black cards became glaringly obvious. The GAA press release tells us that black cards were issued at a rate of 0.8 per Championship game. How many of those black cards were issued in the All-Ireland series?

Four black cards were issued in the eight games of the All-Ireland series. One in the four quarter-finals, two in the three semi-finals, and one in the final. That’s a rate of 0.5 cards per Championship Game among the games that were the highest-scoring of all, contrasted with the 0.8 overall average in the Championship.

The GAA are taking the seanfhocal literally, and are trying to say black is white. End this black card farce now.