Andy McGeady: Warning - contains data, use carefully

Statistics provide great insight but can be dangerous in the wrong hands

Reducing sporting achievement to a few digits, a grade, or a ranking, can seem unromantic. It’s a touchpaper for holy wars in the US over such things as MVP and Hall of Fame debates.

There’s danger in the rage to conclude; X over Y in a pure, black and white manner. Even for the most ardent of sporting numberphiles, the ‘magic number’ way of looking at sports is fraught with difficulty.

Dr Seamus Kelly said that when putting together the various modules of UCD's MSc in coaching and exercise science they looked at gaps in coaching. While they cover general coaching theory as well as experience from specific sports (the university also offers an MSc in rugby management), Dr Kelly sees two key issues in the use of data by sports teams.

Firstly, they need the ability to interpret data correctly. Secondly, they then need to be able to transfer findings to the coaches and players in a meaningful way.

READ MORE

The former Ireland B soccer international said that while data was just one piece of the jigsaw, if used properly it can provide huge insight into players and opponents. But there are dangers. "There can be a massive overemphasis on stats that may not matter being used by people who don't know what they're doing," said Dr Kelly.

Absolute faith

In football the rush to use a player’s distance covered as a measure of workrate and worth has been gently but firmly moved to the dustier parts of the archive. It’s in the corner marked ‘things we thought were good’. But perception changes based on stats can change even more quickly. For example, one should not place absolute faith in the stats released on the full-time whistle.

Football and rugby are fast-moving sports. In football the movement is almost constant during the 90 minutes of play, while in rugby the participants can be quite literally piled on top of one another. Which player actually made the tackle? Perhaps two of them did so jointly? Who actually effected the turnover?

While there’s a thirst for stats available at the full-time whistle, the folks who code matches in real time can be only so accurate. So then there’s a review of the stats, which is quite sensible given the speed of the preceding two halves of action. And that means sometimes those stats can change.

Luke Charteris was initially reported as having made 37 tackles for Wales against Ireland in the 2015 Six Nations but, when you check the stats now, six of those have seemingly disappeared.

In this age of data everywhere all the time, the need for speed might seem to outweigh a key characteristic of good data: accuracy. A cynic might say that the media doesn’t care about that – it just wants content. And sure who’s really harmed?

At the inaugural Sportdata & Performance Forum, a meeting of sporting minds organised by Edward Abankwa in Berlin, a technical director of a European nation's Olympic committee asked a very senior representative of a major stats company whether they had any scruples in delivering this full-time data to the media. The Olympic man viewed journalists as being generally inclined to take the stats, cherry pick something without understanding it and then using that stat to confront a coach after a game.

Less favourable

Charteris being asked for comment about his other-worldly 37 tackles is one thing. It was incorrect, but to the player’s benefit. But how about if the incorrect stat being used by the journalist was something considerably less favourable? And that’s without even getting into the differences of interpretation that exist between the various stats-gathering firms and a team’s in-house analyst.

"Anyone using data should know how to use it" IRFU performance director David Nucifora told the Irish Times. "It should be used appropriately; if not, that can lead to a false perception of a player, or a team."

That first tackle figure for Charteris, while not strictly accurate (by Opta’s definitions and standards), was still useful. It rightly drew the attention to a tremendous tackling performance in a game where Irish players seemed to run at him all afternoon long. This has value.

Magnificent schooling L

ionel Messi’s magnificent schooling of Jérôme Boateng was something anybody can appreciate. Similarly, Agen’s Canadian flyer Taylor Paris scoring from behind his own tryline last weekend. They are great moments. But the appreciation of a moment should not preclude the appreciation of something else.

Data can be used to provide an extra perspective; perhaps confirming something already thought to be known or discovering something new and useful. As with many new schools of thought, there will be wrong turns on the journey. But to cast anything to do with numbers aside would be a mistake.