This Article is 1.37x More Interesting Than Other Articles About Expected Goals
What kind of football person are you? Your feelings about expected goals or xG might just tell you. The latest stat fad for football broadcasters, expected goals has become a flashpoint in the world of football punditry between ‘Proper Football Men’ and ‘sentient stat-borgs’. For those who aren’t clear, expected goals is a way of assigning a difficulty rating to every attempt on goal; if a chance is 0.3xG, it should be scored 30% of the time. To work out a team’s xG for a game, you just need to add up the xG rating of their individual efforts. ‘Proper Football Men’, who have little time for vague possibilities — football is about results — see xG as an existential threat to their clan, an imposition by nerds who dare assert their authority in a sport they were probably rubbish at. Gillette Soccer Saturday’s Jeff Stelling encapsulated this attitude when, surrounded by his very own gang of sniggering Real Football Men, he derided expected goals as “the most useless stat in the history of football”. 1-0 to the real football men. No decimal points here.
But does Jeff have a point? Well, sort of. There is something odd about producing a stat like xG after the game, in defiance of the actual score line. Stelling’s rant was triggered by an Arsene Wenger post-match interview in which he cited expected goals to suggest that Arsenal’s 3-1 loss to Man City was closer than the scoreline let on. And used in the context of a one-off game, expected goals does offer a particularly shallow explanation of a match. It doesn’t, after all, give us any sense of the external factors affecting a game; did one team have an off day? How did their approaches to the game differ?
But then again it was never supposed to, and certainly not for a single match. Without rehashing the numerous, very good articles on the value of expected goals, it goes without saying that what can be a useful analytic tool for coaches and players in identifying patterns and trends, is next to useless in assessing a one-off game. But this is how we, the viewers, usually encounter xG. Match of the Day, for example, puts the xG of a match at the bottom of the screen after every game, but rarely offers the stat in the context of a longer series of fixtures, where it might tell us something of value.
Indeed, Wenger’s interview suggests that the viewer at home might not be the only potential victim of a broadcasters’ sloppy or misleading use of statistics. Broadcasters, to an extent, assume responsibility for defining the vocabulary we use to talk about and analyse football. We might expect better from Wenger, a club manager since 1984 and someone known for his intelligent approach to the game. But is he, or any manager for that matter, likely to worry too much about using xGs misleadingly when broadcasters have practically offered it to him as an excuse? Despite Stelling’s protests, it is his world that put those words in Wenger’s mouth.
Whether Wenger’s appeal to xG is genuine or not, his use, or indeed misuse of it, does speak to a broader concern over the use of data and analytics in popular sports analysis. Sky Sports are particularly guilty of bombarding the viewer with unprocessed stats, divorced from any suggestion of what purpose they serve or how they might inform us about the game we are watching. It is left to the viewer to determine the significance of the fact that Ashley Barnes has scored in two of his last three away games, or that Crystal Palace have failed to go beyond the fourth round of the FA Cup in three of their last six attempts. In a game already overflowing with misleading clichés, broadcasters should be using data to deconstruct some of their viewer’s assumptions about the game. Instead, they seem set on constructing new clichés, this time based on the delusory authority of data for data’s sake.