Oh lovely, another article on xG.
I too am tired of talking about it, but that doesn’t mean it isn’t important. However, I don’t want to rehash what xG means and what it’s for; Bobby Gardiner has done a swell job of that here.
Nor would I dare presume as a layperson to pull a Marek Kwiatkowski and start talking about how and why we must move on from this Dark Ages metric into something vastly more complex fruitful (though Marek is ultimately correct there).
While I don’t want to tell other people what to do, think it behooves stats analysts to take a moment to stop tinkering with the building blocks of football analytics—in this case, metrics like xG and its myriad relatives—and start accepting them as a fait accompli. I think we need more of the holistic analysis that reveals football analytics at its very best.
Now, this is not at all to say no one is doing this sort of thing; of course, they are. Sites like Statsbomb and American Soccer Analysis have been leaders in this space for a while, and Paul Riley, in particular, has done some excellent work over the years in this field, as have Mike Goodman and Michael Caley.
But I think as a whole, football analytics might get a little more traction if it abandoned the endless nitpicking over names, metrics and models and refocused its efforts on more mid-to-long term diagnosis and treatment.
In general, this would mean less adding up xG totals after individual games, and more looking at 5-10 game trends. It would mean less relying solely on xG and it’s derivative stats (xA etc.), and more using xG in tandem with other statistical tools, as well as video and tactical analysis, to present a more vivid theory as to why a team may be underperforming. It would mean less obsession with which teams are ‘riding their luck’ or are ‘unlucky,’ and more to see why certain clubs fail to create or prevent dangerous chances in the first place.
In other words, a football analytics gesamtkunstwerk.
I think this general approach is good for a host of reasons. First, it grounds xG and other related stats in a real football context, which makes it more palatable to a wider audience. In my experience, basketball analytics tends to do this very effectively, by using basic metrics to divide teams by on-court defensive or attacking patterns, and then zeroing in where teams may be failing to implement these strategies.
Second, it helps identify gaps in our understanding of the game and starts giving us a clearer sense of the boundary of our understanding of the sport, both from a tactical and statistical perspective. This may, in turn, lead to new statistical tools to help fill these gaps, something that used to happen frequently in football analytics.
Third, and this is related to the second point, it will help prevent the creation of metrics for their sake. Often the introduction of a new metric is viewed as its own accomplishment, without an understanding of if and how this new metric improves on what’s used today. An analyst must be more than a toolbox, and so new metrics should only be introduced if they provide a vast improvement on the existing ones, or fill in a significant gap in our understanding.
Finally, this can help analysts get out of the habit of staring at ten matches every weekend and feeling pressure to make definitive claims about every single one. In my view, analysts should be wary of getting caught up in focusing too much on individual match outcomes, even when applying metrics like xG. After all expected goals aren’t magic; while they are a better predictor of long-term quality than goal difference or league points, they too take several games to stabilize into a trend. Therefore, discussing which team ‘should have won’ related to single game xG totals is problematic, because even here teams can get lucky on the day.
Again, I realize this is all very high level and that doing this sort of analysis is a lot harder than writing about doing this kind of analysis. But I think if we’re interested in moving xG past a Jeff Stelling punchline, it will behoove football statisticians to once again try to get a better grasp of the whole picture.