Soccer analytics took another step into the mainstream last season. Can we expect further advances in 2018-19?
The slow but steady shift toward analytics in soccer media has had a two-fold purpose: To give fans a more in-depth understanding of the game, as well as to keep up with the coverage of the other major American sports that have increasingly benefitted from the advent of statistical-driven analysis.
From the broadcasters’ perspective, the challenge is giving viewers a better feel for the game through the use of data, without making them feel a degree in statistics is required to understand the analysis they’re being offered.
The vast majority of statistics offered to the American viewer at large are quite basic, consisting of things like shots, shots on goal, corners, fouls, offsides, possession percentage and passes completed. As things stand, you don’t have to be a savant to interpret this data in order to understand which team had the better run of play.
There has been very little differentiation between the likes of NBC Sports, Fox Sports and ESPN in terms of their usage of statistics, as evidenced by the coverage of the last editions of the World Cup, Champions League and Premier League. None of the main soccer broadcasters in the U.S. has shown a willingness to dive deeper into the world of more sophisticated statistics.
Given the wealth of game data already available, there seems to be a tremendous opportunity to offer new, slightly more advanced statistics. As we head into a new season, here are a few metrics that could help improve soccer coverage in the U.S.
Expected goals (xG)
xG, once used almost exclusively by the nerdiest of analytics nerds, took a big step into the mainstream last season, as the BBC began listing xG totals on its Premier League highlights show, Match of the Day.
xG measures the likelihood of a given shot resulting in a goal, based on shots taken from similar positions in the past. If you shoot from 70 yards out, you probably have about a 1 percent chance of scoring (an xG value of 0.01). If you shoot unmarked from 10 yards out, you probably have something closer to a 70 percent chance of scoring (an xG value of 0.7).
The exact xG value of a given shot is based on extensive data on shots taken in the past by professional players, including shot location, the type of pass that preceded the shot and much more.
If a team outperform their xG — that is, score more goals than you would expect based on the shots they took — it likely means they either finished particularly well or got lucky (or both).
xG, like almost every advanced stat, has met with resistance among traditionalists, but as the BBC showed, its ripe for use by major broadcasters, since it can be listed easily next to other more straightforward stats like shots on target. This of course leaves out a lot of explanatory detail, but it can lead interested viewers to do more research into these figures in their own time.
Given American fans are typically more receptive to statistical analysis of sports to begin with, it seems plausible that, say, NBC could begin to incorporate xG into their analysis of the Premier League.
Passing stats 2.0
As the past several years have shown, passing the ball a lot doesn’t necessarily contribute to winning. Russia knocked Spain out of the World Cup this summer despite completing 800 fewer passes. Unsurprisingly, a high possession percentage can also mean very little. It’s not how much you have the ball that matters, but what you do with it.
This isn’t particularly revelatory information, and yet major broadcasters still tend to trot out possession percentage figures without any context. There are plenty of more detailed, and still easily understandable stats, available for free on sites like WhoScored, which measures key passes (passes leading to a shot), short passes and long passes.
Some of these passing stats are clearly correlated with the league table. Manchester City led the Premier League in key short passes last season, while key long passes doesn’t seem to be a (strong) predictor of success (Burnley, Tottenham and West Ham led the way in terms of key long passes per game for 2017-18).
There are other, more advanced, passing metrics available, but the initial use of more simple stats like key passes seems like a valuable step in developing a more statistically literate audience.
Duels won 2.0
Duels won stats include things like aerial duels, dribbles completed and tackles won. These have found a place in mainstream soccer coverage recently, and occasionally featured on Fox’s coverage of the World Cup.
To take thing further, broadcasters can simply present these figures as percentages. Completing a lot of dribbles isn’t necessarily a good thing if a player is failing to complete dribbles twice as often. Similarly, dribbling too much can be a sign of over dependence on certain kinds of attacks.
A more advanced version of these metrics might try to incorporate the degree of difficulty of a particular dribble or tackle. It’s easer, for example, to take on a back-pedaling defender in lots of space than it is to take one on from a standing position in a crowded penalty area.
These are only a few options broadcasters could choose from to expand their use of statistics. Finding a way to incorporate stats into coverage in an entertaining, accurate and informative way is a difficult challenge, and not something that will necessarily come easily to the likes of Alexi Lalas and Fernando Fiore. Or perhaps Fox might replace Dr. Joe, the “rules expert,” with a more seasoned (and lively) stats expert.
Regardless, as stats become more and more a part of our analysis of the game, it will be interesting to see how broadcasters incorporate them into their coverage. It’s likely to be a gradual, baby-steps process, as viewers become more accustomed to, and less intimidated by, these numbers.
However, it’s also likely that if major broadcasters fail to keep up with the level of statistical analysis available on the web, fans will begin to scoff at the current use of statistics, driving demand for a more sophisticated use of data in the mainstream.
Add us as a preferred source on Google