Get the latest news and insights from 21st Club

Please read our Privacy Policy carefully to understand how we use your data. You can opt-out of our marketing communication at any time.

MailChimp Pop-Up

In football analytics, the media isn’t always the message

Despite a few one off columns here and there—see Martin Samuel’s clanger on the alleged superfluousness of the marginal gains provided by analytics in football and other sports—the relationship between football stats analysts and major media organizations covering the sport has been relatively cozy over the last few years, particularly compared to ice hockey in North America. While the development and application of statistical science in the National Hockey League is the subject of fierce and often hostile debate in North American media circles, football analytics seems to be the quiet purview of a few open minded print journalists like Sean Ingle, Adam Bate and Jonathan Liew. Most football writers simply ignore it.

There are a number of possible reasons for this.

One is the inherent conservatism of many soccer analysts, who are generally very careful to temper any findings with the caveat that these predictive measurements like Expected Goal values have yet to be put to a rigorous historical test. As the field is still so new, few are willing to make grandiose claims over their findings no matter how predictive they may be, and that tends to attract less media attention, positive or negative.

A second reason is that there are currently very few public, predictive and repeatable individual player metrics in football to use as talking points in arguments over which team should should buy which footballer. This isn’t the case in hockey, in which some analysts will rake traditional media over the coals for praising players with poor individual metrics, whether Corsi or number of zone entries etc. This kind of argument fuels much of the hostility between traditional media and independent analysts.

A third reason is that many of the findings so far in football analytics are intuitive: counterattacks open up space and allow for better chances, better shooting positions lead to more goals, teams that shoot more as a rule win more games etc. While puck possession for example is often held up by hockey analysts as an ideal mode of playing over the dump-and-chase approach, we’ve yet to see a hard Maginot line form between media and soccer stats analysts on a single, ideal way of playing the game.

This detente could all come crashing down at any moment of course. Perhaps those individual player stats will emerge, leading to a host of players favoured by “numbers guys” and disparaged by traditional media. Or, as is more likely, a club will hire a popular analyst, make the news public—perhaps as a temporary PR move to signal a progressive change, particularly if the club is struggling—and then fail to meet unrealistic expectations sparking a backlash against the use of analytics in football.

There is a concern this will happen in ice hockey following last summer’s public analytics hiring spree by the NHL’s Toronto Maple Leafs and the Edmonton Oilers. Both teams are badly struggling this season, leaving them vulnerable to a very lazy cause-and-effect criticism—”Maybe you shouldn’t have hired those nerds.” While most journalists—even those publicly skeptical over analytics’ impact on hockey—have been remarkably restrained to this point, that might all change at the end of the season.

The problem is stats analysts have always had the deck stacked against them when it comes to media. Though the RAF wing commander Charles Reep is still the subject of scorn in England for using stats analysis to promote long ball football in England beginning in the 1950s, few credit his methods for aiding Wolverhampton Wanderers’ success under Stan Cullis, which included three first division championships.

One problem is that mathematical analysis and literary sportswriting are uneasy bedfellows. It’s hard to wax poetically about how a spreadsheet helped a team to eternal glory on the football pitch. Yet the real issue for sports media in understanding the impact of analytics is that they are only ever getting roughly half the story. To truly understand whether statistical analysis in football is working or not working at a club, we have to know both the process and the results. Most journalists—most of us, in fact—are on the outside of clubs looking in; they only ever know results.

Here is a hypothetical example. A club is largely satisfied with their squad depth ahead of the January transfer window, yet it contemplates making a low bid on a striker who is considered to be just past their peak performance age. The reason is their technical scout has metrics that suggest that while the player won’t likely score many goals or start many games, his style of play has a subtle but dramatically positive impact on defense. It may be worth taking the player on if only to sub him in when in the lead in the last fifteen minutes. So the chief executive calculates a reasonable transfer bid and adds the player to the team.

Lo and behold, the player doesn’t start often and scores zero goals for the remainder of the season yet has a dramatic impact on defense, holding up the ball and tracking back to help secure a lead late in the game. Most sportswriters however are inclined to look at the player and their position and judge their impact solely based on their role—strikers are supposed to score goals. Not only did this player not start any matches for the team but he only came on a sub periodically and didn’t score once, and is therefore given a failing grade in most of the end of season reports at the end of the year. What these critical journalists don’t know—can’t know—is that this player in fact did their “job” beautifully.

There are others situations too where a player clearly fails to succeed on all counts, statistical and otherwise, and yet are still part of an overall club recruitment process that had better odds of success than a team picking players based on the “eye test” alone. Again, unless the media is privy to this process—and proprietary concerns ensure they rarely will be—they don’t have enough information to make a judgment on this, positive or negative. Nevertheless, most journalists will be tempted to conclude the club “screwed up,” when in fact they made the right gamble at the right time and lost anyway.

While the easy solution may seem to be for clubs to simply ignore the media and keep their analytics program secret, most fans read the papers and will likely draw the same conclusions as the journalists writing for them. That can put immense pressure on teams to abandon their analytics processes altogether. As internal statistical analysis continues to permeate professional football in England and Europe, clubs will have to tread a fine line between educating the public on what exactly good statistical analysis can provide, and ensuring they don’t give away too much. Unfortunately, there is no secret formula for getting the balance right.

And, if there is, I’m not telling.  

About Richard Whittall

Richard Whittall has created 28 entries.