The number’s up for those ignoring statistical analysis

Forget splashing out on a big striker. Here’s why many clubs would be better off buying a couple of nerds, a few laptops and a Wyscout subscription.

By Joel Sked

This article first appeared in Issue 4 which was published in June 2017.

quotation mark
The average bottom six Scottish Premiership squad player will be on around £1,000 a week, £52,000 a year. That’s enough to hire two full time analysts, buy a few laptops and get a Wyscout subscription.

Ebbe Skovdahl spent three-and-a-half years in charge of Aberdeen in a spell which is best described as tumultuous. There was a bottom-place finish and an embarrassing European exit to Bohemians, but also two cup finals. Yet, it wouldn’t be denigrating in the slightest to say the avuncular and enigmatic Dane is best remembered for the things he said rather than the things he did. Type ‘Ebbe Skovdahl’ into Google and the first suggestion is ‘Ebbe Skovdahl quotes’.

There were some belters but few stand out more than the one which compared statistics to mini-skirts. The assessment was offered in 2001, after it was put to Skovdahl that Arild Stavrum had more shots on target than Henrik Larsson. His point was that statistics gave you a good idea of the game but they hid the “most important thing”: the score.

There should be no reason needed for a trip down memory lane when it involves Ebbe. However, in this case there is a reason. Sixteen years on, statistics are creeping into everyday discussions surrounding the game in Scotland. A burgeoning community of analysts are watching games from a statistical point of view, foraging through the numbers and presenting their findings in a range of graphs, charts and radars for public consumption.

Taking on Skovdahl’s metaphor would lead us down a dangerous path but the five words which are still pertinent to this day are “they give you good ideas”. No longer should they be seen to “hide the most important thing”, but rather, illuminate it. Instead of being looked upon as a form of obfuscation, data, if collated and presented clearly, can provide guidance and aid us in understanding the most important thing.

In a recent BBC Radio 5live interview with Ted Knutson, owner of statistical company StatsBomb, who has worked with Brentford and Danish side FC Midtjylland, presenter Darren Fletcher introduced the subject with an almost disdainful “it is impossible to ignore number crunchers these days”.

Within the Scottish football landscape, on Twitter at least, three individuals have become “impossible to ignore”. A trio endeavouring to provide fans with greater insight and understanding of what they are seeing on the pitch.

Dougie Wright, Stewart Brown and the Clevland, Ohio-based Matt Rhein are all passionate followers of Scottish football, football addicts if you will. And all have been inspired, one way or another, to watch game after game, and ‘crunch the numbers’, to provide greater depth.

While we all love Scottish football, its propensity to entertain, bamboozle and amuse, we can all agree that unlike the country’s history of innovation, the sport has been slow to evolve and modernise. In certain cases it has failed to do so at all.

Clubs have been glacial in their reaction to developments in the way the game is analysed. They appear ambivalent to what is in it for them. It is a mindset which has left Scottish football, once again, failing to see the bigger picture, to embrace an area which is becoming increasingly important throughout the football landscape.

Statistical analysis could be regarded as being in its embryonic stages in the sport, despite books on the subject being published as far back as 2009. The pioneering Soccernomics, by Simon Kuper and Stefan Szymanski, was the first of its kind. In 2013, Chris Anderson and David Sally released The Numbers Game. They are books which have been an inspiration, opening the eyes of many to what numbers can teach them about the game they love.

And, of course, there is the work of Michael Lewis and the pervasive Moneyball, a book which explored Bill James, Billy Beane and the statistical rebellion undertaken at the Oakland A’s against the conventional approaches in baseball. To say it changed the sport would be hyperbolic but it played a key role in shaping modern thinking within Major League Baseball.

The term ‘moneyball’ has now somewhat incongruously found its way into football parlance. I say incongruously because many who use the term overstate its meaning or simply don’t understand it. While its constant use to describe any sort of statistical philosophy is wrong, it has provided people within football circles the encouragement to explore the edges and gains which can be mined from data.

Through the work of OPTA Stats and StatsBomb, as well as websites such as WhoScored.com and Michael Caley’s Cartilage Free Captain site, analysis from a statistical viewpoint has steadily forged a path into the mainstream south of the border. Informative and entertaining stats-based articles are common, while data is commonly used on Sky’s Monday Night Football programme to back up arguments.

So why has it not seeped into the mainstream in Scotland, as well as being a part of clubs’ framework? Time and money? Most definitely. Resistance and scepticism? Most likely.

Matt, who runs the The Backpass Rule website from the US, says: “I believe a lot of the scepticism about analytics and stats in football in Scotland is due to the fact there has been much less exposure to them than in England. The more exposure and understanding there is with them, the less lashing out against them without understanding there would be.”

This is a mentality, an attitude, which isn’t just specific to statistics. It is a mindset that is prevalent across our football culture, whether it be tactical analysis or, say, the dynamic between a certain young coach and a certain director of football. Sometimes we, as a nation, can be guilty of being closed-minded, suspicious of anything different.

However, not only are stats here to stay – they will become commonplace. Eventually.

Dougie, who volunteers as a performance analyst in the academy of a Ladbrokes Premiership team, recognises that “there’s been a lot of change over the past year or so” with the caveat that there is a “long way to go though”.

While Matt is aware of only “one SPFL Premiership club who are starting to develop an analytics program” while another which uses “the former author of the Rangers Report blog as an analytics consultant”.

“That means there are possibly 20 Ladbrokes Premiership and Championship clubs who likely don’t have any analytics programme and thus 20 clubs you could gain easy advantages over implementing analytics and the ‘low hanging fruit’ associated with it.”

The benefits are bountiful and diverse. Take the clubs and players. What is the one thing we all try to avoid in our jobs? Mistakes.

Stewart, who was inspired to create The SPFLRadar after watching the Sloan Conference, an all-sport analytic conference in the US, says: “Through data teams can have an idea of what on average a player does per game, and when the player produces lower or higher than the base average they can inspect why through video.”

And it is not just the clubs who can benefit. Us mere punters can be enlightened. If we want to be, that is. Some fans just want to watch and be entertained. Others want to know why something happened, how it happened, and whether a pattern is developing.

“There are a lot of people who think any football game can be won by 4-4-2, two quick wingers and a big striker,” says Dougie. “They’re not really interested in over-complicating football, and I’m not going to spend too much time trying to convince them to look at things differently.”

He continues: “It’s an evidence based approach, so I think it will benefit most of the game’s stakeholders when they embrace it. Attempting to quantify what happens on a football pitch means you can start making objective comparisons. Are we better than we were last year? What’s the gap between us and second place? Are we improving because we’ve got better strikers, or a ‘less worse’ goalkeeper?”

Matt adds: “As far as fans and journalists go, I would never hold it against anyone if they don’t care to enjoy their football this way. Some people just want to enjoy what is on the pitch and that is fine. However, some use analytics as some straw man to lash out against something they do not understand. Often times people will make outlandish sweeping statements about stats in football but when I discuss it with them they realise they had a misconception about what exactly they are.”

The most popular use of the term Moneyball is in conjunction with recruitment: ‘Moneyball signings’. At the time of writing Paul Hartley has just been sacked by Dundee. One of the common criticisms of his time in charge at Dens Park was his recruitment. More than 40 players were signed but very few could be regarded a success. This, for Stewart, is where clubs can perhaps benefit the most.

“Data can be used as a filtering system within a recruitment process,” he explains. “For example, Dundee may need a central defender. Due to Dundee not having a lot of possession per game this defender must have experience of being in a side that defends most of the time. A defender playing in such a team will have a high rate of interceptions per game and more involvement in aerial duels due to being without the ball.

“The filtering process has now narrowed your market down to central defenders that may fit your team well. Your scouts could now take a look at these players and narrow down even further to the right prospect. Data isn’t here to take over, its merely more information for a person to use in connection with what they see to make a more informed judgment on a player.”

There is every chance Hartley would still be prowling the technical area at Dens if he opted against signing a Danny Williams or a Nicky Low and the club invested in an analytic programme.

Dougie, an analysis masters student, notes: “The average bottom six Scottish Premiership squad player will be on around £1,000 a week, £52,000 a year. That’s enough to hire two full time analysts, buy a few laptops and get a Wyscout subscription.

“Who’s more valuable: the guy who never gets off the bench and sulks behind the scenes because he doesn’t get a game, or a semi-decent analysis department who can tell you that the team you’re playing on Saturday don’t know how to defend a near-post cross?”

Yet, there is a clear sense Scottish teams have missed an opportunity to bridge the gap between our domestic game and more elite leagues, to get a yard or two on more wealthy clubs. “I think both the ‘proper football man’ and ‘stats nerd’ would agree that there is a large gulf in financial resources available to the SPFL, Celtic included, compared to the EPL, La Liga, Bundesliga,” says Matt.

“Using analytics could have been a way to find inefficiencies and help alleviate some of the financial advantages the bigger leagues in Europe have in recruitment, scouting, and other areas.

“Just look at a club like FC Midtjylland in Denmark. Midtjylland famously used stats to find various inefficiencies and used it to help them win their first league title in club history, as well as achieving unprecedented European success. However, I think that many of the advantages that could have come with an analytics department at a club in Scotland has been lost, at least when compared to the bigger leagues in Europe.

“Many ‘big clubs’ like Arsenal, Manchester City, Bayer Leverkusen and more are using analytics regularly. Scottish football missed a chance to use these innovations to make up for the large financial gulf it faces.”

That, however, should not act as a deterrent for clubs looking to break ground now.

It is acknowledged that to increase buy-in from fans, the media, clubs and players, communication needs to be clear and concise. Especially with players; after all they are the ones going out on the pitch and performing. As noted by Ted Knutson in his Radio 5live interview: “More information…given in a way they like,” is going to make them a better and more efficient worker.

Matt says: “Of course, how you deliver these messages is what is important. If I went up to an SPFL striker and said: ‘Your expected goals is low, it needs to improve,’ I would probably get a response somewhere in between befuddlement or laughter.

“However, if I said: ‘When you shoot from outside the box you only score about once every 20 shots, however if you shoot in the box within the frame of the goal, you usually score about once every four shots. If you shoot more there, you will likely score more,’ you will likely get your message across better.

“Both of those statements are essentially saying the same thing and are from analytical reasoning, yet the use of jargon makes one statement less relatable.”

The use of jargon is something which Dougie tries to steer clear of, using bars and graphs to help illustrate points. He says: “I do believe that a lot of the resistance probably comes from guys like myself not explaining clearly enough what we mean. It’s so easy to take for granted that everyone understands what you’re on about, but that’s not always the case.”

One such phrase is the infamous xG. A quick search of ‘Craig Burley expected goals’ will bring up results of an incredulous former Scotland internationalist losing the plot over the mere mention of expected goals.

To the uninitiated it is shorthand for ‘Expected Goals’. Perhaps the most popular and widely used stat, certainly within analysis from a blogging and fan point of view, it is simply analysing the goals teams or individual players are expected to have scored.

“It was quickly established that corners have zero correlation with goals scored,” explains Dougie. “However, people noticed that the amount of shots you took had a relationship with goals. Then they took that, examined it, and found that there was an even stronger relationship with the quality of shots. And now we’ve got these expected goals models which correlate even more strongly with goalscoring, and those will probably get replaced by something even more robust further down the line.”

Matt adds: “I think everyone should look at a stat and think about it in two ways. First, does this stat make sense mathematically?

“The reason expected goals is so widely cited when discussing football stats is that it has been shown mathematically to be a better predictor of future success than things like goals, wins, points.

“Expected goals certainly has its flaws and I would guess many clubs in Europe have something better they are using but not sharing publicly but it is the best we have now in the public.

“Second and more importantly, does this stat make sense in relation to the game? All of these stats need to make footballing sense. Again to use expected goals as an example, it makes sense to me that a team that has more high percentage scoring opportunities than their opponent over the course of a season will be more successful over that season.”

Maligned though it may be in some quarters, America’s Major League Soccer gets a lot of things right: marketing, broadcasting, engagement and more. It is a league that isn’t scared to embrace new ideas and innovate.

This season it has brought stats, such as expected goals, into its coverage, showing it in the same way that our broadcasters would display possession and attempts.

Dougie would like to see it taken a step further, and when it is, he feels substantial progress can be made. “The national association should probably carry some more responsibility,” he says. “There’s no reason why they can’t have someone filming every league game, doing some basic analysis, then sending it back to the clubs.

“Although the clubs would benefit from doing more themselves, there’s no reason why the SFA couldn’t nudge them in the right direction. Hopefully by then you’ll have had a club who, rather than sign a couple of squad players, really invests in a well functioning data analysis department.

“You pair that up with a coach who understands how it works and what it does, and there’s a massive distance that club could go. Once you’ve got that first mover proving to be successful, the others will follow.”

Bravery. It is an intangible quality which resonates with Scotland and Scottish football. Scotland the Brave. Project Brave.

“If Scotland is to regain its previous reputation as a footballing nation,” says Matt, “it needs to look past traditional methods. Traditional methods have gotten us to where we are now, we need to be brave, pun fully intended, to get out of that position.” 

This article first appeared in Issue 4 which was published in June 2017.

Issue 11
out now

Subscribe here Buy a gift subscription Back copies