When Leicester astounded the world with their 2015-16 Premier League title triumph interested parties were not slow in coming forwards with their theories as to how the ‘little’ club from the East Midlands had so spectacularly defied the odds. Those with a vested interest in the status quo can explain it away as a freak year, a black swan event, that comes along literally once in a blue moon.
The conspiracy theorists, unable to explain the sudden rise and (this season’s spectacular fall) of a squad of previously journeyman players, have looked for a dark side link, and to doping as the means to explain away the improbable and perhaps impossible improvement of mature players at Leicester last season.
And there’s certainly fuel to add to any hearsay or anonymous speculation due to the link to an Italian sports science consultancy. Claudio Ranieri, who also managed Mapei Centre client’s Juventus in 2007, has apparently worked with the same sports science consultancy on his Leicester project. And Mapei are an organisation that definitely split the crowd due to their previous involvement with high profile cycling and Italian football clubs.
Certainly, even anecdotally, it seems mildly jarring that while the EA Sports speed index had noted just one Leicester player in the list of quickest Premiership players in the previous campaign, en route to the League title, three previously unplaced players (Vardy, Schlupp and Albrighton) filled the index’s first three berths. In addition, captain Wes Morgan, listed an improbable eighth place within the EA Sports speed index, ahead of flying machines like Kyle Walker, Theo Walcott and Yannick Bolassie. There’s something amiss there certainly and perhaps we will never know the truth.
But there is definitely one group within the game who would love to make capital from Leicester’s success. And they are the earnest young analysts with their post grads and laptops who swear allegiance to all things data, Billy Beane, Moneyball et al.
It seems like a stretch to me. From a recruitment perspective, Moneyball might explain the unearthing of an unheralded N’Golo Kante from Caen in France. But surely it can’t provide the explanation as to how or why Leicester City won the league title in 2016?
But let’s first of all dig into what Moneyball actually is.
Moneyball, in economic terms, means ascribing a monetary value to potential recruits and current players. It means acquiring assets that are undervalued by others but which can add a value to your project and selling the assets you hold that are overvalued by others in such a way that there is not negative detriment to your roster.
Now that we’ve defined the basic theory, let’s dig deeper so as to understand just how these so-called assets are classified.
What is an asset? How are they classified?
According to Moneyball theory, an asset is not a player as such. Rather it is the combination of attributes that a player possesses.
Take for example a concept such as ‘being caught out of position’. If a player boasts superb positioning, it is a playing characteristic that can either be an undervalued asset or an overvalued output depending on how you chose to interpret it, what you are looking for as a manager, and how it relates to the player’s role and the overall blend of playing personnel.
If you are Jose Mourinho, a player like Alberto Moreno of Liverpool, who is often caught out of position, will be overvalued (a toxic asset) for a manager hung-up on the management of space.
But for a boss such as Pep Guardiola – who wants his players to dominate possession – the same technical football ability of Alberto Moreno could be an undervalued proposition, albeit with a little bit of tweaking needed on the defensive side of his game.
So, any sense that Moneyball provides a value-free, objective assessment of player is a nonsense. As per this example of Moreno, it is merely an efficient way to sort players in relation to the actual skills they possess, relative to the manager’s requirements.
Using stats to arrive at a solution to implement a system
Whatever tools you use to augment player recruitment the two fundamental questions for a manager are always the same namely:
1) What do I already have to work with that is positive?
2) How do I want my team to play (either as a reflection of existing personnel or as an expression of personal taste).
Taking in the broad sweep of European football, it was notable that championship winning Leicester were the only English Premiership side to feature prominently (top 20) as a notable counter attacking sides in terms of their goals scored from break-outs in 2015-16.
It shouldn’t really surprise that Leicester were an anomaly reflecting an ‘English’ style of football where most teams still prefer dominating possession as a means to foster goal scoring chances. And it may explain why they were so difficult for their rivals in the league to master last term.
Under Ranieri last season, Leicester had the lowest possession stats of any Premiership champion since 2006–7 (42%). And this, I would hazard a guess was a deliberate tactical ploy by the Italian manager, based on wider trends for underdog teams.
Leicester’s primary strengths were the ability to exploit turnovers and then enjoy open, end to end games of the kind that are typical in the Premiership. Their scoring returns flagged up goals scored via through-balls and assists from wide areas. They had a consistent game plan: cross and shoot often, play dynamically and at high tempo, turnover possession. They used the same XI as much as possible.
They were not especially good if asked to break down opponents who were equally disciplined, resolute and unambitious in their approach.
Then as now, Leicester’s primary weaknesses are physical ones: defending set-pieces, challenging in the air, retaining possession through individual strength under pressure.
To beat them it made sense to slow down the tempo of the game, possibly sit with a deep back line and compressed midfield and give Leicester a lot of the ball – look to play them at their own game. Deny them space to break into. But there opponents only came to that realisation too little, too late.
With that as a starting point, there is a lot that a Moneyball approach might help with once the profile of the required reinforcements has been established for a project like Leicester City. So, before a ball has been kicked and the Moneyball filters applied, the manager and football staff have to first decide upon what system to use and then how it is to be enacted.
Leicester settled upon a no frills 4–4–2 with Albrighton and Mahrez as wingers who will bomb forward supporting fast breaks.
It goes without saying: Mahrez’s game will be negatively impacted whenever he is burdened with defensive work.
The speed merchant Vardy needs a supporting striker who can distract defenders and create gaps for the main man to burst into.
At the same time Leicester cannot play high defensive line as the key to their game plan is winning the ball back where the opponents have overcommitted men ahead of the ball. So Leicester have had to sit back in and hit on the counter, often from the edge of their own box.
This means, in recruitment terms, that Leicester would need to build their game around one type of player in particular. That is a player they would need to buy in. The key man is a midfielder who has both tremendous explosive energy and the stamina to cover acres of ground in front of the back four for 90+ minutes – games after game after game.
Now with this primary target identified, the next part of the process is identifying appropriate subsidiary players who might fit the bill.
Apply stats to find a player
Leicester’s title winners also needed a dominant centre half, a player who can win headers from crosses and a man to be a strong partner for Wes Morgan. Despite his limitations, Robert Huth is perfect for the role. Ex-Chelsea man Huth isn’t athletic or a great reader of the game. But as he showed under Tony Pulis’ management, he can win things in the air and doesn’t shy away from a physical battle. This is exactly what is needed for a compact, structured defence that allows the opposition to have possession and whose only remit is to clear its lines at all times and start attacks with a short, simple pass.
Remember we spoke about Leicester needing a midfielder who is energetic first and foremost as they don’t want to burden Mahrez with defensive duties?
Well, now the search begins. N’Golo Kante is that key player who dominated the stats chart as the most athletic of players in the Premier League as Leicester drove on to the title. Kante played a total of 37 games and missed just one through suspension en route to a title gong, international recognition and a megabucks contract at Chelsea in the summer.
Kante has other attributes than legs and a nose for emerging danger. His tackling and interceptions would have caught the eye in videos of his games at Caen. Kante isn’t much of a goal scorer or a creator though – just one goal in 75 games for the club in Lower Normandy’s historic capital. He is no thwarted No. 10. But neither Ranieri nor indeed Antonio Conte require goals, assists and flamboyant passing from the French Malian.
That creative spark is reliant upon Mahrez ,Vardy and his supporting striker and the other wide man Albrighton. And that is where the goals and assists emanate from in a simple solution for Leicester.
Moneyball and its evolution
Scouts have an amazing eye for talent and are if not well trained formally, then steeped in the game and its nuances and details. But the Moneyball theory goes that inevitably these self-same scouts are also products of all too human bias. What they really need according to the Moneyball credo is the checks and balances of an objective measure of outputs. That is they need to be told what players ‘actually’ do as opposed to what the scout’s perception sees them doing.
Michael Lewis’ Moneyball: The art of winning an unfair game is an excellent read irrespective of whether you are baseball fan or football fan, and it has certainly allowed team sports to be seen in a new light, and with a new supporting industry of number-crunching MBAs. In truth it is something of a cult, and a fundamentalist cult too.
Billy Beane an ex-baseball player turned scout, turned sabremetrics guru, is the messiah and also the inspiration behind Lewis’ book and an Oscar-nominated film from 2011 starring Brad Pitt (as Beane).
And there is no denying, Beane has an excellent CV based on notable success as a Major League Baseball General Manager and in his broader influence on American sports.
In common with so-called old school scouts – in football and other sports – Beane believes there are certain aspects of the game that can be developed in a player and others that cannot.
Characteristics like strength, endurance or dribbling or passing range can all be worked upon but there are certain aspects of the ‘mental’ game that cannot be developed at the same rate as others such as patience amidst the chaos (effective distribution), calm decision-making, creativity, decisiveness and so on.
Beane in his turn, is an acolyte of the sabermetrician and Baseball writer Bill James.
Sabermetrics is a way to provide reports using statistical analysis of a player or a game’s significant events. Beane’s big idea is to play Moneyball – to use stats to identify cost-effective player targets who were already developed in the parts of the game that he required to build a cohesive team unit.
In football the name of the game is that you need to score more than the opponent to win. But how you get there encompasses an almost infinite number of individual data point per game. That’s data points around which a narrative must be woven to tell a story of cause and effect, outputs and outcomes.
To arrive there, at a practical level, winning games, managers implement different systems they believe will create team cohesion and invest in people – their personnel.
Since the release of the book in 2003 a lot has changed. Technology has evolved so much that it is now possible to connect videos to statistics and generate an entirely new style of scouting report. Such is the rate of travel, that new ways of making calculations are brought forward every day in what is a sporting equivalent of the old Cold War space race. This time the prize is total domination in a chosen sport by way of a secret code, a new ‘way of seeing’. Though the concept of Moneyball remains the same, the battery of weapons it can draw on in the form of drilled-down numbers grows ever greater. Fine you might say, and I do. But there is always the nagging doubt that this kind of sterile, academic big data exercise is not a natural fit for football, a game too dynamic for large or small datasets and a game naturally resistant to presumptuous form analyses.
Moneyball and football
Moneyball, the original concept which is used in baseball cannot be directly applied to football.
This is because football is fluid, human, obscure in the meaning of its outputs, low scoring and very, very unpredictable.
By way of an example, a stylish central defender might boast a 90% passing accuracy but is this really especially relevant as a stat when what’s required is a possession based game tomorrow against a high octane, high pressing premier league side that won’t let him rest for a second? That passing stat encompasses all sorts of inputs – time on the ball, the blend within a back four, not to mention the quality, distance and tempo of those said 90% successful passes.
As such there simply must be caveats, various parameters applied to each definition, to each number. In programming, we call this method overloading. For football, as I understand it, this overloading needs to be applied in a serious and scrupulous way if the data is to be anything other than useless at best and specious at worst – robbing good, brave men of the testimony of their eyes due to ‘the weight of numbers’.
Moneyball can’t be the be all and the end all. It can only ever be part of the picture and even then, not without qualification.
Moneyball probably works best as a sense check to a scouts opinions – and vice versa. Though scouts can continue to flourish alone (as they have always done) the path of fully automated scouting and recruitment could be a road to hell driven by corporate hubris on the one hand and the historic tension between suits and football people that is an eternal battleground within every single football club, throughout the world, on the other.
Why use Moneyball?
Moneyball cannot guarantee success. It cannot make you a champion. But it can inform more rounded recruitment decision-making. Moneyball’s USP is its promise of a ‘slight field levelling’ for clubs with less resources, without the wherewithal to compete.
It’s credo will not have won The Premier League for Leicester, nor could it.
But, in the highlighting of players such as Kante and Mahrez it may have allowed the club to recruit the right specialists – players specifically suited to the manager’s purposes as a tactician and team-builder.
What Moneyball can’t do is organise teams on the field of play, it can’t reveal the depths of players hearts and minds and personalities. It can’t bind them to a common cause. No spreadsheet derivative can do that.
And football at its very heart is a game of human beings, a game of hearts, minds, blood, sweat and tears and it is sparked by social interaction.
At another level, Moneyball promises a better allocation of resources: the right players, in the right positions, at the right time and at the right price.
And that is something that, if it works, really can help football clubs punch above their weight.
It’s power if it has any is in generating scouting lists of players for the in-house experts to run the rule over, do their background checks on, and pass judgement on their suitability for purpose.
Like most innovations, Moneyball is a labour-saving and potentially money saving tool – specifically for recruitment, in this case. But while it might assist in the avoidance of costly transfer mistakes it can’t put a winning team of players on the park. That is an act of charisma – something only football expertise, talented people can enact.
Determining undervalued stocks using statistical analysis is something that is highly tried by economists with little or no success. It is a Holy Grail as such, and it would be a passport to inevitable riches for everyone in on the secret. If it could be found.
The footprint generated by the Moneyball project is huge and is a subject of endless learned research, bitter dispute and no-end of swanky international conferences. That is to say, it is a subject, akin to an article of faith, something that both raises debates and divides opinion.
It is also of course, a subject that can be easily interpreted from differing standpoints, to serve different ends and masters. As such, the so-called objective standard, that the concept was originally ‘sold’ as, appears to be shrinking from view. The one-time panacea championed by the likes of Damien Comolli at Spurs and Liverpool, has become just as slippery, just as partial as the scouts’ hard-won perspective that Moneyball’s most militant disciples would love to fatally discredit.