How we have the best scouting system and best medical record in England.

[KamikazeBlue 337]

Big Data and development performance systems at Chelsea

An interesting read

Source:bigdatainsightgroup.com

There has been a growing fascination in data among football fans in recent times, so much so that it is becoming somewhat of a sexy subject for the next generation of tweeting and blogging supporters. Companies such as Squawka and Opta have only helped to bring consumer friendly, real-time data analytics and visualisations further into the mainstream, too.

It’s not just the arm chair pundits who are eager to glean new insights from clever statistical analysis either; many of the world’s leading football clubs are also realising the value that data can offer to the way they run their clubs and manage their players.

The work the ‘development performance systems’ department, of which Ben Smith is a part, delivers at Chelsea FC demonstrates the varied benefits provided by the intelligent use of data. Moreover, while some business people may scoff at what inspiration they can take from a football club, there is plenty to be learned from the way Chelsea collects, analyses and feeds the insights from data back into its training and match operations.

Anyone who pursues even a parting interest in football will know that when in June 2003 the Russian billionaire Roman Abramovich bought Chelsea FC, the south-west London club underwent a significant transformation. In the intervening nine and a half years since the takeover, there has been an increased emphasis placed on performance – something big data plays an integral part in.

“When Mr Abramovich initially took over there was an understandable priority on the first team,” Smith says. “Shortly after the focus was expanded to include developing our own 1st team players and the club began putting a lot of focus on improving the quality of the academy.”

For Smith, the reasoning behind applying big data analytics to help in the process of monitoring and improving player performance and progression is clear. He states: “The professionalisation of sport has been dramatic over recent years and it’s only going to continue. There’s a huge amount of money and drive within the industry today; the rewards are massive for those getting things right and they’re substantial for getting it wrong – data analytics helps us ensure we do the former and avoid the latter.”

Since embarking on his role at Chelsea, Smith notes that “there has been a massive influx in data’s involvement in the sport. I think it’s already providing an opportunity to go into incredible depth in really interesting areas. But it’s very important from our perspective to stay focused on what we’re trying to achieve and not getting lost in a wealth of irrelevant data.”

Tackling big data

There are two main types of data put to use in player development. The first of these is what Smith refers to as ‘objective data’. This comprises of physical measurements which track the movements and conditions of each player – such as heart rate and GPS – during training sessions and matches. This data set is generated through a variety of sensors worn by the players wirelessly transmitting masses of data every second. The second type of data the club collects and uses is ‘qualitative data’. This is made up of staff assessments on each player, their strengths and weaknesses, and how they feel they’re performing and progressing.

The big data elements of the work come from the unique challenges these two different data sets present. On the one hand they must be able to collect, store and analyse the massive amount of objective data being generated by all the players involved in a training session or match. Combined with this there is the added challenge the qualitative data presents which, although much smaller, comes in varied and complex forms such as text, making it more difficult to analyse and report on.

Furthermore, timing is a critical aspect of the task of turning all this data into something insightful and valuable for Chelsea's staff. Smith says: “After every training session or match we have around an hour turnaround time on our data to put it into an engaging report that will tell that team’s coach: ‘this is your key information about the individual and team’.” These reports, like those in any business, are vital to the process of reviewing the successes and failures of ongoing objectives, thus enabling any problems are addressed moving forwards.

“Real-time is always our preference,” Smith continues, “and if we can’t do that then we want to be as quick as possible. Often there is a small window of opportunity from when the players leave the field in which staff want the information, before their attention moves onto the next task. We always try to be prepared for when a coach wants to engage with the information – it has to suit and work around them.”

To help them in their goal of delivering as close to real-time insights as they can, the development performance systems team performs all the management of their data onsite. “The hardware,” he explains, “is based at our training facilities at Cobham [in Surrey] as well as at the club’s stadium Stamford Bridge [located just off west London’s prestigious King’s Road]. There is also an in-house IT team which supports performance systems analysts.”

And while the analysis is largely done on-site too, specialist consultants around the world are also called upon, particularly for the analysis of the objective data. “To simplify the process of using external data specialists,” Smith continues, “we have automated the process, via a ‘black box’ in our servers, of getting the data from the games or training sessions and sending it immediately for expert analysis. Their results, reports and insights are then fed back in to us as soon as they’re completed.”

Turning football coaches into fans of data

Beyond the complexity and time constraints placed on the analysis, another major obstacle faced in the job – like that faced by so many people entrusted with big data within an organisation, football club or otherwise – is to make data useful, accessible and engaging to colleagues who have little interest or experience in dealing with numbers. In the context of football clubs and their coaches, however, this issue probably becomes somewhat exacerbated.

As already mentioned, delivering performance reports from training sessions and matches promptly is one important way of making it useful to the coaching staff; for it to be beneficial to them they need it in their period of reflection, not when future plans and preparations have already begun.

One key tool used at Chelsea to make big data analytics valuable is visualisation software. Smith says: “Visualisation is a critical issue in football because you’re dealing with people who don’t typically have experience in data and often are not particularly passionate about it – you have to bring the information to life.

“Numbers are really, really dry and people from a coaching background, even the modern coaches, are not often data driven. If you can present the numbers in a way that means they quickly understand its direct relevance to the things they’re trying to achieve then they will appreciate the significance of what it’s telling them.

“We’ve got a fantastic set of staff at the club, really progressive and great coaches, but everyone has their own characteristics. It’s about understanding the ways which are going to deliver something meaningful to that individual, and data visualisations are really important because if you give our coaches a series of numbers it doesn’t have a huge amount of significance to them. But when you start to put it in a way which is aligned with the way their coaching brain works then it has a much stronger impact.”

Through comprehensive reports on each player, which marry together the objective data about their physical conditioning and positioning on the pitch with the subjective reports from the expert coaches, a holistic picture can be built up for each player over time.

For the coaches at Chelsea FC, Smith says, big data “tells part of the story in incredible depth. However, it’s a very complicated picture when you look at the performance and progression of a player and that’s ultimately what ensures the coaches have the final say on what we do; they can bring together all the factors, some of which you can’t measure.”

In much the same way, getting insights from data can also play an important supporting role for scouting teams. Smith explains that although you would never sign a player based purely on stats and facts, it can narrow down targets and effectively help the club “decide how to deploy a scouting networking to get the most out of its available time”.

Blood, sweat and tears... and data

But it is not, of course, just the coaching and scouting staff that benefit from the big data analytics being carried out at the club, the players are also reaping the rewards of the work across the club. He says: “Every one of Chelsea’s Academy players from the age of nine has a personalised development programme.

“It’s really important to look at and treat our players as individuals so that they’re managed in the appropriate way to match their specific needs, whether that’s physically, mentally, medically or technically. If we treat all the players the same then we will deliver a successful programme to a few but be wide of the mark for many.”

In addition, by allowing players to see an objective breakdown of their own performance – including things like key decision outcomes, how they performed their positional responsibilities, and physical conditioning throughout – they gain a much broader understanding of their own game and how they can improve in the future. ‘Knowledge is power’, as they say, and here data can empower the players in their quest to improve and develop into the stars of tomorrow.

When it comes to managing the physical aspect of the game, big data is extremely valuable too. Smith says: “All the players go through different rates of physical development which means that understanding the physical exertions that are placed on their body is extremely important. The objective data we collect and analyse allows us to understand their physical development and, more importantly, it tells us what we need to do in the short, medium and long term to that player’s programme to maximise their training but also to minimise their risk of injury.”

Indeed injury prevention and rehabilitation is one area of football in which Smith believes big data can be “exceptionally useful”. He explains: “By using data we can reliably predict when someone is entering a ‘high risk injury zone’. And because we know how to manage the different demands of match and training scenarios, we can lower their workload or alter their training schedule accordingly. And then when they’re returning from injury we can understand how far along that recovery path they actually are, as opposed to simply assessing ‘do they look ok on the eye?’

“It’s all about the quality and depth of data; when you’re doing preventative and rehabilitative work regarding injuries, the depth of data you get surpasses what you can do with the naked eye.”

What other businesses can learn

When asked what he thinks the business world can learn from the data-based work taking place at Chelsea FC, Smith modestly states: “It’s difficult for me to answer having had very little experience outside of football.” He does, however, explain the key mantras they stick to which help them turn data into a valuable asset at the club. And these dos and don’ts certainly ring true off the football pitch and in the business environment.

The first piece of advice Smith has is to ensure that data doesn’t completely upset an organisation’s staff or processes. Although big data is a disruptive force which can transform an organisation, it is still important to maintain the principles and ethos that are critical to the character and success of your organisation. He also stresses the importance of aligning any big data projects with your business objectives so that time and manpower is allocated as effectively as possible in order to help achieve the organisation’s bigger goals.

In his own words: “It’s really important for us to view data and analytics in a supporting role within the club. We have very clear objectives of what we need to do and we have to understand how data can support those things.

“Everyone has to have a very clear intention of why we’re doing something and what we’re ultimately trying to achieve. We then focus on what we need from the data to help everyone in their roles.

“Furthermore,” Smith continues, “it’s incredibly important to know what data you have at your disposal and what you could potentially do with it. When you combine these things to understand the relevance of the data you have to what you’re trying to achieve it simplifies any data project – the task of using the right sets of data to support the decision-making and development processes becomes far more effective and efficient. Of course it also spares you the time of messing around with the wrong or less useful sets of data.”

This is certainly a best practice tip that experts have often shared with Big Data Insight Group. Any big data or advanced analytics project needs to be aligned with the organisation’s business objectives. Moreover, understanding the data you have and then ensuring it is visualised in the right way and delivered to the right people when they need it is a certainly going to be a productive process to underpin any successful big data initiative.

Becoming clinical with big data

The involvement of data within sport is growing all the time. It’s not just football that now realises that it would be foolish not to use big data to uncover things that even the most experienced physios, coaches and managers can’t spot or predict. It is not going to render the Sir Alex Fergusons or Jose Mourinhos of this world obsolete but it is certainly going to prove, and is already proving at many clubs, an invaluable tool to aid player growth and progression behind the scenes.

Indeed, while things like the famous book-turned-film Moneyball may have helped to further popularise the subject and to get fans and onlookers doing number crunching of their own, football clubs are now using data in such an advanced and business focussed way that it rivals any other sector. Chelsea's work behind the scenes pays testament to the value and worth of doing just that.

From creating comprehensive, in-depth player reports which can support or positively challenge a coach’s expert judgement, through to creating highly personalised training regimes for every player to accelerate their development and helping in injury prevention and rehabilitation, big data analytics substantially enhances the work that happens on and off the pitch at the club. In a game when the margins for error are so small, and the rewards for the victor are so great, it is no surprise that big data has already become a marquee signing at Stamford Bridge.

[The Skipper 20,593]

Very interesting read. Thanks.

[kellzfresh 7,229]

That's so long

Ok just finished, was a nice read. Better way of analysis, nice.

[Fernando 6,464]

Fucking great.

I love read like this!

Thank you for bringing it. :blue scalf:

[Tomo 21,733]

A thread titled "Mourinho will ruin Chelsea" by a user who named himself "out with Abramovich" got 6 pages off response while this (one of the best threads on TC for me) get's ignored for hours

Fantastic article mate, our long term plan is ridiciolously good.

[1chelsea 864]

That something good to have our club getting the best out of player and coaches