“you need to create the storyboard of information” – barry mcneillBack
Football fans will always argue for their sport’s exceptionalism; but then, football fans will argue about almost anything.
That’s something that lies at the very core of what does make football different – everything, every single thing, is up for debate.
There isn’t a fan, player, manager, chairman or referee who will ever see a game the same way as another. It must be the only televised event that is watched by millions, none of whom can agree about what they just watched happen.
“I think there was some research done recently which suggested that a coach can only remember a maximum of 52% of events in a game,” Barry McNeill told me. “And therefore if they’re basing pretty critical decisions about performance on the game, then they’re basing them on half the facts.”
Those decisions have only become more critical as the game has filled up with television money, and in such a highly scrutinised environment, where the stakes are so large, anything that can help the decision making process is valuable.
That’s how come Prozone, founded in Leeds in 1998, has become so integral to modern football. It’s now so inherent, and fans are so used to possession stats, heat maps and scoring charts, that Prozone is almost a shorthand for all football stats and data.
Barry McNeill joined Prozone in 2000, first as a consultant introducing the system to Premier League teams and developing it with their input, and became executive director before leaving the company in 2013.
“Serendipity definitely played a part,” he says. “I had played for Lancashire Boys, but never thought I could convert my passion for sports and football into a career.
“I went down the academic route, undergraduate and post-graduate, and got a job working for the NHS as a researcher – that was my first foray into managing data. A friend that I played football with came out of Liverpool John Moores University with a sports science degree, and he was approached by a startup in Leeds called Prozone.
“Prozone required some sports scientists to help them convert a technology idea – of a 2D animation of a football match as it happened so you could see the movement of every player at any one moment in the game – and take that to football coaches and help them utilise it.
“I had a coaching background, so I had my coaching badges in place, and I had a bit of experience managing data, and I just thought it was a huge opportunity to take a product to market and do something new. A great test of that was that my parents didn’t quite understand it, which for me was a good measure of something that was exciting and innovative and future proof, versus something that was safe.”
Prozone was adopted first by the coaching staff at Derby County, including future England boss Steve McClaren, who took the idea with him when he was appointed assistant at Manchester United.
“We’d install a series of cameras in a stadium, and they’d be fixed on a portion of the pitch; we would then track a player moving from zone to zone, and that would knit together in the technology to give us a player’s movement profile for every 1/10th of a second for the whole game.
“We could link the back four together, for example, and watch how they moved in operation with ball events – about 2,500 events per game, passes, shots, throw-ins etc – that, combined with the movement profile, gave the coach an animation of the game with a video for reference in the corner.”
From early resistance – one manager on his way to a team meeting turned around and left when he saw Barry was there to make a presentation – the application of data analysis to football, which by its very nature has always seemed mysterious and impervious to analysis, has become commonplace. The challenge remains, though, of making sure that the data is used and analysed properly, something Barry thinks still doesn’t happen as often as it should.
“The minute a scientist produces a document that is static, and hands it to a player or a manager to then communicate to another player or manager, you’ve lost the value,” he says. “That’s where context is lost.
“When people are confronted with new technology and certainly new data, they’re quite lazy at using it effectively, and volume becomes the focus rather than quality, because quality is more challenging.
“The great example is the idea that 12 passes are better than 11, or that running 12km is better than running 11km. My job as a consultant was to say the coach that no, we don’t put a list of top ten passers up on the changing room wall.
“Instead I’d sit down at Aston Villa with David Ginola and the coach and we’d look at his passes. His passes would be on Prozone’s tool, a pass map with arrows everywhere. But the minute you hover over a pass you can see the second and third ball, you can click on the pass and watch the animation, so you can see his options, and then you can watch the video. That’s where the data is best used – when it’s purely contextual.”
Often, and especially given the increasing profile of stats websites that produce graphics for Twitter and forums, that contextual aspect, the interpretation of data with regard to the player and the scenario, is lost beneath pass completion rates and counts of tackles and headers won, regardless of their type, quality or outcome.
“I think in the media space, consumers – fans – are getting static data that is a bit meaningless – a lot meaningless. And unfortunately it’s misleading, like any data. I think number of passes, number of crosses in a game… like in any business, if you take any one isolated piece of information and form a judgement on it, there are huge dangers there.
“If broadcasters and media channels and distributors are trying to say that 74 passes for Xavi vs 63 for Gerrard is any different, or that anything can be read into it, then that’s false, that’s wrong. You need to be able to interact with the information, and create the storyboard of that information.”
Understanding the limits and capabilities of technology that, twenty years ago, didn’t even exist, has been part of Barry’s journey with Prozone, and part of the challenge of making pure data a meaningful part of football.
“I had a great experience with Paul Jewell when he was at Wigan and they had just got promoted to the Premier League. Leighton Baines was the left back, and we used to have benchmarking data for the Premier League and the Championship that showed the biggest difference between a Premier League team and a Championship team was the left back position. Fewer passes, fewer balls received, fewer final third entries, less high intensity distance covered, and a few other key performance indicators. The documents and the information that I was showing to Paul suggested that he maybe needed to look at replacing his full backs.
“And Paul made a really smart point. He suggested that all the data showed was what Leighton Baines had done; but not what he was capable of. He said that he had wanted to play a certain style of football in the Championship, to get the ball forward to where they had Jason Roberts and Nathan Ellington scoring for fun; but he had no hesitation in believing that Leighton Baines could play a different style of football in the Premier League. And obviously Baines has gone on to be hopefully playing in the World Cup in the summer at left back.
“So that for me was a great example of where the fusion of the data and the coaching process, and the insight that the coach as an expert brings, is powerful.”
Football as the ‘game of opinions’ has traditionally seemed opposed to a data tool like Prozone, that by supplying pure data has threatened to nullify the power of those opinions. What Barry passionately believes, after more than a decade of using and improving Prozone, is that the data needs the opinion and the debate to thrive.
“I think consumers need to be heard, and fans need to be heard at the moment. There are a couple of projects I’m working on that will transition stats from lots and lots of noise, to some meaningful insights into how we measure luck, how we create one ‘top-trumps’ metric for a player, all with very very strong algorithmic foundations. That’s a project called Scorem.co.uk, a live polling tool that will be about trying to create this fusion of a metric, a Prozone or an Opta score, and a fan score, and start a debate and a discussion about the discrepancy.”
The future holds more fan engagement with data, and it also holds more data for the fans to engage with. “Imagine if England get to the penalty shoot out in the semi-final of the World Cup,” says Barry, “and Steven Gerrard walks up to take the fifth penalty. If he scores we’re through, and there’s a heart rate monitor on him, with maybe some neural pathway activity – and imagine that was your second screen application. We’d all watch it, wouldn’t we? The technology exists – so I don’t think it’s that far away.
“So I think the future is pretty exciting. There will obviously be more data, so I think therefore there is an arms race to make more meaning from the data, rather than in data provision. I’m looking to try and get one club to work with – Leeds would be my preference, because it’s where I live – and to help them connect those two areas: data, and how the coach and players improve performance. That is really where I want to spend my time.”
Originally published in The City Talking: Leeds issue 11