Crunching fantasy football numbers to become big business

It’s not often that several million people will be concerned over whether Chelsea’s Romelu Lukaku starts a match against Hull City.

However, as the first week of the Premier League season arrives – and with it the dawn of a new fantasy football year – Lukaku's relatively low fantasy league price (£8 million), married to having a "double game week" due to a midweek fixture with Aston Villa, means many will agonise over whether he will usurp the misfiring Fernando Torres at centre forward.

Fantasy sports is a "billion dollar industry" in the US thanks to more than two decades of fantasy baseball, NFL and NBA games, with this side of the world slowly catching up on the idea.

New England-based Soccermetrics chief executive, Howard Hamilton, says the decision by the official Premier League fantasy game to launch a "pay to play" format this season means a number of data analytics companies may now see fantasy football leagues as a "viable route to make money".

Testing ground
Soccermetrics itself specialises in creating algorithms and software to build products such as "player prospect reports" where data is collected on player performance from leagues around the world in order to judge the current and potential worth of each player, with Premier League clubs among its clients.

“We view fantasy sports as a testing ground for some of the algorithms and approaches we want to implement,” says Stanford graduate Hamilton. To create a predictive model which helps fantasy players’ view how their team will perform over a season is difficult though, he adds, due to a lack of access to the “right data”.

Founder of Wisconsin- based fantasy sports data company Rotowire, Peter Schoenke, says that "access to proprietary data is the issue" when it comes to creating a fantasy football version of the predictive analytical tools now commonplace in finance and other areas of business.

“Premier League clubs don’t usually release that data,” adds Schoenke, who notes that while American sports teams previously viewed sharing player information as being part of “a type of open-source project” they too have now become more secretive as the “power of analytics” has grown.

So without a piece of software to help you judge your best team, what’s the best way to go about spending an imaginary £100 million on 15 players for your fantasy squad?

"Believe it or not there's a science to [fantasy football] rather than just a bunch of people sitting around a table in a pub arguing about how much Gareth Barry is worth," says Jon Trigg, MD with Silent Manager.

A UK group which has spent a decade building fantasy leagues across 14 sports for newspapers and major brands, the science employed by Silent Manager begins with predicting final league table positions and dividing the league into five tiers from Champions League certainties to relegation fodder.

“There’s a probability in a team’s performance and that influences each player’s rating,” says Trigg.

“Then you look at that player’s influence within the team. You examine last year’s data and we have six or seven years’ worth of information on a number of players in our Premier League game and you can look at their trends for point scoring.”

Interchangeable
Prof Chris Anderson, of New York's Cornell University, recently co-wrote The Numbers Game: Why Everything You Know About Football is Wrong, a book which attempts to cut through the "decision biases" common among football fans, pundits and even managers.

He notes that many fantasy league players are susceptible to relying on conventional wisdom over the type of proven data Trigg is speaking about.

"Take players coming in from Spain, Italy or Germany, " he says. "There's a lot of credence given to playing games in the Premier League as an indicator of quality. What we find is that across the top four leagues in Europe [player performance] is pretty interchangeable.

“There are tactical differences but in terms of basic outputs by the workers on the shop floor so to speak, it’s very similar,” he adds.

As for picking his own team this year, Anderson says that without any fantasy data tools to hand, “I look forward to being just as wrong as anyone else this season!”

Fantasy picks: Gary Hooper, Wilfried Zaha and Romelu Lukaku
The official Fantasy Football game from the Premier League sees participants handed £100 million to spend on 15 players. You win points based on a player's performance which is ranked using a number of criteria such as goals scored and minutes played.

Interestingly, the official Premier League game has
this year introduced a new bonus points system which rewards a player with three points for a pass completion rate of more than 90 per cent in any game. Of the experts The Irish Times spoke to, here are some of their fantasy picks this season:

Andrew Laird, senior soccer editor, Rotowire: "We like Gary Hooper at Norwich, Wilfred Bony from Swansea and Arouna Kone from Everton, specifically because of the available time on the pitch they'll get."

Jon Trigg, MD, Silent Manager: “Wilfried Zaha at Manchester United. His pro- bability of playing is fairly high, his probability for assists is high and the chance of getting the odd goal is also high.”

Chris Anderson, author The Numbers Game: "Lukaku at Chelsea. If early in the season he gets in the team I can see him having a terrific season."

Steve Watt, Opta Sports: "If you're looking for any new guys on the cheap I really fancy Emanuele Giaccherini at Sunderland, Norwich's Ricky van Wolfswinkel and Nicolas Anelka at West Brom to be great buys. – the latter dependent on whether or not he can be bothered on that particular week!"