Workforce Management Lessons from Fantasy Football
Author: Israel Beniaminy
Some people actually get to play fantasy football during work hours. In this case, I’m talking about the British fantasy football league, which our American friends will know as Soccer. The lucky people who not only get to play the game but also to openly tell everyone about it are scientists from the University of Southampton, who developed an impressive Artificial Intelligence (AI) software program for playing the game.
How impressive is it? The researchers, Dr Sarvapali Ramchurn, Lecturer in Computer Science, PhD student Tim Matthews, and George Chalkiadakis, visiting researcher at the University of Southampton, experimented to see how well their software would have performed in the previous season. Very well indeed: It consistently performed in the top 1% of the 2.5 million players in the official English Fantasy Football League, run by the Barclays Premier League. At one point, it placed within the top 500 players (top 0.02%).
A brief (and unavoidably oversimplified) review is in order: Fantasy football is a highly popular game in which the players create an imaginary football team, by choosing from among real life footballers. Every week, each player’s team is scored according to the performance of the team members in actual game play, taking into account football statistics such as scoring goals or making assists. As you create your team, you need to assign real-life footballers according to their roles, so that your team includes a goalkeeper, defenders, midfielders and forwards. Naturally you’d like to select the best footballer for every role, but equally naturally the game imposes some limitations. The chief limitation is total budget allowed per team (that’s fantasy budget, of course).
Regardless of the game rules and systems (and there are many similar games with many variations), good players predict how well each footballer will do, and use their predictions to form a team, within budget, in which each selected footballer’s strengths contribute to the team’s overall score, while the weaknesses are covered by strengths of their other selections.
Does this remind you of something? If you’re a manager or team leader, you have already experienced this challenge many times. Whether forming work crews, or hiring new people, of deciding whom you can bench (say, for vacation or for training), you have to form the best team you can, assigning each team member to one of the roles you need, and doing it all within budget. Some of these decisions are relatively long-term and you can take your time considering all the possibilities, as in hiring. Other decisions need to be made many times every day, such as crew selection and approving vacations. Wouldn’t you want some intelligent software to help you do that?
Let’s take one objection out of the way: isn’t it an important part of the manager’s job to do exactly this? If so, is such AI software destined to replace the manager? That’s highly unlikely. Here’s what the Southampton researchers have to say on the human element: “Our previous tests have shown that a machine working on its own will perform better than millions of humans. But a machine can’t take into consideration if a player is injured (and still plays), has low morale or has personal issues and may not perform at his best.” For this reason, they decided to add a human element: “we will be using humans and the machine working together as a team so that the humans can add this subtle information into the system and, together with the software’s extensive analysis, it will hopefully improve the machine’s success rate.” In other words, let humans deal with the personal issues. In fact, they’ll have more time for human interaction once large parts of the decision process are automated.
Soon, players of British fantasy football will be able to use this software as a web app, either playing against it or using it as an adviser. Players of other fantasy sports games won’t be left behind: See for example the DraftOpt app created for American football, supporting draft-day decisions.
And what about the serious game of managing your workforce by creating and managing the best teams? That’s one of the challenges we at ClickSoftware developed our software for. What is your experience with this challenge? What would be on top of your wish list for software supporting your team management solutions?