As if the Premier League weren’t popular enough as a product on its own, millions of people also play free Fantasy Premier League (FPL) games on various platforms, including the league’s own. Since we’re offering resources for fantasy players on smarterscout.com, we thought it was time to get involved directly. So this season, for the first time, we have a team – and a special strategy to go with it.
Unfortunately, we didn’t have time to write a complete algorithm to optimize our team. When we do, we’ll make the results available to everyone on our site. But for this season, we just ran a few regressions using points totals from the past couple of seasons to see which individual metrics – the very same ones you can see for free on our site – seemed to predict FPL points in the next season.
For simplicity, we didn’t use any team metrics, though we have a lot of good ones in the “Fantasy” section on our site; we just stuck to individual metrics. But we did use individual metrics from other leagues adjusted to a Premier League standard, so we could also predict points for players who had transferred into the Premier League this summer. And then we went down our ranked lists of players at each position, trying to fill our squad with the most predicted points.
Minutes played was the most predictive metric carried on our site. That’s not surprising, since players can’t score points unless they’re on the pitch. And for peak-age players, minutes tend to be pretty persistent from one season to the next. Among the other important metrics were finishing and shot-stopping skill for non-headers in open play, attacking output, involvement in moves leading to a goal, and, for defenders, skill in ground duels out of possession.
All that said, team selection – as you’ll find out below – wasn’t a very important part of our strategy. We didn’t even play around with different formations. We just tried to wedge the top players on our lists into a 4-4-2. For our starting XI, we avoided players on promoted clubs (though some of them were quite high on our list!) and stuck to players who looked likely to be starters for most of the season. So here they are:
Our team is what fantasy players sometimes refer to as a “zombie”, because we’re not planning to make any changes to it unless a member of the starting XI suffers a season-ending injury or is transferred out of the league. We used the Triple Captain chip right away when selecting the starting XI. (Giving it to Harry Kane proved to be a pretty good decision, as he scored twice and racked up 39 points!) We’ll save our Bench Boost for the last Gameweek, so that we have the maximum time with no chips used. And we won’t use any of the other chips.
Why would we handcuff ourselves this way? Well, it’s an experiment. We want to gauge the overall benefit of actively managing a squad versus doing almost nothing.
In finance, there’s a longstanding debate – and plenty of research as well – on the question of whether actively managed investment funds outperform passive funds that just follow an index of the stock market. The same question applies to individual investors: day traders versus those who buy and hold. Because of the cost of trading – commissions, time invested in research, etc – and the difficulty of predicting share prices, the consensus is that passive management is usually better, on average, over long time periods.
FPL also has a cost to active management, in the form of the four-point hit to swap out a player. So will passive managers also outperform active ones in FPL? That’s what we hope to find out, by tracking our team’s rank from week to week.
Assessing the results
After Gameweek 1’s bonus points were added, our rank was 750,059. That’s nothing to write home about, but we don’t care! The result of our experiment depends not on our rank itself, but rather on whether our rank changes significantly as the season wears on.
All other things equal, our rank should probably drop. Latecomer players will enter the game, and some will perform better than we do. Also, we’ve used our Triple Captain chip already, and we’ll be left behind by some players who use it later on. But if our rank were eventually to rise, it would suggest that either we were extremely unlucky early in the season or that passive management may actually be superior.
Of course, we would have to enter a whole raft of passively managed teams into the league in order to conduct a more complete experiment. We don’t have enough time for that, alas. Yet it should still be fun to track our rank, and we’ll keep everyone updated on our Twitter feed.
A final thought
One thing is clear to us after just one Gameweek. From a marketing standpoint, fantasy is a great way to increase fans’ engagement. Even if you support Manchester United, you might have Mohamed Salah in your FPL squad… and you’ll be more likely to watch a Liverpool match. Multiply that by XI, and you might be rooting for and following a lot more players than usual. No wonder the Premier League offer their game for free.
We hope the metrics on smarterscout.com will help you to keep track of all those players from different clubs. And if you’re playing fantasy in a different league, there’s great news. We offer Fantasy metrics for every player in our database, covering 40 leagues around the world. Just hop on over to smarterscout.com to check them out – it’s free!