planetfpl
patreon
đź—‚ Simplifying FPL
🕑 Added 2022-08-02 21:33:41 +0000 UTCWritten by Advanced Tier Patreon Mini League Winner Cole Bonzer
We’re all chasing the same goal in this game: to score more points than everyone else. What makes FPL so fun to all of us is that there’s more than one way to succeed, and there is no secret to sustained success for everyone. Each person has to find their own path, and what works for one person won’t work for all.
I should have no advantages to picking a better set of eleven players to score as many points as possible. The vast majority of even somewhat serious FPL managers have watched football for longer than me. You all have superior historical knowledge. You all have seen players come and go, and teams rise and fall. So the odds have never been in my favor to achieve a very favorable overall rank, but through spending a massive amount of time watching the game and analyzing the aspects of FPL itself (and large amounts of luck), I have had what I’d call 3 successful seasons out of 4.
In order to do this, one of the things I’ve made an attempt to do is look at FPL on the whole to see if there’s any edge I could get that other managers overlook. One of my favorite movies is Moneyball, and Jonah Hill’s character in the film says “... I see an imperfect understanding of where runs come from.” in reference to a particular baseball player. With that mindset, I set out to explore if there’s an imperfect understanding in the community of where FPL points come from. Going into this task, I was completely aware that FPL in many ways is extremely over-analyzed. So I wasn’t about to come up with some unhidden gem of a statistic, nor some way of using the eye test that isn’t already common knowledge in the FPL community. I was looking for something simple that everyone sees, or is capable of seeing, but maybe not fully appreciated within the community.
So starting from the ground up, what action gets the most FPL points? Goals and clean sheets (will focus on the former here). Goalscoring defenders are amazing, and I chase after them like a lot of people every season (looking at you Marcos). However, they don’t score nearly enough to make a meaningful difference across a whole season. Midfielders get 5 points per goal and 1 possible clean sheet point, and forwards get 4 points per goal and no clean sheet point. Among other reasons, this is how De Bruyne can outscore Kane 196-192 over the season when Kane had more goals and assists all while playing 1000 more minutes than De Bruyne. When a midfielder and forward have similar goal scoring numbers, the midfielder will have a very clear edge. Putting that on hold for a second, but this creates another question: Is there any correlation between actual goals and any statistic I can look at?
When sorting by different types of offensive statistics, two statistics to me show the strongest correlation to actual goals scored: shots in the box, and big chances. Through analyzing these two statistics for the past two seasons, I have grown a slight favoritism towards shots in the box rather than big chances. I haven’t always agreed what determines a big chance versus a regular chance, but shots in the box as a statistic is concrete. There’s no room for interpretation. xG, xGI, and the related statistics do show strong links to goals, but all of these statistics are very closely related anyways so that is to be expected.
Now why shots inside the box rather than total shots? Even the most prolific finishers like Kevin De Bruyne and James Ward-Prowse only had 5 goals from outside the box last season, with no one bettering the total. But these are outliers, not the norm. It’s no surprise to see that Mo Salah led midfielders with shot attempts in the box last season by a large amount, so it would never be surprising to see that he’s the top point scorer in the game. A sampling of that list after Salah includes Mane, Saka, Bowen, Son, and Zaha. All of them had excellent seasons for their clubs, and except for Zaha they were among the top 8 midfielders for total points last season.
So how did I apply this very simple analysis to the complicated act of managing an FPL team? I started catering my decisions towards who was looking the most likely to get the most shots in the box, and paid less attention to the more solely creative midfielders, or players who don’t go quite as far forward consistently. But doing this isn’t as simple as looking for whoever had the most shots in the box, job done. Bryan Mbeumo had 56 shot attempts in the box last season while only scoring 4 goals. If one were to base FPL decisions solely on shots in the box, they would be led to believe Mbeumo would be an incredible asset when he simply was not. But looking at shots in the box is only one individual metric for predicting goals scored.
Another simple factor that impacts goals/shots in the box for a particular player is what team you play for. Manchester City scored 99 goals last season, more than double the goal tally of half the league. They are able to do this through creating shooting opportunities in good positions against all of the teams in the league. Whereas a team such as Wolves aren’t likely to create a large amount of chances in every match they play. Even Heung-Min Son, who was arguably the most devastating finisher in the league last season had a goal conversion of 26.7%. So, at the very least he needed 4 shots to score 1 goal on average. If Son played for a team like Wolves who don’t create a large number of shots in the box, it might take Son 2-4 matches to get the 4 shots on average to score. But at Tottenham, Son gets more chances to score per game, and thus he is capable of sharing the Golden Boot.
Combine these simple observations, and you get that the ideal FPL asset is a midfielder that has a high number of shots in the box that plays on a good team. We all know that Salah has been the best FPL asset for years now, and he fits this description perfectly. So this gave me confidence to put a larger amount of my 100m+ total into midfielders last season, and the simplified way of looking at FPL worked out well in the end.
Now, one might read this and think that my way of playing the game is more stats based rather than based on the eye test. But for me, that’s simply not true. The goal of searching for the statistic that best predicts goals was not so that I could look at that table every week and make transfer decisions based off of that table. The goal was to find a statistic that could better inform my eye test. The statistics that every manager looks at will only take you so far, as they can identify players who on paper are capable of consistent returns, but past results don’t guarantee future performance. There are a lot of players in the Premier League who are fully capable of having a great run of form over 6-12 weeks, and you definitely can’t have them all. So for me, watching as much football as I can is the best way to be informed on who is most likely to do well in the future. And at the end of the day, that’s what all of us playing FPL are trying to do.
Comments
itsrosco
Great insight Cole!
Stuart Oxlade
Love Moneyball. I've been meaning to do some similar analysis myself but rarely find the time. Great post!
