FPL a Moneyball Perspective

When tinkering with my GW1 team this preseason, I had trouble finding sub £7.0m midfielders or forwards that I really liked. This made me consider trying a new formation this year, switching the normal 3-4-3/3-5-2 to a 4-4-2/4-3-3. This led to more questions – how many points can I reasonably expect from a defender versus an equally priced midfielder or forward? If I had one spare million, would it be better to upgrade a goalkeeper, defender, midfielder or forward? Is it generally better value to upgrade a £5.5m asset to £6.5m or a £9.0m to £10.0m? As my head kept spinning, it dawned on me that I had little to no idea on how many points I could reasonably expect from a given price point or position. To mitigate this black hole in my FPL knowledge, I pulled out a spreadsheet and started looking at numbers from previous seasons. In other words, this article will not outline a tried and tested strategy but hopefully, you will find some food for thought and might even end up gaining a new perspective on some aspects of FPL.

Method

I started out by loading the numbers from FPL seasons 14/15 through 19/20 into a spreadsheet. For each season, I included all players having played more than 1600 minutes – 1372 players in total. I’ve used the GW1 price as a reference when comparing points to a player’s price. As a higher probability of starts and minutes on the field is part of what you’re paying for with higher-priced players, I have used total points per season instead of per 90 minutes played.

Granted, history shouldn’t be used to predict the future and the key to a great FPL season can’t be derived from calculations in a spreadsheet, but knowing what has happened in the past and which trends that tend to repeat themselves should lead to more informed, and thus better, decisions. At least in theory.

Minutes on the field vs total points

The first hurdle was deciding how many players to include in the analysis. There is no right or wrong, but it is important to be aware of how it affects the outcome. Too few minutes played includes many fringe players and ultimately lowers the average total points, whereas a threshold for minutes played that’s too high results in a small sample size. I compared the total points of players playing on average 60, 70 and 80 minutes a match. Unsurprisingly, total points per season generally increased with playing time. There were of course individual differences between the price points, but on average going from 60 to 80 minutes played per gameweek resulted in an increase of 20 points per season for defenders and midfielders and 30 points per season for forwards. It might not look like much, but across the 10 outfield players in your team, it will make a difference of 200 points a season. This is by no means a discovery worthy of a Nobel prize, but it underlines the importance of picking players that are bonafide starters and remains on the field for the full 90 minutes.

I ended up using 2500 minutes played (on average 65 minutes per gameweek) as the threshold for further analysis, making the sample size 631 players across the 6 FPL seasons.

Value represented as points per million

At first glance, FPL is a simple game; Pick 15 players within a £100m budget and get as many points as possible. But even that simple premise has a couple of caveats. Firstly, £100m is not your active budget per gameweek. Your bench will not score any points and will probably take up about £17m, leaving £83m that can be turned into points. A different perspective would be to say that in order to enter a valid team, you have to spend a minimum of £64m – two goalkeepers and five defenders at £4.0m and five midfielders and three strikers at £4.5m. Granted, that team will not score many points, making how you utilize the remaining £36m all the more important. Each million must be used to gain the most value and value in FPL terms is points, leading to points per million being a key factor when assessing players for your team.

Whenever you try to achieve something, it is always good to have a goal in mind. 2500 total points would have placed you close to the top of the previous FPL seasons and even won a couple, so that will be our goal in this article. 2500 total points equals 66 points per gameweek or 6 points per player in your starting eleven every gameweek. On average. For every blank, you need another player stepping up with a double-digit haul. In terms of points per million, every million you spend of the active £83m must net you 30 points over the course of the season. But none of those millions start repaying you before you spend four of them and by then you’re already 120 points in the hole. FPL doesn’t look that simple anymore.

As players are generally kept for 5-10 gameweeks and then swapped for others with better fixtures, it’s easier to look at what those 30 points per million means in terms of points per gameweek. As a rule of thumb, a player should be able to net you on average 0.8 points per million for each gameweek he’s in your team.

Cheatsheet for how many points players in each price point should provide on average each gameweek.

Achieving 2500 points a season with your FPL team is not easy and neither is it for a player to get 30 points per million or higher. In the six seasons I’ve looked at, 47 players made the cut and Rhiyad Mahrez in Leicester’s title winning campaign was the standout performer, making a mockery of his £5,5m starting price to the tune of 43.6 points per million.

The top FPL assets in terms of points per million over the previous six FPL seasons. 

Mahrez is one of just 11 midfielders or forwards among the 47 and I bet most FPL veterans have fond memories of the other ten. The remaining 36 are goalkeepers and defenders. That certainly makes me think twice the next time I consider a £5.5m midfielder for my starting eleven.

The eleven midfielders and forwards that achieved more than 30 points per million in the previous six FPL seasons.

Lowering the threshold of points per million to 25 closes the gap somewhat, but there are still more than twice as many goalkeepers and defenders (50 and 59) compared to midfielders and forwards (25 and 12). Even if there is about 100 more MID/FWD compared to GKP/DEF within the sample size. Lowering the threshold yet again to 20 points per million and GKP/DEF still has MID/FWD beaten by 257 to 171. 5-3-2 anyone?

What the midfielders and forwards with over 30 points per million have in common, is that none of them were priced above £9,0m going into the season and only Salah was priced above £6,0m. Alexis Sanchez is the player with a price tag of £10,0m or more in GW1 that achieved the highest points per million with 24 in Manchester United … eh … Arsenal in 16/17 (oops, it’s easy to get millions per touch confused with points per million). That’s a number bested by 191 other players over the six seasons. A player priced £10.0m or higher has achieved over 20 points per million only seven times in that timeframe and three of those were named Eden Hazard.

So what does this all mean? Should you shy away from all premium assets in the hunt for value? No, because those are valuable captaincy options and far more likely to reward you big returns. But there might be a case for going with an extra defender in the £5.0m-£6.0m price range instead of a cheap midfielder or forward. The numbers also highlight the importance of rotation, as few players manage to deliver enough value over the whole season. Defenders seem to provide the best value, so increasing the number of defenders could give you two benefits:

• Higher points per million compared to midfielders or forwards.
• Fewer attacking options to rotate, making transfer management easier.

A deeper dive into the numbers for defenders, midfielders and forwards will be the topic of a future article. Speaking of dives, let’s have a look at goalkeepers – probably the most overlooked player in the average FPL team.

Goalkeepers

“Set and forget” and “Put as little money as possible into goalkeepers” are common statements in FPL forums, articles and podcasts. Clean sheets are difficult to predict, but rotating £4.5m defenders is common so why not goalkeepers? Your average cheap defender not named Lundstram doesn’t offer that much attacking potential anyway. Goalkeepers do have the possibility of offsetting a lost clean sheet with save points in tougher fixtures, evening things out, but there surely must be some value here as well?

Looking at the numbers, the “put as little money as possible into goalkeepers” holds true. In fact, the average points per million decreases from 28 at £4.5m to 21 at £6.0m. The same trend can be seen for the maximum points per million, which decreases from 38 at £4.5m to 24 at £6.0m.

The average points per million has also been fairly consistent across the six seasons, with £4.5m being the price point with the highest value in four of them and the second best in the remaining two seasons.

Then what about “Set and forget”? Spending “As little as possible” usually means a starting £4.5m goalkeeper and a £4.0m on the bench. The difference between this setup and one with rotating £4.5m goalkeepers is only £0.5m. Looking at our target of 30 points per million, that extra £0.5m spent must provide 15 extra points over the course of the season to give value.

I usually prefer to rotate goalkeepers and did last season as well, mainly between Pope and Henderson but I had Ederson for project restart. I took a look at my record and compared it to how many points I would have scored if I had picked one of them as a set-and-forget option. My goalkeepers provided me with 199 points last season. Compared to Pope or Henderson as set-and-forget, that is 29 and 39 points better. I got the rotation wrong in a third of the gameweeks, meaning I benched more points than I had on the field. Even so, I came away with at least 29 extra points for that additional £0.5m spent on goalkeepers, which translates to 58 points per million spent. One season is hardly any proof and I probably had some luck with how the rotation turned out, but I definitely think that goalkeeper rotation is worth more consideration than most FPL managers seem to give it. If you fret the agony of benching the wrong goalkeeper, leave that decision up to the bookies and follow the clean sheet odds when deciding who to play in a given gameweek. You should end up being right more often than not.

In addition to the points upside, it also gives you coverage if your chosen set-and-forget goalkeeper gets injured, suspended or falls completely out of form, saving you a valuable free transfer. Speaking of out of form, like many others, I started this season with Alex McCarthy as my set-and-forget option at goalkeeper. I didn’t really like the other options and I liked Southampton’s defensive form in project restart and their fixtures at the start of this season. We’re only two gameweeks in, but Southampton looks terrible from a defensive point of view. At the same time, I like how Brighton’s new backline is looking as well as Aston Villa’s signing of Emilliano Martinez. Together with Mat Ryan, he provides a £4.5m rotation duo that looks promising.

Brighton and Aston Villa’s matches from GW3-GW15.