I approached this analysis with a hunch that the latest season of FPL in 2016/17 was more difficult than the previous two seasons I have been involved in since I started playing again. There seemed to be more difficult choices to make, more players in form and less consistency throughout the season. Personally I felt I was frequently making wrong calls where I would transfer out a player only for him to go on a run of points.
To test the theory, I broke it down into three hypotheses to be tested:
- There was more rotation, meaning key players were getting dropped more frequently
- There was fewer budget players scoring high point totals
- The form of players was fluctuating more than in previous seasons
Below is a breakdown of what I have found.
Hypothesis #1: There was more rotation, meaning key players were getting dropped more frequently
Figure 1: players involved in Premier League games, 2014-2017
The table above appears to show that the first of my assumptions was incorrect. There was only a slight increase from 2015/16 to 2016/17 in the number of players playing more than the equivalent of 30 games, and an equal amount playing more than the equivalent of 22 games. There has been a drop in the last two seasons of the number of players playing less than six games’ worth of minutes in a season and an increase in those playing more than 30, indicating that Premier League managers are actually slightly more settled and less willing to experiment than they were in 2014/15.
If we isolate the data to include only the traditional ‘big six’ clubs (both Manchester clubs, Tottenham, Arsenal, Liverpool and Chelsea), we see that there is a more pronounced pattern towards less rotation. The number of players playing the equivalent of more than six games has decreased indicating that there are fewer players involved, and the number playing more than 30 has increased season-on-season. This suggests that there are more key players amongst the so-called big teams, but that the number of squad players playing getting a chance has decreased since in the last couple of seasons.
Figure 2: players involved in Premier League games, ‘big six’ clubs, 2014-2017
There has been an increase in the number of players playing an active role (the equivalent of 22 games or more) since 2014/15, however this amounts to just 0.5 players per club.
In summary, there was not more rotation in 2016/17, and in fact the number of core players amongst the big teams is steadily increasing. Where it has been more difficult regarding these teams – and indeed the whole league – has been the fact that more players are scoring big points season on season. Figure 3 below shows that the number of 100+ point scoring players has increased each season, suggesting that the points are being shared around more. The jump from 2015/16 to 2016/17 has been more pronounced amongst the ‘big six’ (+20%, compared to +8% for all clubs)
Figure 3: number of players to score more than 100 points, by season
Therefore, it is not that there was more rotation last season, but rather it just seemed that way because the players scoring points were rotating throughout the team.
Hypothesis #2: There was fewer budget players scoring high point totals
On the surface this hypothesis seemed anecdotally easy to prove: the emergence of Leicester in 2015/16 champions, as well as Watford’s Ighalo and Tottenham’s Alli meant that it was a season where the successful budget players made it easier to get the big hitters in. But in truth, that was only half the story.
Figure 4: value (£m) of the top 10/20 point scorers in ascending order, by season
Figure 4 highlights that there was a dearth of quality budget options in 2016/17. The top 20 point scoring midfielders, for example, had only three budget players (under £6.5m) whereas the previous season there had been had been six. 2016/17 similarly didn’t have any sub-6.5m forwards in the top 10 whereas 2015/16 had four. It should be noted however that 2014/15 and 2015/16 were similar overall, meaning that the lack of budget players performing in 2016/17 was not a return to normal service after the Leicester anomaly.
The other half of the story though was the return to form of the big-hitters.
Figure 5: the median rank of total points amongst positional peers of premium-priced players, by season (table)
Figure 6: the median rank of total points among positional peers of premium-priced players, by season (chart)
Figures 5 and 6 show that there was a concentration of premium-priced players towards the top of the points charts in 2016/17, more so than in either of the previous two seasons. The ‘median rank’ is the point rank of the middle player in the sample amongst his positional peers. In simple terms, taking the forwards as an example, there were 11 £9m+ forwards in the game this season (taking their starting price as the base) and the rank amongst the forwards of the middle player – the 6th in this sample of 11 – was 7th. Put another way, that means that there was only one non-premium forward (Jermain Defoe) in the top seven points scoring forwards. In 2015/16, the median rank was 10th from seven players, meaning that there were only four premium forwards in the top 10. In short, the lower the median rank, the higher the concentration of premium players at the top of the scoring charts.
As we can see, 2016/17 was a bumper season for the premium players in every position. This is brought into sharper focus in figure 7, which shows the cumulative value of the top 30 players in each season.
Figure 7: cumulative value of the top 30 players, by season
The data shows that to afford the top 10 players (at their starting price) in 2015/16 you would need £76.5m, but that figure jumps to £94m (+£18.5m) in 2016/17. To get the top 30 players, you would need an extra £36.5m in 2016/17.
This, coupled with the relative lack of high-scoring low-value players put significant strain on the budget of FPL managers. When we combine this knowledge with that from the rotation section of this report – which showed that more players were scoring 100+ points – it shows that fitting in the correct players at the correct times proved remarkably difficult.
Hypothesis #3: The form of players was fluctuating more than in previous seasons
A key factor when bringing a player in during the season is whether you believe that player will go on a run of form over the coming games to justify the investment (indeed, it is the only factor aside from whether you can afford him). I have set here the criteria of ‘form’ to mean an average of six points per game over a run of six games (fewer as the games run out at the end of the season). For example, if a player went on a run from gameweeks 1 to 6 scoring 38 points (6.33 points per week), that would be considered a ‘run of form’ by this methodology (note: if that same players scored 40 points from gameweeks 2 to 7 at an average of 6.66 points per week, that would also count here as a run of form).
Figure 8: number of ‘runs of form,’ by position and season
The data shows that my initial assumption was incorrect; it was in fact more opportunities to get in a player who was on, or about to go on, a ‘run of form’. This was not true for goalkeepers, who had a more torrid time of it this season, and there were only slightly fewer runs of form for forwards than there were last season. However, the sharp increase in runs of form for defenders and an even sharper increase for midfielders showed that 2016/17 was actually easier to find consistent players.
However, when we factor in this data to the previous findings, we understand that a lot of these runs would have been from premium players. Therefore this added to the illusion that it was a more difficult season: if you have a player who has just blanked, it’s likely that a big hitter somewhere else didn’t making you feel like you have a) missed out and b) cannot afford to bring him in anyway. This can make the game seem more difficult.
The data shows that if 2016/17 was – or seemed – more difficult it was because of the premium players hitting the mark more frequently than in previous seasons. There was not much more rotation of players, and in fact there were a lot more players getting good point total and more opportunities to find a form player in 2016/17 than previously. However, the problem was that the budgets of FPL managers were stretched to the limit because of the lack of budget players performing well and the number of premium players scoring big.
A couple of personal anecdotes from my season – captured in this article – highlight this problem. On separate occasions I transferred out heavy-hitting Eden Hazard and Harry Kane for Alexis Sanchez and Zlatan Ibrahimovic, only for the players I transferred out to go on great runs of form after that. The fact was that many of the big name players were scoring frequently at various points of the season, and whereas in previous seasons you could rely on at least a partial template of outstanding performers this season there was almost too much variety. This lead to frustration and more perceived difficulties as the season progressed.
So what can we learn from this going into the 2017/18 season? Ultimately it is difficult to know because there is nothing to suggest that next season will be like this season. However, I would expect there to be a continuation of points being shared around the team, particularly from the big teams. Last season there were isolated examples of there being a ‘must-own’ player from a particular club who eclipsed everyone else; Alexis Sanchez and Romelu Lukaku were the two main protagonists here. However, for the most part the attacks of the top teams were loaded with threat and fluidity making it difficult to know who was going to crop up with a goal. There is no reason to suspect this trend won’t continue so it will be important to look at the performance of a player over time rather than in isolated games. It is my opinion that differentiating between an expensive player having a quiet game and a loss of form that needs to be addressed will be the toughest test of an FPL manager’s analysis skills next season.
As always, I can be reached on Twitter @artemidorus_1