Statsfindings

Now that the FPL season has come to an end, people may want to take a break or review how their season went. One aspect of the game people may consider if they analyse their decisions is chip strategy. Due to the large number of double gameweeks and the extra Free Hit Chip, there were lots of possible chip strategies this season.

Thanks to Sertalp who has collected data on when members of the Analytics XP mini-league played their chips, it’s possible to analyse how playing chips at different times affected Points totals at the end of the season. This analysis concerns the use of the Triple Captain Chip and Second Wildcard, but further analysis could be performed on use of the Free Hit and Bench Boost Chips.

Analytics XP League

The Analytics XP league is a mini-league with 499 players. These players tend to rely heavily on statistics, such as expected goals, plus model predictions to assist their transfer and captaincy decisions. The use of optimisers to find the best possible transfer, based on predicted points, is also common. FPLReview is a well known website with a model and optimiser. It’s likely that players in this mini-league have their own models, including me.

Members of this mini-league who you may recognize from Twitter include Trout, Simon and Sigurd.

The Data

From these 499 players, only players who had used all their chips by the end of the season were included in the final analysis. This was done to try and filter out ‘inactive’ managers i.e., those who had given up.

Removing those managers left 428 teams to analyse. FPL points instead of Massive Data points (from FPL Review) were used to analyse chip strategy against final outcome. This is because FPL points are easier to understand and more realistic than Massive Data points, plus there’s a strong correlation between the two variables (r = 0.83) in this sample, so the choice of outcome variable shouldn’t have too much effect on the analysis.

One limitation of this dataset is that FPL points (or Massive Data points) were only available after GW37 instead of GW38, the final gameweek of the season. However, one week’s worth of points shouldn’t make a big difference to the analysis.

Opportunities to Play the Chips

Just to refresh your memory, the main conundrum on when to play the Second Wildcard and Triple Captain Chip revolved around GW26. This is because Salah had a Double Gameweek with Norwich and Leeds at home.

However, GW26 was followed by double gameweeks in 28 and 29. This raised the question of whether the wildcard should be used in GW26 to set up a team for all three double gameweeks. Similarly, one option was to wildcard in GW28 to prepare for that week and GW29, with the advantage being that you could use the triple captain chip in GW26.

The consensus at the time was that managers using a wildcard in GW26 should play a Free Hit Chip in GW27 (popular picks in GW26 blanked in 27) and managers playing their Triple Captain Chip in GW26 should save their wildcard for GW34 (or GW35), to prepare for Bench Boost in GW36.

When Were the Chips Played?

Starting with the Triple Captain Chip, approximately two-thirds of the sample used the chip in GW26.

Using a wildcard in GW34 or 35 to set up a bench boost in GW36 was the most popular option, chosen by 37% of managers. From those choosing to Wildcard in GW26 or GW28, both options were just as popular, chosen by 19% and 16% of managers. Finally, that leaves 27% of managers who played their Second Wildcard in a different week.

What Was the Effect on FPL Points?

Managers who played their Triple Captain Chip in GW26 scored slightly more FPL points than those who didn’t.

Unlike the Triple Captain Chip, the timing of the Second Wildcard had a greater effect on FPL points scored. Managers who wildcarded just before GW36 or in GW26 scored the most FPL points, with pre GW36 wildcards just edging it. GW28 wildcards were also a viable option, although managers who played their wildcard in GW28 scored noticeably fewer points than the previous two groups, on average. Finally, Managers who played their wildcard in any other week scored a lot fewer points than the other wildcard strategies.

Now focusing on chips played in GW26, GW26 wildcarders ended up with marginally more points than managers who played their Triple Captain Chip that week.

However, when considering chip strategies as a whole instead of just the chips played in one week, we can see that there was almost no difference in final FPL points between GW26 triple containers or GW26 wildcarders. Whilst managers who wildcarded just before GW36 performed better on average than those who used their Wildcard in GW28, all three of these strategies scored many more points on average than managers with an no clear chip plan, or an unconventional plan.

Perhaps one lesson we can learn from this is that whilst managers frantically deliberated on whether to Wildcard or Triple Captain in GW26, as long as you follow a sensible plan, it doesn’t make too much difference on your final points total. One lesson to be learnt could be to try and avoid an early wildcard and wait until the double gameweeks to use it.

Limitations

The main problem with this analysis is that it’s on a specific sample of managers. Perhaps the fact that managers in this league often use expected goal statistics and model predictions to aid their transfer decisions has a bigger effect on their final FPL points than the chip strategy they use.

A better analysis of which chip strategies were the best could be conducted by choosing a larger sample of managers at random.