Player value and the most efficient use of the FPL budget: Part 2

Following my previous article looking at the most efficient way to use the FPL budget, I’ve rerun the numbers looking at the new season pricing with last year’s scores.

As a reminder, the approach I’m using is to take a base team of 11 of the cheapest possible players, then make upgrades to that team that offered the best value (points per million) until the team value was at £100m. You can read the full approach and some thoughts on its limitations in the previous article.

2019/20 Prices

Running the approach using 2019/20 prices and 2018/19 scores, the formula spits out the following team in a 4-5-1 formation as the optimum with a score of 2079 (excluding captain points):

GK: Ederson | 6.0m | 169
DEF: Virgil Van Dijk | 6.5m | 208
DEF: David Luiz | 6.0m | 164
DEF: Aymeric Laporte | 6.5m | 177
DEF: Andrew Robertson | 7.0m | 213
MID: Ryan Fraser | 7.5m | 181
MID: Gylfi Sigurdsson | 8.0m | 182
MID: Paul Pogba | 8.5m | 179
MID: Luka Milivojevic |7.0m | 166
MID: Mohammed Salah | 12.5m | 259
FWD: Jiminez | 7.5m | 181

Interestingly, the second best formation comes out as 5-4-1 at 2074, with Marcos Alonso (6.5m | 161) instead of Milivojevic.

FPL obviously price up players based on expected performance, so when we look back retrospectively, the underpriced gems rise to the fore and gives us more budget for the expensive midfield and forward options.

What would be interesting would be to see what the predictive picture would have looked like at the beginning of previous seasons and compare that to what things actually looked like retrospectively at the end of that season. This may give us some hints as to how the picture for 2019/20 will look compared to this predicted model.

2018/19 ‘Prediction’

I’ve rerun the formula with the 2018/19 prices, and the 2017/18 season scores. This emulates what we would have seen if we had done this same process at the start of last season. If we had done so, we would have had the following suggested team in 4-5-1 formation:

GK: Lukasz Fabianski | 4.5m | 157
DEF: Cesar Azpilicueta | 6.5m | 175
DEF: Marcus Alonso | 6.5m | 165
DEF: Antonio Valencia | 6.5m | 146
DEF: Nicholas Otamendi | 6.5m | 156
MID: Mohammed Salah | 13.0m | 303
MID: Abdoulaye Doucoure | 6.0m | 136
MID: Raheem Sterling | 11.0m | 229
MID: Pascal Gross | 7.0m | 164
MID: Luka Milivojevic | 6.5m | 144
FWD: Jamie Vardy | 9.0m | 183

The second best formation would have come out as 5-4-1, just a single point behind swapping out Valencia and Doucoure for Ben Davies (6.0m | 143) and Jan Vertonghen (6.0m | 138).

This mimics the same two top formations that come out of the 2019/20 predictive model.

Rerun 2017/18 ‘Prediction’

Running the same method as if we were at the start of the 2017/18 season with 2016/17 points totals gives us, yet again, 4-5-1 as the best predicted formation:

GK: Tom Heaton | 5.0m | 149
DEF: Charlie Daniels | 5.0m | 134
DEF: Marcus Alonso | 7.0m | 177
DEF: Gary Cahill | 6.5m | 178
DEF: Cesar Azpilicueta | 6.5m | 170
MID: Christian Eriksen | 9.5m | 218
MID: Gylfi Sigurdsson | 8.5m | 181
MID: Robert Snodgrass | 6.0m | 133
MID: Deli Alli | 9.5m | 225
MID: Sanchez | 12.0m | 264
FWD: Josh King | 7.5m | 178

Second best, only one point behind was 4-4-2, swapping Sigurdsson for Roberto Firmino (8.5m | 180). A further one point behind we see 5-4-1 appear again with Snodgrass dropping out for Gareth McAuley (5.0m | 131).

Comparing predictive to retrospective

Formation

The first thing that stands out, is that the predicted best line up using this method every year is 4-5-1, with 5-4-1 following closely behind.

What came out for the retrospective looks at the 2017/18 and 2018/19 seasons was a 3-5-2 (see the previous article for the full result). Each time 4-5-1 wasn’t far off, however 5-4-1 dropped a considerable distance behind (6th best formation each time on the retrospective).

Price Points

The reason that 5-4-1 looks like it will be a good formation at the start of the season is, simply, that the premium midfields/forwards that scored well are now really expensive! That means its hard to get many in and therefore much easier to get a better team with five big hitters at the back.

Why does 5-4-1 end the season so far off the pace? Those underpriced gems that emerge means more money to spread around. This also seems to be why 3-5-2 is better in retrospective than looking forward at the start.

So, for example, in 2018/19 Alexander-Arnold (5.0m | 185), Fraser (5.5m | 181), Wilson (6.0m | 168) and Jimenez (5.5m | 181) all outperformed their price significantly leaving enough money to upgrade the goalkeeper to Ederson (5.5m | 169), and a premium defender to Hazard (10.5m | 238)

In 2017/18 it was Salah (9m | 303), Sterling (8m | 229) and Gross (5.5m | 164) who allowed budget to upgrade a cheap forward to Firmino (8.5m | 181) and a cheap defender to Vardy (8.5m | 183).

Actual Players

In both seasons, there are players that look like they will be good value for money at the start of the season, and do in fact turn out to be that good. In 2018/19 it was Salah and Sterling, it was Eriksen, Alonso and Azpilicueta.

The aim of the exercise is not to predict which players are best to get, but more the themes that repeat season-after-season. Your team needs to be flexible to jump on the specific players that emerge and knowing the pattern of past seasons can help you prepare.

What to expect in 2019/20

At the start of the season the 6m and under midfielders and forwards look uninspiring, yet most seasons one or two options emerge scoring around the 170-180 mark.

At the start of the season, it looks like we need to cram in a couple of 11m+ big hitters, yet there’s almost always one or two in the 8-10m range that hit over 200 points.

For both those reasons, five at the back (often 5-4-1) looks tempting at the start of the season. In reality, once those overperforming players emerge, we simply can’t spend enough money in that formation to make it count and 3-5-2, 4-5-1 or 4-4-2 become better options.

Don’t start the season with a formation that doesn’t easily allow you to jump on a 5-6m midfielder/forward and an 8-10m midfielder/forward. Those are the price points you need to be able to jump to quickly.

For my team, I think Vardy and De Bruyne could be this season’s 8-10m men who hit the 200 mark. I really don’t know who the 6m or under player will be who hits the 170-180 range, but I’m fairly confident there will be at least one. I will start with four premium defenders at the back so that one of them can easily become a ~6m midfield or forward when that gem emerges.

Still to come…

A limitation of this approach is that it doesn’t account for rotating pairs in defence and attack and whether they can offer better value for money upgrades over keeping a single player in your team long term.

I am hoping to put together a follow-up article that will delve into the data and consider the case for rotating pairs.