Fantasy Football Stats

Presenting a value for money team

In Game Week 20 I created a team based on “value for money”. The squad value was £98.8m and I had £1.2m in the bank. It scored 93 points and achieved a Game Week rank of 3,498! What can we learn from this, and what should we look for going forward? Can this help managers on a wild card?

What is “Value for Money”

The most commonly used calculation to work out “value for money” is to divide a player’s points scored by his current cost. I have seen this used on many sites. Indeed, we show this value on Fantasy Football First on our player tables under the column “Points Per Million”, or PPM.

The problem I see with PPM is that good players who have not played many minutes get low PPM scores. For example, a player who has missed 10 weeks through injury will have a lower “Points Per Million” than a player who has played 20 games. Fortunately, we have “PPM Last 4 Weeks” on our site tables allowing us to measure the impact of a player returning from injury, however, PPM is not a good indicator of Value For Money over the long-term.

At Fantasy Football First we use Points Per Million Per Game, or PPMPG, as a measure of “Value For Money”. This is a better indication of value than PPM because it allows us to compare a player with only 10 games to a player with 20 games. Another measure we could use is “Points Per Million per 90 minutes”. This allows us to see the worth of a super-sub as if they regularly played 90 minutes. Both measures have merit as a measure of Value For Money, and us FPL Managers can use either statistic depending on if we want to exclude impact players or not.

In this article I am going to use Points per Million Per 90 Minutes, or PPMP90.

My Team

Site Team

The standout performance was captain Giroud. Also, selecting Daniels ahead of Alonso resulted in a net gain of 10pts. The question is why did I select the team, was it a one off performance and can the tools used, be used again in subsequent week?

Value for Money Tables

Forwards

Rather than showing PPMP90 as a table, here I show it as a grid of the top 20 scoring forwards at the end of Game Week 19, where total points scored is plotted against their PPMP90. At the intersection between the two I show their price (at the time of writing). Also, plotted on the left of the table is their minutes played and teams selected by.

Click to download

Click to download

At the top of the table is the median of points (63.5) and median of value for money (£0.69). These are then used to colour code the table. Players who score above both medians are in the green section. Players below both medians are in the red section. Players above one or other median are in one of the yellow sections.

The first point of interest is that most players fall into the yellow sections. The top right yellow section has the premium forwards, who tend to score the most points, have high ownership and have played the most minutes; but offer poor value for money (excluding when you consider them as a captain). The bottom left section tends to have forwards who have played less minutes, scored less and have lower ownership, but offer the best value for money. This section could contain hidden gems that if they got more playing time they could return substantial points.

In theory, the green section could contain ‘must have’ players however, in practice it only has two forwards Defoe & Negredo. They only just make it into the section and both play for teams in the bottom six. I have a bias where I will not select players from teams that are in the bottom six or I think will end up in the bottom six.

The red section contains players to be avoided. So I did.

In selecting the forwards I went for balance and picked one from each of the yellow sections. The standout player in the bottom left section was Giroud. He had the second highest PPMP90. Only Iheanacho was higher but his game time is limited by Aguero. Due to Walcott’s injury I thought that Giroud’s chances of playing increased. Also Arsenal have good fixtures. If you want to compare PPMPG with fixture difficulty, try the FPL Player Rotation Checker on Fantasy Football First.

From the top section I rejected Costa, even though he has the highest score and relative good PPMP90. This is because there is better PPMP90 in Chelsea defence and midfield. Zlatan and Aguero were rejected for their low PPMP90.

Lukaku and Kane have the same PPMP90 and are in the middle of the section. I went with Lukaku as he is cheaper and Everton has good fixtures.

Midfielders

The Midfielder table is below and it is this table that convinced me to go with the 352 formation.

Table Mids

Click to download

The points scored median (88.5) is higher than the forward median (63.5). There are four genuine players in the green section (unlike the forwards section where we were starved of players).

I selected three out of these four for my team: Lallana, Alli and Phillips. Walcott was rejected due to being injured. He may come into the team in later weeks.

Like the forwards, the other two midfield positions were selected based on balance. Hazard because he offers the best PPMP90 from the premium midfielders. Antonio because he is just outside the green section and West Ham has good fixtures in Game Week 21 & 22.

Interestingly, the red section contains Sigurdsson and Payet and both of these players are in my regular team. Maybe time to swap them out for players in the green section, as only Phillips out of these is in my regular team.

Defenders

The defenders table shows there are many value for money options. Mainly, Chelsea and Tottenham premium players.

Table Defs

Click to download

As I picked PPMP90 forwards and midfielders, the budget was available not to have to compromise on the defence and I was able to select three premium defenders and Charlie Daniels.

Rose was selected over the more popular Walker, as if they get the same playing time between now and the end of the season I expect Rose to outscore Walker – interestingly the RMT Optibot thinks the same too!

A popular discussion in the FPL community is the validity of doubling up on Chelsea defenders, which was discussed on this week’s FPL Podcast. For my team I was able to double up on Chelsea defenders, playing Azpilicueta and Alsonso. Cahill was rejected as Alonso has better PPMP90 and Azpilicueta has higher ownership. Although, I think any combination of two out of three would have been ok.

Goalkeepers

Table Goalies

Click to download

I may have made a mistake with the keepers. I went with Grant and Boruc. Grant because of his PPMP90 and Boruc because Bournemouth has good near-term fixtures. However, Heaton does appear to be the stand out player. After Bournemouth’s good fixtures I may correct this mistake.

Interestingly, the most selected keeper in the game, De Gea at 35%, does not seem to offer value for money – although I note United’s form has changed in the last 10 games compared to the first 10 games.

Limitations

I think there are three limitations of this method:

  • The tables are only looking at the top 20 players in each position so I might be overlooking some real gems. One example is Yaya Toure. He was frozen out in the first half of the season but is now a regular starter. He will not be subject to Champions League rotation, as he is not in MCI’s Champions League squad and he now appears to be the first choice penalty taker. PPMPG Last 4 Games will help with this, and is available on most stats pages off the Player Menu at the top of the site.
  • The January transfer window could bring some excellent new players to the Premier League who will not appear in these tables.
  • The stats do not take Captaincy into account, which gives a player double points per million per 90 (or per game).

Going Forward

In the introduction, I asked “was the performance of my FPL team sustainable?”. My gut feel is a gameweek rank of top-5k every week is incredibly unlikely. However, I do feel this Value For Money approach will give me game week scores far greater than the average.

Going forward, I will monitor the team performance, update my tables and make transfers based on the updates. If after 4 or 5 weeks this team outscores the game week average then I may play the second wildcard early in my regular team. You can follow a similar process using the free tables on Fantasy Football First.

Best of luck and please ask if you have any questions on the process or on player value for money.

Authors

  • Bio
  • Latest posts

InfernoSix

Fantasy football manager who is using stats to try and score 2500 points in 2016/17 season.

40 thoughts on “Presenting a value for money team

    • @ChorleyRocks Thank you for posting the article. Thanks to everyone below for all the positive comments. Greatly appreciated.

  1. Interesting article. Nice to see the stats being considered by the human rather being blindly followed…
    It might be an idea to consider yellow yards in future. Rose has 8 so could get a ban soon, Walker has only 5, therefore I think Walker would have been the better choice.

  2. Liverpool goalkeeper Loris Karius has had to deal with heavy criticism in his short Anfield career and was dropped after a nightmare performance in the 4-3 loss at Bournemouth.
    Kumast, can you explain the situation :question:

  3. A good read and well-researched article. Very interesting, had me nodding, scribbling down notes and waving my pipe around in a scholarly way.

    • @Jbugman you can use PPMPG from the Player Menu at the top of the site, for example, for Mids click on the player menu, then select Midfielders. Once the page has loaded click on PPMPG to sort by Points Per Million Per Game.

        • @ChorleyRocks Is there a way to use these analytics in combination with the predictions of the bot?

          Excuse me if I sound audacious and greedy.

        • @Zwischenzug So Historical PPMPG on one axis, Predicted PPMPG for the next 6 weeks on the other axis, and then a magic quadrant of players who have delivered Value for money and are expected to keep delivering value for money? Or something different?

        • @ChorleyRocks Yes, exactly that. Of course in a table that shows overall points, TSB, team and MP.

          It shouldn’t be hard with someone that has access inside the bot.

          And maybe be able to alternate the 2 variables. For example, PPMPG with PPMPG L4W or PPG L4W with Predicted PPMG for the next 6 weeks.

          (Now in order to be as greedy as it gets:
          May we, please, have it before this deadline so I can do my wildcard?)

          I actually believed I was close deciding my wildcard team and then InfernoSix decided to publish his article.

        • @ChorleyRocks Oh. Well, there’s no need to add to the existing costs despite the fact that it will be helpful.

          In this view, since I am extremely fond of the communities that work for the community per se, please do share if there is any workload that can be shared within the community and that we can actually help internally. Unfortunately, I am not a programmer but maybe we can help in other ways.

          Thank you!

  4. InfernoSix,

    Thank you for a most fascinating article. Will you be providing updates to your team and its progress?

Comments are closed.