Community

Where have all the clean sheets gone?

The season started with a trend among the FFS community for all premium back lines, indeed I advocated that approach in my article based on previous season’s stats. As the gameweeks have ticked by, the apparent lack of clean sheets has led to a shift to budget defenders. But is this season really that different to those gone before?

Total clean sheets

Looking at number of clean sheets in the first 8 gameweeks we see:

2019/20: 41 clean sheets
2018/19: 44 clean sheets
2017/18: 59 clean sheets
2016/17: 36 clean sheets
2015/16: 44 clean sheets
2014/15: 43 clean sheets

The obvious outlier is 2017/18, but otherwise the number of clean sheets at this stage actually seems pretty usual.

Teams with clean sheets

But what about the spread of clean sheets – so far this season, no team has reached 50% clean sheets, how does that compare to previous years?

2019/20: 0 teams
2018/19: 4 teams
2017/18: 6 teams
2016/17: 1 team
2015/16: 4 teams
2014/15: 1 team

Ok, so this season is a little strange with no one having 4 clean sheets in the first 8 gameweeks, however 2016/17 and 2014/15 are pretty similar.

The big six

What about the big boys that usually command the premium prices for their defenders? I looked at the average clean sheets at this stage for the big six teams over the past few seasons:

2019/20: 1.8 clean sheets
2018/19: 3.3 clean sheets
2017/18: 4.5 clean sheets
2016/17: 2.5 clean sheets
2015/16: 3.3 clean sheets
2014/15: 2.0 clean sheets

Again, this season does seem to be on the low side, but its not a million miles from 2014/15 and 2016/17.

What to expect going forward

So, is it still worth having the premium defenders this season?

I’m going to exclude the 2017/18 from this analysis, because it is quite clearly a freak season for clean sheets. For the other season, let’s look at the end result for number of clean sheets:

Season: After GW8 – After GW38

2019/20: 41 – ?
2018/19: 44 – 207
2016/17: 36 – 214
2015/16: 44 – 215
2014/15: 43 – 224

And the average clean sheets for the big six by the end of the season

Season: After GW8  – After GW38

2019/20: 1.8 – ?
2018/19: 3.3 – 12.8
2016/17: 2.5 – 14.5
2015/16: 3.3 – 14.2
2014/15: 2.0 – 13.0

The deviation from the norm is remarkably low, and the correlation between the first eight gameweeks and the season is pretty weak. By the end of the season I would expect to see an average of 13-15 clean sheets for the big six teams as per usual despite the poor start to the season.

For my team, I’m sticking with premium at the back and believing in a regression to the average

29 Comments Post a Comment
  1. Rotation's Alter Ego
    • Fantasy Football Scout Member
    • Has Moderation Rights
    • 12 Years
    4 years, 5 months ago

    Brilliant, thank you for this Hedge.

    It's a tricky one - it's certainly felt that the defences have been far less predictable, but the clean sheets stats don't seem to particularly support this. Perhaps we were just spoilt by the attacking returns of Mendy, Robertson and Alonso in the opening games of last season?

    1. Hedge
      • Fantasy Football Scout Member
      • 12 Years
      4 years, 5 months ago

      I think there is a large element of recency bias. Last year we had similar clean sheets but more weighted to the usual suspects. And the year before we had unusually high clean sheets overall

  2. GreennRed
    • 12 Years
    4 years, 5 months ago

    Defenders can't defend as well as the used to as teams strive to be 'entertaining'. No need for stats to tell you that.

  3. GreennRed
    • 12 Years
    4 years, 5 months ago

    What's a 'community' article anyway?

    1. Rotation's Alter Ego
      • Fantasy Football Scout Member
      • Has Moderation Rights
      • 12 Years
      4 years, 5 months ago

      An article submitted by someone from the community through here:

      https://www.fantasyfootballscout.co.uk/submit-your-article/

      1. GreennRed
        • 12 Years
        4 years, 5 months ago

        Cheers. I'll go back to the regular articles for now.

        1. St Pauli Walnuts
          • Fantasy Football Scout Member
          • 5 Years
          4 years, 5 months ago

          Ingrate is the most suitable word I reckon.

  4. Kuro-o
    • 7 Years
    4 years, 5 months ago

    I want Vote Up button for this article

  5. FPL_Crisis
    • Fantasy Football Scout Member
    • 13 Years
    4 years, 5 months ago

    nice work Hedge. Some obvious team like Man U and Spurs regressing end of last season, start of current will impact CS stats. Liverpool playing much higher up the line, City don't have the same quality without Laporte and Kompany. Also - who knows what the current 'big six' really is!

    1. Hedge
      • Fantasy Football Scout Member
      • 12 Years
      4 years, 5 months ago

      Yes, good point about big six, seems mixed up this year. Would still expect all of the traditional six except arsenal to be in the top 6-7 for clean sheets by GW38 though... we will see

  6. Splatter\\\'s Omerta
    • Fantasy Football Scout Member
    • 11 Years
    4 years, 5 months ago

    Excellent article. Well done. It confirms my sense that this season is more unpredictable for the DEF/GKs. Last couple of seasons I went for 3 heavy hitters at the back, which worked beautifully well (and against the common thread of the Scout 'template'). I started this season the same, but have now gone to all cheap defenders - most expensive being 4.7. There are a number of

    1. Splatter\\\'s Omerta
      • Fantasy Football Scout Member
      • 11 Years
      4 years, 5 months ago

      ...cheapies that are playing regularly and delivering, albeit inconsistently (but so are the expensive ones). I have now dropped back to 5 playing cheapies, most expensive 4.7m, allowing me to spend more up front. Maybe that changes back to type and I'll regret it, but it's fun trying to be a little contrarian. Adapt or Die!

      1. Hedge
        • Fantasy Football Scout Member
        • 12 Years
        4 years, 5 months ago

        The average CS for the big six at GW8 Vs GW38 is the key for me. Regardless of whether the average was 2.0 or 3.3 at this stage they ended on 13-14 by the end.

        Doesn't prove anything, but does make it likely that premiums will come into their own again soon.

        Ultimately, you should be looking for a 10 points per 0.5m (https://www.fantasyfootballscout.co.uk/2019/07/19/player-value-and-the-most-efficient-use-of-the-fpl-budget-part-2/?hc_page=1&hc_sort_by=comment_date#hc_comment_20448853) so it comes down to if you think someone like VVD will score 40 more than someone like Soyunchu

  7. Majestic Chanka
    • 7 Years
    4 years, 5 months ago

    Interesting article, I’m still happy holding my Pool defenders as those CS will come. Also, all those with a cheap defence who have been getting away with it will be found out.

    One point, I don’t think it’s good practice to say, x data is an outlier so I’m going to remove it, unless you remove the extremes from both ends.

    1. Hedge
      • Fantasy Football Scout Member
      • 12 Years
      4 years, 5 months ago

      Probably true, though there was no obvious outlier at the other end. I'm not a statistician though.

      Agree on pool. I'm on TAA, Digne and Otamendi at the moment, but planning to go back to double pool soon

      1. Majestic Chanka
        • 7 Years
        4 years, 5 months ago

        I guess it’s this season! I just think it would interesting to see if the end of season totals were back in line with other years despite the fast start? Good stuff all the same!

  8. The Yorkshire Pirlo
    • 7 Years
    4 years, 5 months ago

    People with Son ditching after the international break?

    I’m looking at moving Son, Pukki to Maddison, Wilson for free.

  9. mookie
    • 10 Years
    4 years, 5 months ago

    Thanks for sharing!

  10. Bring back Rafa
    • 4 Years
    4 years, 5 months ago

    My guess is that teams are playing it from out the back more often than ever. Especially the top teams have been very aggressive with this and are pressing harding than ever.

    When teams get this wrong the opposition scores an easy goal (like we've seen several times this year) and when they get it right they're already in their opposition's half with plenty of space.

    This is why there is less CS this year especially with the 'big 6'.
    Some teams who aren't pressing as hard and are still getting just as many CS , like Burnley & Crystal Palace which is why on average across the competition it's only slightly down

  11. The Rumour Mill
    • 7 Years
    4 years, 5 months ago

    Great article, thanks. Been holding Liverpool double up all season and in terms of expected points they're performing perfectly well for their price. Just being let down in terms of actual points by freak goals (Willems) and the individual mistakes of a second choice keeper (Ings, Maddison).With the underlying stats remaining strong and Allison's return imminent, I'm confident of a return to actual points from them. Facing the worst United attack in living memory next is a good place to start too!

  12. gillespb
    • Fantasy Football Scout Member
    • 12 Years
    4 years, 5 months ago

    My sense of your analysis is that you (perhaps unconsciously) have bent the statistics to produce a result the justifies your intention to keep your premium priced defenders.

    The important statistic IMO is not the absolute number of clean sheets in a season. It is how the clean sheets are distributed. And this year they are more evenly distributed across the league and not dominated by Man City, Liverpool and the other big clubs.

    The price difference between a premium defender and a budget defender is around 2.0 to 2.5 million. That is a lot of money that could be used to upgrade to higher scoring mids and forwards.

    Right now a backline consisting of AA, Robertson, Maguire and Otamendi has tallied 120 points at a total cost of 25.1 million. A back line of consisting of Van Aanholt, Lundstram, Pieters and Dunk would cost you just 19.5 million (even less if you had bought them a few weeks into the season) and has produced 139 points. I commend your effort but IMO your analysis is unsupported by your own facts.

    1. Hedge
      • Fantasy Football Scout Member
      • 12 Years
      4 years, 5 months ago

      You may be right about unconscious bias, however I did address the issue of distribution. While this year has shown a lower average for the big six (1.8), it's not a million miles away from 2014/15 (2.0) which ended up within 1 CS of the average by the end of the season.

      Hindsight is all well and good, but I doubt anyone considered starting the season with PVA, Lundstram, Pieters and Dunk! Returns for the cheaper players are less predictable. While the returns for Liverpool and Man City defenders are certainly lower than normal, I think the data does bear out that clean sheets, at least, will tend towards the norm.

      Either way, the data is there for you to draw your own conclusions - no need to agree with mine 🙂

  13. BRKFCTrueRover
    • Fantasy Football Scout Member
    • 11 Years
    4 years, 5 months ago

    I would be more interested in a CS analysis of the first 19GW and the second 19. I think the likes of city, Liverpool and possibly Chelsea will have really tightened up by then and come into there own for the run in.

    1. Camino Aleatorio
      • Fantasy Football Scout Member
      • 5 Years
      4 years, 5 months ago

      Liverpool maybe

      MCI is playing with 2nd and 3rd options due to injuries. Wonder if they can beg Kompany to come back. Especially if Anderletch fires him as player/coach

  14. Camino Aleatorio
    • Fantasy Football Scout Member
    • 5 Years
    4 years, 5 months ago

    Teams are trying to generate goals. Wingbacks that attack and lots of teams, even small teams play with 4 attacking players.

    How Newcastle scored more than 2 goals shows that attacking even with relegation-level talent can consistently soil clean sheets

  15. Baps hunter
    • Fantasy Football Scout Member
    • 6 Years
    4 years, 5 months ago

    For statistical analysis we should also compare number of goals. If it has risen, it should predict something (more goals means less clean sheets imo).

    Total number of clean sheets is quite irrelevant also. If even let us say four to five teams are solid at the back, we can have decent defense with them. Especially because of Tomori, Soy and Lund. Add also more expensive TAA + Robbo and Ota + Zinch to this equation and it should be quite decent pool to choose from for the "defensive spine".

  16. the Penman
    • Fantasy Football Scout Member
    • 12 Years
    4 years, 5 months ago

    Nice article. There is a general point as well, isn’t there, about how we have less defensive style teams in the league now such as Chelsea or even Spurs or Man U of previous years. The big teams are more committed to pouring forward and keeping the opposition on the back foot than shutting down their own back line, so perhaps it follows from that there’s less predictable clean sheets? And that doesn’t just go for Pep and Klopp at the top, it goes for Potter and Farke at the bottom of the table too...

  17. Nickemon
    • 14 Years
    4 years, 5 months ago

    I’ve always felt that the number of clean sheets in the start of a season is higher after a tournament FREE summer. And yeah, by that I mean a World Cup or a European Championship and not a Copa America.

    My theory is that it gives teams and defences more time to gel and train together.

    But considering this season’s unpredictability and low no of CS’s the theory is either just wrong or the tactics have changed. There is really not that many defense first kind of coaches managing the bigger teams now, wouldn’t you agree?