634x258 113
Community Submissions October 1

One Stat To Rule Them All

113 Comments
Share

Jonty often brings up on the Scoutcast about how he loves his shots in the box stat, or chances created for midfielders – doing what midfielders should. Of course there is community regular Doosra with a stat table always at the ready to answer every question. Statistics are clearly a staple part of many Fantasy managers psyche, but which single stat is the best to follow. Here I’ve sought to find the best correlation between goals and assists over the last four seasons, crunching some numbers to find out what is the best stat to use to predict Fantasy returns.

Results

stats1

 

The above chart shows the average R squared per stat over the last four seasons. R squared is basically how well the stat explains the differences in performance (a more technically correct definition is how it charts the differences from the mean, in this case a linear trend line joining the stat to goals and or assists). I worked out how well the shot based stats per player correlate to goals scored over the season, the touches and passes received to goals + assists and the chances created and throughballs to assists.

What the data shows over four seasons (2011-2014) is at least over a season that Jonty and other shots in the box fans are better off looking at shots on target. Also touches in the area and passes received type involvement based statistics are pretty poor predictors of goals and assists relatively (especially when used in isolation) and chances created and really good chances created (throughballs) are really poor at explaining assists.

We can also expect reversion to the mean, which means players that start out with lots of goals but poor shots on target numbers, such as Anthony Martial, over a whole season will revert closer to what the shots on target numbers predict. Conversely, players with high shots on target numbers but a low goal tally look to be good acquistion targets if they can maintain their shots on target statistics.  This bodes well for Sergio Aguero owners in particular.

This Season

Looking at the season so far for the best shots on target and big chances (two highest R2)

Shots on Target:

soTarget

 

Big chances:

bchances

Of course you can and should use stats in combination and small sample sizes for this season caveat but I thought it was an interesting exercise and is hopefully useful for some people. It is clear that with shots on target there really is one stat to rule them all.

rakkhi Love my football, love my stats, hoping to improve each year. Go the gunners! @rakkhis on Twitter Follow them on Twitter

113 Comments Post a Comment
  1. J0E
    • Fantasy Football Scout Member
    • Has Moderation Rights
    • 14 Years
    8 years, 7 months ago

    Excellent analysis. Shots in Box promptly switched to shots on target. Cheers. 🙂

    Open Controls
    1. rakkhi
      • Fantasy Football Scout Member
      • 13 Years
      8 years, 7 months ago

      Thanks for the edits and posting Jonty. Just a couple of spelling mistakes in the edits: sort to fund should be sought to find. Cheers!

      Open Controls
    2. Albertini
      • Fantasy Football Scout Member
      • 10 Years
      8 years, 7 months ago

      Jonty, the second chart picture shown is the same as the first one (again shots on target instead of big chances).

      Excellent stuff Rakkhi, thanks.

      Open Controls
      1. rakkhi
        • Fantasy Football Scout Member
        • 13 Years
        8 years, 7 months ago

        Posted big chances link below. Flagged mod to update.

        Open Controls
      2. J0E
        • Fantasy Football Scout Member
        • Has Moderation Rights
        • 14 Years
        8 years, 7 months ago

        That's been changed...may take a few minutes to kick in. It's also available via the link in the text.

        Open Controls
  2. ​
    • 10 Years
    8 years, 7 months ago

    Yes, shots on target are an explanatory statistic of where the goals have come from, but which is actually the best predictive statistic over the next period?

    Open Controls
    1. rakkhi
      • Fantasy Football Scout Member
      • 13 Years
      8 years, 7 months ago

      Surprisingly similar. A good expected goals model followed by shots on target. http://statsbomb.com/2015/09/learning-from-early-games/

      Open Controls
  3. Sheeeit
    • Fantasy Football Scout Member
    • 8 Years
    8 years, 7 months ago

    Great article! I love using the data, but I'm not always sure I'm interpreting it correctly. It's quite easy to get twist the stats to fit in with your own misguided methodology.

    So more articles like this please.

    Open Controls
    1. Barry Woj
      • 11 Years
      8 years, 7 months ago

      I highly recommend reading http://www.statsbomb.com and following some of those involved on Twitter.

      Open Controls
  4. Latchmere Wave Machine
    • 9 Years
    8 years, 7 months ago

    Great stuff, if you did the same kind of thing with Defenders what would be the stat to follow?

    Open Controls
    1. J0E
      • Fantasy Football Scout Member
      • Has Moderation Rights
      • 14 Years
      8 years, 7 months ago

      Tough one I'd imagine as clean sheet points are their bread and butter and that relies on four other blokes doing their job too.

      Open Controls
    2. rakkhi
      • Fantasy Football Scout Member
      • 13 Years
      8 years, 7 months ago

      Be interesting to look at shots in the box conceded and shots on target conceded which are the main one I use maybe also with crosses and chances created. Problem is no one seems to publish the full points per player for FPL for 5 years, not even stats here as FFS only has Opta stats. That would be the most interesting correlation. I'll see if I can do clean sheets vs the other metrics and post the graph.

      Open Controls
      1. Woy of the Wovers
        • 13 Years
        8 years, 6 months ago

        I use a combination of SitB conceded and Defensive Errors to predict goals. Needs a sort of "queue" model to convert goals conceded to clean sheets.

        Open Controls
  5. Qubit
    • 14 Years
    8 years, 7 months ago

    Good article 🙂 Ive been using shots on target for some time as my own analysis showed it to be the best indicator. Should really keep information like this to ourselves to be honest 😉

    This is a good read (but not exactly light!):

    http://onlinelibrary.wiley.com/doi/10.1111/j.1756-8765.2008.01006.x/full

    I actually strongly suspect that shots on target in the box would be the best single indicator, but we don't have access to that data - although i believe it exists as you can see in the comparison tool.

    Open Controls
    1. RedLightning
      • Fantasy Football Scout Member
      • Has Moderation Rights
      • 13 Years
      8 years, 6 months ago

      That's a very long paper, but an illuminating one.

      tl;dr: The best single heuristic is often more effective than overcomplicating things by trying to take too many different factors into account. So keep it simple, because less is more!

      Open Controls
      1. Essaraba
        • 8 Years
        8 years, 6 months ago

        this td;lr is very much appreciated thanks 🙂

        Open Controls
      2. Qubit
        • 14 Years
        8 years, 6 months ago

        Im glad someone else found it interesting 🙂 Dont blame you for the tl:dr 😉

        Open Controls
        1. RedLightning
          • Fantasy Football Scout Member
          • Has Moderation Rights
          • 13 Years
          8 years, 6 months ago

          I read most of it, but I put in the tl:dr for the benefit of others who might find it too long to read it all themselves.

          I could have added that the reason that the single best heuristic wins is that adding others sometimes merely increases the background noise and detracts from the effectiveness of the best one for forecasting purposes.

          Open Controls
          1. Qubit
            • 14 Years
            8 years, 6 months ago

            Yes, effectively taking you further away from the 'truth'. Whether or not SOT here is the best really needs further investigation, i think lumping assists in with goals is probably not wise for the same reason, they need separating. I still really want to have a look at the shots in the box on target data, but doesnt looks likely.

            Open Controls
            1. RedLightning
              • Fantasy Football Scout Member
              • Has Moderation Rights
              • 13 Years
              8 years, 6 months ago

              Would Paul Riley's expected goals model quoted below by Barry Woj be any use for this?

              Open Controls
    2. Camp No No
      • 10 Years
      8 years, 6 months ago

      tl;dr "casuals rule"

      Open Controls
      1. RedLightning
        • Fantasy Football Scout Member
        • Has Moderation Rights
        • 13 Years
        8 years, 6 months ago

        Casuals may well rely on simple factors such as Total Points, points or goals scored in the most recent game-week, and picking players who they have heard are good footballers, and basing their selection on these should be rather better than making a completely random choice.

        However, if more information is available then none of these factors is likely to provide the single best heuristic. Anyone who is even aware of Shots on Target is unlikely to be a Casual.

        Some of the examples quoted in the paper are for simple cases, such as arranging Swiss cities in order of population when the most relevant thing you know about them is how often you have heard them mentioned.

        But it does also mention a case where a 12th degree formula can model the past data exactly, but where a 3rd degree formula is better for forecasting purposes because it eliminates more of the randomness and noise.

        Open Controls
  6. Christina.
    • 14 Years
    8 years, 7 months ago

    Yet we are all going Kun(c).

    😕

    Open Controls
    1. rakkhi
      • Fantasy Football Scout Member
      • 13 Years
      8 years, 7 months ago

      Not sure why you say yet. The analysis supports Kun has had quite a lot of shots or target and big chances relative to his goals. If there is a large regression to the mean like GW8 last year then people that captain will benefit

      Open Controls
      1. Doosra - ☭DeclanMyGeniusâ…
        • Fantasy Football Scout Member
        • 14 Years
        8 years, 7 months ago

        http://members.fantasyfootballscout.co.uk/my-stats-tables/view/939/

        Simply not true, in any way.

        http://members.fantasyfootballscout.co.uk/my-stats-tables/view/11091/

        Open Controls
        1. ​
          • 10 Years
          8 years, 7 months ago

          I'd rather take the full sample

          http://members.fantasyfootballscout.co.uk/my-stats-tables/view/10778/

          Open Controls
          1. Doosra - ☭DeclanMyGeniusâ…
            • Fantasy Football Scout Member
            • 14 Years
            8 years, 7 months ago

            OK. 🙂

            Open Controls
    2. Pacer.
      • 9 Years
      8 years, 7 months ago

      And perhaps rightly so. Not only are Aguero and other players due a mediocrity regression but Aguero almost constantly outperforms peers. No doubt people have scored excellent points by being brave and putting the (C) elsewhere, but taken on a weekly basis there's again a very good chance of a big haul for Aguero (C) players

      Open Controls
      1. Christina.
        • 14 Years
        8 years, 7 months ago

        *takes tongue out of cheek

        😉

        Open Controls
  7. A.T
    • 13 Years
    8 years, 7 months ago

    One stat to rule them all
    One stat to find them
    One stat to bring them all
    And in the darkness bind them

    👿

    Open Controls
    1. Doosra - ☭DeclanMyGeniusâ…
      • Fantasy Football Scout Member
      • 14 Years
      8 years, 7 months ago

      The whole point about stats is variety, though ... 😉

      Open Controls
      1. rakkhi
        • Fantasy Football Scout Member
        • 13 Years
        8 years, 7 months ago

        Yeah worth considering a full picture but I often find sorting on one or two measures to get who the leading player is that I should get in or for finding the weakest link in the team so that's useful for me at least.

        Open Controls
        1. Doosra - ☭DeclanMyGeniusâ…
          • Fantasy Football Scout Member
          • 14 Years
          8 years, 7 months ago

          Here is one for you, then:

          http://members.fantasyfootballscout.co.uk/my-stats-tables/view/7270/

          Open Controls
          1. rakkhi
            • Fantasy Football Scout Member
            • 13 Years
            8 years, 7 months ago

            Sorry Doos not really clear what I'm looking at here. Think was said on other articles but that is the weakness of the table formats here - data visualization really needed to get the understandable message. Hopefully something for Mark's todo list in the summer! 🙂

            Open Controls
            1. Doosra - ☭DeclanMyGeniusâ…
              • Fantasy Football Scout Member
              • 14 Years
              8 years, 7 months ago

              Well, these are the attributes that make for a good holding forward - a Sparky wannabe, say. Usually a good guide to the all round footballing attributes of some of the Premiership donkeys. Even Giroud has got to fifth on this list, once. 😀

              Open Controls
              1. rakkhi
                • Fantasy Football Scout Member
                • 13 Years
                8 years, 7 months ago

                Sure but does that translate to points which is surely what we are most interested in for FF right? Fair enough if you mean general football such as what player should a club recruit.

                Open Controls
                1. Doosra - ☭DeclanMyGeniusâ…
                  • Fantasy Football Scout Member
                  • 14 Years
                  8 years, 7 months ago

                  No, but I find that teams with a good one of these tend to score ...

                  Open Controls
                  1. rakkhi
                    • Fantasy Football Scout Member
                    • 13 Years
                    8 years, 7 months ago

                    Bit more complicated maths but you could test that relationship to see if the data bares it out or it is just a feeling

                    Open Controls
      2. d1sable
        • Has Moderation Rights
        8 years, 7 months ago

        No the point of stats at least to me is to use them to try and predict points for my fantasy team.

        This article helps to simplify it for those of us who don't have the time to route through them anymore and is much appreciated.

        If four seasons worth of data says this in the best indicator then sign me up

        Open Controls
      3. A.T
        • 13 Years
        8 years, 7 months ago

        Interesting stuff nonetheless. If stats can be made simple and broken down to 'one stat to rule them all' then that'll help someone like me who doesn't have a lot of confidence in them generally. I'll certainly keep an eye out for it in the future.

        Open Controls
    2. The Kolarov Mane âš”
      • 11 Years
      8 years, 7 months ago

      'til some tricksy little hobbitses toss your stat into a fiery grave...

      Open Controls
  8. d1sable
    • Has Moderation Rights
    8 years, 7 months ago

    Great work Rakkhi and much appreciated, lets hope the trends continue!

    Open Controls
  9. Pacer.
    • 9 Years
    8 years, 7 months ago

    time to get walcott

    Open Controls
    1. rakkhi
      • Fantasy Football Scout Member
      • 13 Years
      8 years, 7 months ago

      Yes!

      Open Controls
  10. KeepCalm
    • 10 Years
    8 years, 7 months ago

    Captain:

    A. Aguero (NEW)
    B. Haz (SOT)
    C. Sanchez (MUN)

    Open Controls
  11. J0E
    • Fantasy Football Scout Member
    • Has Moderation Rights
    • 14 Years
    8 years, 7 months ago

    Here's a handy Members table sorted by Rikkha's key stats - shots on target.

    Also included is Big chances, plus goal and assists and shots in the box.

    http://members.fantasyfootballscout.co.uk/public-stats-tables/view/11275/

    Open Controls
    1. Qubit
      • 14 Years
      8 years, 7 months ago

      Hi Jonty, ive asked before, but is there any chance the members area data could display shots in the box that are on target? Or perhaps it can be looked at for the future. I think it would possibly be quite useful. Cheers 🙂

      Open Controls
    2. RedLightning
      • Fantasy Football Scout Member
      • Has Moderation Rights
      • 13 Years
      8 years, 6 months ago

      That's a good table for last season, but quite a few of the players appear to have regressed since then, and not all of them will repeat last season's form. How accurate is that for predicting what will happen this season?

      Has anyone tested the stats for the first half of a season to predict goals in the second half? And might goals scored in the first half perhaps be a better predictor than at least some of the underlying stats mentioned?

      The percentages quoted are all for hindsight, and goals actually scored would beat them all with 100% correlation - but which stats work best for prediction?

      Also, it is said that outliers tend to regress towards the norm, but by how much, and is every player's norm the same? And is it not also true that all predictions of future goals, whether based on goals scored to date or on one or more of the underlying stats, will improve as the season progresses and the number of matches played increases?

      Open Controls
      1. rakkhi
        • Fantasy Football Scout Member
        • 13 Years
        8 years, 6 months ago

        It is the same for prediction a good expected goals model followed by shots on target : http://11tegen11.net/2015/01/05/the-best-predictor-for-future-performance-is-expected-goals/

        Open Controls
        1. RedLightning
          • Fantasy Football Scout Member
          • Has Moderation Rights
          • 13 Years
          8 years, 6 months ago

          Thanks, Rakkhi. Some useful graphs there. Also an interesting discussion between Tegen and Dino.

          Open Controls
      2. Jose's Magic Omelette
        • 10 Years
        8 years, 6 months ago

        Hit it spot on redlightning. I'm preparing an article to see which variable is the best predictor of future goals and assists including past goals and assists.

        Open Controls
  12. Demí
    • 13 Years
    8 years, 7 months ago

    Great stuff Rakkhi

    Open Controls
  13. FPL Daniel
    • 14 Years
    8 years, 7 months ago

    Thanks. I expected more harm to be done to us stat nerds 😛 Phew..

    Open Controls
  14. PurpleCobras
    • 9 Years
    8 years, 7 months ago

    If we're looking at defenders, would the team stat "shots on target conceded" be a good indicator of clean sheets? Is this stat readily available in the members area (I couldn't see it myself)?

    Open Controls
    1. J0E
      • Fantasy Football Scout Member
      • Has Moderation Rights
      • 14 Years
      8 years, 7 months ago

      That's a great point. I think it would.

      Open Controls
      1. Jose's Magic Omelette
        • 10 Years
        8 years, 7 months ago

        Agreed, is it possible for this to be added or to compare more than 2 players in the comparison tool? Quite a pain going through several pairs of players in the comparison tool as it is.

        Open Controls
      2. Guy Demel's SH
        • 11 Years
        8 years, 7 months ago

        Unfortunately I don't think this stat is included in the members area. I emailed the FFS support email back in April requesting it.

        Open Controls
  15. AxeMinion
    • Fantasy Football Scout Member
    • 12 Years
    8 years, 7 months ago

    What say ye:

    A) Sako (CPL) + Silva (-8pts)
    B) Dann (CPL) + Kolorov (-8pts)
    C) Silva + All Out Attack (-4pts)

    Open Controls
  16. Pacer.
    • 9 Years
    8 years, 7 months ago

    I am big fan of stats (and I think my record against friends demonstrates the advantage they hand you), but don't these results just demonstrate the blindingly obvious?

    Variability in number shots on target largely explains variability in goals scored.

    Duh!

    😉

    Open Controls
    1. rakkhi
      • Fantasy Football Scout Member
      • 13 Years
      8 years, 6 months ago

      Yup seems obvious but a lot of people me Jonty etc were using stats like shots in box to evaluate players, the big numbers article often highlights things like Penalty area touches. While all those are useful in providing a rounded picture I think what this shows is for best correlation with goals you should use shots on target and it is the change in this stat more than any other that we should pay attention to me. That was an useful insight to me at least and changes the way I play FF

      Open Controls
    2. Woy of the Wovers
      • 13 Years
      8 years, 6 months ago

      It is a bit obvious for one clear reason. It is the foremost requirement for scoring a goal. Without a shot on target, a player cannot score.

      Open Controls
      1. Pacer.
        • 9 Years
        8 years, 6 months ago

        Except for own/freak goals, but they only make up a very small % of goals.

        It's nice to see a visualisation confirm it, but common sense will tell you the relationship between goals scored and the stats will get stronger through a typical sequence attack sequence

        through-balls result in passes received in the final 3rd
        PRITF3rd result in touches in the box
        touches in the box lead to (or constitute) shots in the box
        shots in the box lead to shots on goal
        shots on goal lead to a goal

        and as you say, it's almost inevitable that shots on target will have the strongest relationship

        Open Controls
        1. Dino
          • Fantasy Football Scout Member
          • 14 Years
          8 years, 6 months ago

          Sorry but this kind of statement is why the likes of this article needed to be posted. Passes received in the final third, touches in the box etc have very little correlation with goals scored yet you mention them as if it they are patently obvious indicators just like shots on target.

          Too many people have been spouting rubbish on here saying x player is due to create chances because he has so many chances created so far or is going to score goals because he has so many penalty box touches or so many shots. What this article shows is shots on target and possibly shots in the box (although the quality of the chance from which the shot is taken can make this stay very misleading) as well as big chances are the only real indicators of whether a player is "due". Even these metrics are wildly variable as it depends on the teams the players were playing when the stats were gained with no guarantee that the same output well be possible v different opposition.

          Tldr stats are good but context is key

          Open Controls
  17. Barry Woj
    • 11 Years
    8 years, 7 months ago

    People who like this article may enjoy Paul Riley's work. He's made an expected goals model based on the location of shots on target. Every shot on target is assigned a probability of scoring based purely on its location.

    Here is an explanation of his model:
    https://differentgame.wordpress.com/2014/05/19/a-shooting-model-an-expglanation-and-application/

    Here is a public tableau of his data. If you click through the players and hover over the shot locations you can see the expected goal value for their shots as well as their ranking.
    https://public.tableau.com/profile/paul.riley#!/vizhome/PremierLeague201516xGMap/PremierLeague201516ShotonTargetxGDashboard

    Open Controls
    1. Barry Woj
      • 11 Years
      8 years, 7 months ago

      Oh and thanks for the great article, keep up the good work!

      Open Controls
    2. Cowboy John
      • 11 Years
      8 years, 7 months ago

      Thanks for the links - really interesting stuff. 🙂

      Open Controls
      1. Barry Woj
        • 11 Years
        8 years, 6 months ago

        No worries mate, just wish I could come up with some of the stuff on my own!

        Open Controls
        1. rakkhi
          • Fantasy Football Scout Member
          • 13 Years
          8 years, 6 months ago

          Yup Paul Riley's work is great. For us FF fanatics he just doesn't update his expected goals model that fast week to week. I really want to build an expected goals model but the shots with x y coordinate data is quite expensive for Opta

          Open Controls
          1. Barry Woj
            • 11 Years
            8 years, 6 months ago

            Ah crap I thought he did! It's still a good thing to check periodically I guess. I also enjoy Michael Caley's xG diagrams after the games, I find they give you a better idea of what went on than most match reports.

            Yeah that would be great, I suppose copying the information from the FFS player profiles would be far too labour intensive? If you have a viable way of doing it, you could surely contact Mark and the guys at FFS? I would have thought they have the raw opta data in a more usable format and it would be a metric that benefits everyone on the site.

            Open Controls
            1. rakkhi
              • Fantasy Football Scout Member
              • 13 Years
              8 years, 6 months ago

              Yeah the stats Mark pays Opta for I doubt includes the x y coordinates of the shots. That is in their performance premium package. I got a price for it and will email Mark and Jonty maybe it is something they are can do. Even the FFS player profiles doesn't have that data

              Open Controls
              1. Barry Woj
                • 11 Years
                8 years, 6 months ago

                So how are they producing the images if you do this:

                Player Profile > Show > Goal attempts (Shots on target in green).

                Open Controls
  18. Chicken-Mandy
    • Fantasy Football Scout Member
    • 8 Years
    8 years, 7 months ago

    Would you play Darmian (Arsenal away) or use AOA and play Scott Sinclair (Stoke home)?

    Open Controls
  19. mm0507
    • 10 Years
    8 years, 7 months ago

    So, maybe you make one big formula called chance to score that will be:

    0.77*SoT + 0.7*BC + 0.69*SiB + 0.66*Sh....

    And I have one question, big chances are all chances involved in (created+missed+scored) or without created ones?

    Great article. Thanks

    Open Controls
    1. rakkhi
      • Fantasy Football Scout Member
      • 13 Years
      8 years, 6 months ago

      Yeah the problem is getting those weighing right. You could back test all the variables but that is a fair bit of work. There is always the temptation to add more complexity but you then also introduce model error.

      Big chances only relate to goals. They are where the Opta trackers judge as a very high chance to score shot. Bit of human judgement but the data seems to have a high correlation to actual goals. Measures both scored and missed but not chances created.

      Open Controls
      1. mm0507
        • 10 Years
        8 years, 6 months ago

        Yeah, I get it. You can divide every addend by league total (LT) (because there is more shots than shots in the box) to get all weighing right. Then you have something like this:

        0.77*Sot/LTSot + 0.7*BC/LTBC + 0.69*SiB/LTSiB + 0.66*Sh/LTSh....

        If you divide all this with 0.77 + 0.7 + 0.69 + 0.66... and multiple by league total goal scored you can get something like expected number of goals scored for every player.

        Thanks for response.

        Open Controls
  20. Bezzer
    • 8 Years
    8 years, 7 months ago

    Cool stuff.
    This just makes me want to get Walcott and Diouf back in!

    Open Controls
    1. Teddy Brewski
      • 9 Years
      8 years, 7 months ago

      Diouf may miss one or two games coming up though with Senegal

      Open Controls
      1. Bezzer
        • 8 Years
        8 years, 7 months ago

        Ah! Great shout Cheers Teddy!

        Open Controls
      2. Apprentice-1
        • 8 Years
        8 years, 6 months ago

        Thanks for the info, do you know fir which GW this might be? I just bought him..

        Open Controls
  21. Guy Demel's SH
    • 11 Years
    8 years, 7 months ago

    Great article, thanks for this. Confirms what I thought too.

    Sorting The3rdTurds bubble graph to SoT and SiB on each axis gives you a good view of who is due a few goals.

    Open Controls
  22. Diva
    • 9 Years
    8 years, 7 months ago

    Thanks for the article. It's reconfirmation that the stat I lean on most for assessing player goal threat is still the best.

    Open Controls
  23. Jose's Magic Omelette
    • 10 Years
    8 years, 7 months ago

    Really like this article. On question, over four seasons there will certainly have been players falling away from the trend line, were there any as far away from the linear trends as Walcott/Agüero are currently?

    Open Controls
    1. Jose's Magic Omelette
      • 10 Years
      8 years, 7 months ago

      *one question

      Open Controls
    2. rakkhi
      • Fantasy Football Scout Member
      • 13 Years
      8 years, 6 months ago

      I only looked at the total season stats for 4 seasons but yes there were players quite far from the trend line would need to go back and see what the largest gaps were over a whole season

      Open Controls
      1. Jose's Magic Omelette
        • 10 Years
        8 years, 6 months ago

        The reason I ask is that understanding how common it is for players to be that far away from the trend line over a whole season compared to over a short period of a few weeks will help us to be more or less confident about players like Agüero and Walcott 'catching up' on goals.

        Open Controls
  24. Jose's Magic Omelette
    • 10 Years
    8 years, 7 months ago

    Also if you're going to do any future articles I would be really interested to see this broken down per position.

    For example for midfielders are shots on target closer to the regression line for goals than chances created are for assists?

    Open Controls
    1. rakkhi
      • Fantasy Football Scout Member
      • 13 Years
      8 years, 6 months ago

      Yup not that hard to do. I'll take a look

      Open Controls
  25. mookie
    • 10 Years
    8 years, 7 months ago

    Thanks Rakkhi.
    So... both Martial and Vardy are overachieving while my strikers are top of the underachievers camp.

    Open Controls
    1. rakkhi
      • Fantasy Football Scout Member
      • 13 Years
      8 years, 6 months ago

      Yeah so some regression to the mean can be expected the timing is the challenge. Vardy though has a high shots on target to his goals so he could well continue his good form

      Open Controls
  26. get_in
    • Fantasy Football Scout Member
    • 11 Years
    8 years, 6 months ago

    I've done some dickering around this evening based on the principles in the article. I was interested to see if there was a weighted combination of stats that could be used to hone in on correlation. Looking at the last four seasons, ICT index weighted by % goal involvement comes out with > 93% R-squared vs total goals and assists, > 85% R-squared vs goals scored.

    Early yet in the season to be using % involvement, but weighted ICT has the following players as the current top 10 to watch:

    Mahrez
    Ighalo
    Mata
    Sako
    Pellè
    Lukaku
    Morrison
    Payet
    Sánchez (Alexis)
    Vardy

    Nothing revolutionary here, but thought it was worth sharing.

    Open Controls
    1. Woy of the Wovers
      • 13 Years
      8 years, 6 months ago

      The problem with goal involvement is that it is a direct result of scoring goals. And you'd get perfect correlation for goals v goals.

      Open Controls
    2. Jose's Magic Omelette
      • 10 Years
      8 years, 6 months ago

      Wondered if this might be the case, thanks for doing the analysis!

      Open Controls
  27. Scotty B
    • 12 Years
    8 years, 6 months ago

    If I get rid of Yaya I will have 10.2 to spend, do I get:-

    A - KDB
    B - Silva

    Thanks

    Open Controls
  28. Sterling Chap
    • 9 Years
    8 years, 6 months ago

    Great article

    Open Controls
  29. Sjaugen
    • 8 Years
    8 years, 6 months ago

    How do I upvote an article?
    Love this one

    Open Controls
  30. Gargamel
    • Fantasy Football Scout Member
    • 14 Years
    8 years, 6 months ago

    Hey Rakkhi, awesome article.

    Open Controls

You need to be logged in to post a comment.