The difficulty of comparing/ranking players

Discussion in 'The Beautiful Game' started by schwuppe, Jan 3, 2019.

  1. schwuppe

    schwuppe Member+

    Sep 17, 2009
    Club:
    FC Kryvbas Kryvyi Rih
    The bread and butter of every football discussion and it always seems everybody (me included) has a high level of confidence in his opinion about certain players. To clarify: all this is concerning player performances and not their actual skillset.

    As someone who loved researching and learning more about certain players I figured out the most accurate (but very time consuming) method would be to actually go through every single career game and keep track of your finding about it somewhere.

    So this is kind of a long term "project" for me right now and not even close to finished, however I already learned a valuable lesson.
    “The more you know, the more you know you don't know.” or however that saying goes would be fitting to describe it.

    After spending some time on certain players - some which I had very strong opinions and was very certain about (examples are Rivaldo/Figo or Iniesta/Xavi ) - I'm seriously reconsidering that they were wrong all along. And those are high exposure players who played recently, careers that I actively followed and I watched them hundred of times!

    So what do I actually know about how good some semi-famous players from the 60s were? I have no fecking clue.
    Unless we are talking about the extremly obvious standout players like Pele and Maradona (let's be honest 99.9% of people also know feckall about those players outside of the WC, but you don't need to go into details to figure out that they are at the top of the pyramid) or recently Messi and Ronaldo most stuff just looks like slightly better guesses to me. Once we go down further the list and reach the end of the top 100 alltime I'd give every player ranked a margin of error of +/- 50 spots or so, getting larger the further down we go.

    Furthermore, an additional thing I'm doing is keeping track of current player performances using all possible material I could think of that might be useful (watching games, player ratings from several different sources, stat based player ratings like whoscored/sofascore, reading writeups, regular stats)
    It made me realize that unless you have an eidetic memory and watch every single game it's impossible to make a decent ranking besides the handful of clearly outstanding candidates. Heck even with all that information some players are just so close that several choices are virtually tied and would seem reasonable.
    Using whoscored only might be alright for attackers, but I feel it doesn't work that well for midfielders, defenders and goalkeepers.

    Still I can tell you for sure that if I would post my findings on a very active discussion board like reddit people would dismiss them as garbage and stand by the opinion they already hold. I assume a lot would be just supporters of certain clubs who feel their players are underrated, another thing that always draws scrunity is ranking no name players who fly under the radar (both playing for a small team and statswise because of their role), but are actually playing well using the "eyetest".

    So what's my point? I think that people in general (me included) heavily overestimate how much they know about players and to make a decent ranking of current players A LOT of time investment is needed. Now if we're talking about an alltime ranking I can't imagine how many years (more like decades - I mean look at my registation date) it would take to make one I would be satisfied with and confident in calling pretty good.
     
    Sexy Beast and comme repped this.
  2. Ariaga II

    Ariaga II Member

    Dec 8, 2018
    The advantage of player-ranking is that it isn't an exact science. You can get away with most anything if you can back your views up properly. Once you start doing an all-time ranking or comparing players from different positions it gets completely arbitrary, anyway. It doesn't really matter if some fanboy gets triggered because the #36 in your all-time list should actually be in the top 30. There's a reason I'm doing my own rankings in a rough "these are the world class players" method instead of a straight top 50 or something.

    I think of player-rankings as a fun hobby. You can do it with the knowledge you have at your disposal at the time, and then revise it when you find new information. It's a bit of a never-ending road, but hey, would you wanna give up your hobby, anyway?
     
    Sexy Beast, ManiacButcher and schwuppe repped this.
  3. PuckVanHeel

    PuckVanHeel Member+

    Oct 4, 2011
    Club:
    Feyenoord
    I'm interested in what you've collected/processed for these players and what your ideas are.

    (I'm in the Figo and Xavi camp :D )
     
  4. JoCryuff98

    JoCryuff98 Member+

    Barcelona
    Germany
    Jan 3, 2018
    I have high memory, but unfortunately I haven’t watched many 60s and early 70s matches barring from Cruyff. In fact, due to lack of footages it’s impossible to view those old era matches. Footballia is doing a pretty good job uploading those black and white games.
     
  5. comme

    comme Moderator
    Staff Member

    Feb 21, 2003
    Inherently it's an impossible job. Football is so subjective and there is so much of it that it's really not possible to come up with anything that can really be considered robust.

    We're trying to compare people across different roles in different eras and different circumstances.

    At the same time it's a lot of fun. I'm just starting a new project to go back through the Ballon d'Or (like Tom Stevens and PeruFC have done), watching more matches, going back through contemporary reports and then writing it all up. A lot of it will be subjective and guess work but will hopefully give me a better view on it all.
     
  6. comme

    comme Moderator
    Staff Member

    Feb 21, 2003
    Just as an example of this, we can't even get any consensus over a single game.

    Pep Guardiola just said of Bernardo Silva's performance against Liverpool last night "Bernardo is incredible. I've not seen a performance like that, in all terms, in a long time."

    Personally I thought he was excellent. He got a 6 in the Guardian. For context:

    Guardian ratings:

    Ederson 6; Danilo .6, John Stones 7, Vincent Kompany 4, Aymeric Laporte 5, Fernandinho 8, David Silva 5, Bernardo Silva 6, Raheem Sterling 8, Leroy Sané 9, Sergio Agüero 9

    L'Equipe Ratings

    Ederson 6; Danilo 5, Stones 6, Kompany 7, Laporte 6, Fernandinho 8, David Silva 5, Bernardo 7, Sterling 6, Sane 7, Aguero 6

    If two experienced observers can't even come remotely close to each other on a game happening now, we have little hope for assessing players from 50+ years ago.
     
    poetgooner, carlito86 and Edhardy repped this.
  7. carlito86

    carlito86 Member+

    Jan 11, 2016
    Club:
    Real Madrid
    That is why statistics are objective and perception is subjective
    Guardiola maybe a credible observer however in this case being his manager he is a biased one

    If Bernardo had an average game he’d call it a good performance and if he actually had a good performance Guardiola would call it phenomenal
    Really it’s just a boost of confidence for him to hear his manager say great things about him

    Guardiola did say Messi was the greatest player of all time in 2009.
    Hardly anybody out side of the Barcelona camp shared this view.
    Even if Messi did turn out to be exactly that it is irrelevant because Guardiola wasn’t talking about potential but his current state

    It also depends on how highly you rate tangible aspects of a players performance over intangible ones
    A goalscorer,especially one who scored a goal with a degree of technical difficulty(ie aguero) could rank higher than someone who scored a easy finish(ie firmino)
    A different observer would be in his right to say for example a winning dribbling goal vs 5 or 6 guys isn’t objectively worth more than a winning tapin goal

    Also we don’t know the criteria different sources use for rating players
    There seems to be a consensus about Fernandinho and I don’t think anybody who actually watched the game would differ about how influential he was
    Agueros match ratings seem to be boosted by the fact he was on the winning team
    Outside of the goal I don’t remember him exactly running riot yesterday albeit he did seem to be a threat at times
     
  8. comme

    comme Moderator
    Staff Member

    Feb 21, 2003
    Just to expand on this, the assessment of historical players is an even harder task because we are having to take so many things into consideration.

    Beyond simply rating the 22 players on the field over 90 minutes, we are trying to make assessments and take into account the pitch, the ball, the general conditions. Then we're trying to make a judgement about the standard of the game, the quality of the opponents and their teammates.

    That is a hell of a lot of factors to take in.

    Meanwhile we have almost none of the factual or statistical anchors to help judge quality and even where we do, they don't give us a balanced picture. We can't use times like people in athletics can, we can't use averages like in cricket or baseball, we can't use raw stats very well like in basketball.

    All the while some people will uphold the pretence that the absolute standard in the past was just as good as it is today.
     
    schwuppe repped this.
  9. PuckVanHeel

    PuckVanHeel Member+

    Oct 4, 2011
    Club:
    Feyenoord
    I read a column by Matthew Syed where he claims OPTA data show passes per game have gone up by 200 compared to 15 years ago, and players cover 2.5km more distance per match compared to 5 years ago (which to me sounds improbable).

    Is there an original source for this? Would like to see that.
     
  10. comme

    comme Moderator
    Staff Member

    Feb 21, 2003
    One for Duncan Alexander on Twitter I would say. He is normally very helpful.
     
  11. PuckVanHeel

    PuckVanHeel Member+

    Oct 4, 2011
    Club:
    Feyenoord
    Of course no reply to this.

    Shame.
     
  12. schwuppe

    schwuppe Member+

    Sep 17, 2009
    Club:
    FC Kryvbas Kryvyi Rih
    I will post everything when it's finished, but essentiell I underestimated 1) How much longevity edge on a top level Figo has 2) How inconsistent Rivaldo was (this was already something known, but I wasn't sure to which degree)

    In the the second case it slipped my mind that Iniesta already played a decent part in the 2006 CL win and was considered a top level player much earlier than Xavi, who was good, but not really great for a long time (or underrated due to playstyle?). The team situation might play a big part of it as Barca went through a difficult phase when Xavi emerged, so it might have been just luck.
     
    PuckVanHeel repped this.
  13. schwuppe

    schwuppe Member+

    Sep 17, 2009
    Club:
    FC Kryvbas Kryvyi Rih
    #13 schwuppe, Jan 8, 2019
    Last edited: Jan 8, 2019
    Your point in general is ofc correct, but quite frankly from my experience so far ratings from UK magazines tend to be pretty random. Some more than others, especially tabloids (Mirror you might as well just roll a dice)

    This is what I have tracked which would clearly point to him having a really good game.

    Allison however looked rather interesting....
    (for clarification if anyone isn't familiar with those magazines: FF means France Football, LE L'Equipe, GdS Gazzetta dello Sport, CdS Corriere, Tutto Tuttosport)
     

    Attached Files:

    PuckVanHeel repped this.
  14. schwuppe

    schwuppe Member+

    Sep 17, 2009
    Club:
    FC Kryvbas Kryvyi Rih
    Yes, I have to say Tom Stevens and PeruFC actually did a great job. Tier rankings are doable, direct comparisons get really difficult.

    The thing is there is A LOT of possible information available that wasn't uncovered (on here at least.... - I found some interesting stuff from the 50s I will post soon) and imo (even if it remains subjective ) a pretty good (at least era specific) ranking is possible. Italy in particular seems to be well documented and offers a lot to work with.

    I'm also interested in other sports and participated in various rankings, Football really stands out to me as particularly difficult.

    You already mentioned the stats thing. Not only aren't they really documented (except for goals which might only be decent for attackers), stats also seem more useful for other sports. Just looking at goals and assists (they track secondary assists since the 50s or something) for a Hockey forward is common practice.

    Another problem which makes ranking football players so hard in comparison is that there are 5 - 15 or so different leagues (helps if you can speak different languages...) worth researching at any point which change in quality all the time. Then there's the WC, Euro, CA, UC, CWC, ICFC and God knows what kind of cups/regional leagues South America has.

    If you're attempting to make a Baseball, Basketball, Hockey.. ranking about 90% of the relevant stuff happened in the same league.

    Those leagues also had all-star teams, awards, rankings etc. dating far back. Sources like newspapers are easier to get and you'll be fine with English only.
     
  15. schwuppe

    schwuppe Member+

    Sep 17, 2009
    Club:
    FC Kryvbas Kryvyi Rih
    I'm sorry for the delay.
     
  16. PuckVanHeel

    PuckVanHeel Member+

    Oct 4, 2011
    Club:
    Feyenoord
    How have you gotten access to papers like Telegraph that are behind a hard paywall?
     
  17. PuckVanHeel

    PuckVanHeel Member+

    Oct 4, 2011
    Club:
    Feyenoord
    Disagree for Peru FC.
     
  18. PuckVanHeel

    PuckVanHeel Member+

    Oct 4, 2011
    Club:
    Feyenoord
    Thanks. Yes, ratings have always been very favorable to Iniesta. Less so for Xavi but there are some who (at hindsight) believe he already reached a world class level between 2003-2005 (in 2005-06 he was injured ofc).

    In general it is accurate to say ratings were less favorable to him, from what I've seen. Don Balon graded Iniesta a lot higher (up to being seen as the best of his team) than Xavi between 2005-2008 (before Guardiola). Pravda paper Mundo Deportivo too I think.


    My take on Rivaldo in relation to his team-mates is here (I agree 100% with you):

    http://www.bigsoccer.com/threads/the-most-underrated-players-you-have-watched.2087099/#post-37136655
    http://www.bigsoccer.com/threads/th...you-have-watched.2087099/page-2#post-37145165
    http://www.bigsoccer.com/threads/th...you-have-watched.2087099/page-2#post-37144443

    One of the main things with him is that on first sight his stats look amazing (for his time), but at a second glance the edge erases a bit.
     
  19. schwuppe

    schwuppe Member+

    Sep 17, 2009
    Club:
    FC Kryvbas Kryvyi Rih
    #19 schwuppe, Jan 10, 2019
    Last edited: Jan 10, 2019
    Why? Does he underrate Dutch players? I'm semi-serious here, but I do think some Dutch players look underrated historically. Krol for example?

    I didn't look at it too closely tbh, but it seems better than Dearman's effort.
     
  20. schwuppe

    schwuppe Member+

    Sep 17, 2009
    Club:
    FC Kryvbas Kryvyi Rih
    #20 schwuppe, Jan 10, 2019
    Last edited: Jan 10, 2019
    Weren't there talks that Barca wanted to get rid of Xavi before the 2008? I don't know how the rumour started though.

    Barcelona's had a terrible CL record at that time in general. 3 first group stage exits in 5 years from 97/98 - 01/02. The Bayern/United and maybe Leeds/Milan group was easy, but the 97/98 showing was embarrassing. They were 2nd in ELO at that time. I don't know what bookies odds were, a team like Kiev (outside top 100 at the start, rising quickly) was surely underrated because it was the start of their first great season.

    Rivaldo had a bunch of top tier performances like vs United and Milan Away, but a lot of times he disappeared and doesn't really have a single great KO game against strong teams.

    Figo isn't setting the world on fire either, during his CL winning campaign in 01/02 it looks like he was one of the weakest Real players.
     
    PuckVanHeel repped this.
  21. schwuppe

    schwuppe Member+

    Sep 17, 2009
    Club:
    FC Kryvbas Kryvyi Rih
    #21 schwuppe, Jan 10, 2019
    Last edited: Jan 10, 2019
    https://www.pressreader.com/

    It goes back to Oct 2005 for Telegraph. It is a paywall, but you get access for 2500 newspapers goining back to 2004 sometimes.
     
  22. PuckVanHeel

    PuckVanHeel Member+

    Oct 4, 2011
    Club:
    Feyenoord
    That is one part of it, but probably a side effect of priorities elsewhere. There is a tendency to take something away after the passage of time, despite at the time itself media/sponsors etc. already naturally being geared towards the bigger markets and established powers. It is a double knock-on effect.

    It is more like South American players get easily a free pass, for example when they are still playing in SA kids leagues (and don't put the national team on fire) or - irony - still play in the Eredivisie, including Suarez.

    The main problem/issue I have though is that it often doesn't square with other parts of the equation. Taken all parts together, it doesn't add up well imho.


    Yes in the main I agree with this. In his first two seasons there were two first group stage exits, followed by three semi finals (one of them in the UEFA Cup, after exiting the group stage).

    There were some legit reasons or excuses for what happened (injuries, suspensions, players arriving, clownesk goalkeeper, starting CL matches with 7 bench players etc.) but main problem is Rivaldo was overall significantly outproduced and outperformed by a couple of his team-mates in the 2nd group stages and knock-out stages (UEFA Cup included). With the freedom he had he should have been above his team-mates, that was his role.

    His national team career is of course great, no doubt, although his 1998 Denmark game is a level above the rest he showed (against teams of similar or higher level).

    Figo his best CL campaign is 1999-00 quite clearly in my perception.
     
  23. PuckVanHeel

    PuckVanHeel Member+

    Oct 4, 2011
    Club:
    Feyenoord
    From what I remembered I thought his 'Football's Greatest' episode (from about that point on) is quite okay on this (oh, now see accidentally another good section around the 8:00 minute mark, referring to team situation).
     
  24. comme

    comme Moderator
    Staff Member

    Feb 21, 2003
    Yes, I agree the English ones are poor and I've spoken previously on here about the way it's largely an afterthought in England.

    To be honest, during the World Cup I thought the ones from Kicker and France Football were pretty poor as well, or at least they diverged very heavily from my own ratings.

    The French ones have a tendency to go over the top in their ratings of bad performances. For instance I think L'Equipe gave Lovren a 2 for the game against Man City and while he was poor, he wasn't awful to my mind. A display meriting a 2 shouldn't be merely poor, it should be utterly inept.
     
  25. schwuppe

    schwuppe Member+

    Sep 17, 2009
    Club:
    FC Kryvbas Kryvyi Rih
    #25 schwuppe, Jan 10, 2019
    Last edited: Jan 10, 2019
    Need to check out kicker a bit more. From what I know their average is pretty much exactly 3.5, which seems counterintuitive as 3 is the average grade per definition in the German school system.

    True.

    I wanted to make a post about this at some point.
    Essentially their scale works in a different way. For Italian's an average performance starts with 6, while in France it's 5. This seems universal with France Football, L'Equipe and Le Parisien.

    They also tend to use the lower half much more frequently, while getting a higher grade like 7 or 8 seems to be have the same likeliness as with GdS/CdSp. A bit like people tend to do on fansites like redcafe or something, who love giving out 1s or 2s.

    So what can be done is to adjust them to Italian ratings to make them directly comparable (done with excel "regression" function). The result of this is that anything from 1-2 are a 4, 3 is 4.5, 4 is 5, 5 is 6, 6 is a 6.5 and the higher grades stay the same.
    This probably isn't scientific in any sense, but from what I've seen after adjusting them they mostly pass the "smell test" and are quite useful. A 4 here seems harsh for Lovren too though.

    If I try doing that with "Sport" or "Mirror" the results are just random with close to 0 correlation.

    The only situation were this is problematic is when a player would get a rating lower than 4 from GdS, but this is incredibly rare and glaringly obvious that the player was horrendous.

    This is all under the assumption Italian ratings are the most correct ones, which is ofc subjective.
    An idea how to assess that would be to test whether the team with the most points, gets the highest ratings over the long run. If a team that gets relegated always tends to have higher rated players than the league winners there is obviously something wrong with the ratings.
     
    msioux75 repped this.

Share This Page