FIFA World Ranking

Discussion in 'Women's International' started by jonny63, Mar 17, 2006.

  1. SiberianThunderT

    Sep 21, 2008
    DC
    Club:
    Saint Louis Athletica
    Nat'l Team:
    Spain
    I'm glad USA-CRC was a friendly and not a competitive match because it'll "only" be a loss of ~7pts... I've never before seen such a dominant performance end in a nil-nil draw.
     
    blissett repped this.
  2. lil_one

    lil_one Member+

    Nov 26, 2013
    Nat'l Team:
    United States
    Certainly the most dominant performance from the USWNT without a win in at least a decade:

     
    blissett repped this.
  3. kolabear

    kolabear Member+

    Nov 10, 2006
    los angeles
    Nat'l Team:
    United States
    Reminder heading into the Olympics — I would add 250 to 300 points to the African representatives to try to get a more realistic idea of what to expect during the tournament. Adding 300 points to the last published rating for Zambia gives them an provisional rating of 1697, which puts them just ahead of New Zealand's 1680.

    Adding 300 to Nigeria's rating gives them a provisional rating of 1916 putting them just slightly ahead of Australia's 1890.

    It's also not far behind Japan's 1976 and Brazil's 1959. If that provisional rating is realistic, it makes Group C really strong and unpredictable.

    ***
    The estimated homefield advantage of 100 puts France effectively at the top of the list by a small margin , 2130 points to Spain's 2100
     
  4. fire123

    fire123 Member+

    Jul 31, 2009
    Can someone calculate the ratings gained by the US recently including the Olympics?
     
  5. SiberianThunderT

    Sep 21, 2008
    DC
    Club:
    Saint Louis Athletica
    Nat'l Team:
    Spain
    #1705 SiberianThunderT, Aug 10, 2024
    Last edited: Aug 10, 2024
    Back of the envelope:
    Each match worth up to 60pts
    1-goal wins worth ~0.85 of the result; 3-goal wins worth ~0.96
    USA favored in every match except the first match against GER
    Let's assume that the ZAM and AUS matches were roughly break-even versus "expected", so ~0pts gained (or lost)
    First match against GER: 60*(~0.96 - ~0.45) = ~30pts gained
    Other three matches, estimate 60*(0.85-0.65) = ~12pts gained each
    My guess would be a gain of ~66pts on just the Olympic matches themselves
    Not quite enough to break 2100, but it's possible USA retakes #1 since I bet Spain's 3 expected wins are outweighed by the draw and loss to much-lower-ranked sides.
    BIG asterisk on that, though, since I'm not considering any team's warm-up matches right now.
     
    kolabear repped this.
  6. SiberianThunderT

    Sep 21, 2008
    DC
    Club:
    Saint Louis Athletica
    Nat'l Team:
    Spain
    #1706 SiberianThunderT, Aug 10, 2024
    Last edited: Aug 10, 2024
    Just checked the Cambio De Juego Twitter account, and they've confirmed (yesterday) that USA would've been #1 in the next ranking regardless of the result of the Gold medal match. Was a gain of 46pts (some points from AUS but losses in the warm-up matches) before today, so maybe around 56pts total instead of the 66pts I estimated.

    USA, CAN, and BRA with huge gains. ESP, FRA, and AUS the big losers of this OG tournament...
     
    lil_one, Lohmann, blissett and 1 other person repped this.
  7. fire123

    fire123 Member+

    Jul 31, 2009
    Wow, just wow! Emma deserves that drink and a few more!!
    I think I will join her !
     
  8. shlj

    shlj Member+

    Apr 16, 2007
    London
    Club:
    FC Nantes
    Nat'l Team:
    France
    We are going to lose 8 places from 2nd to 10th.Incredible and deserved freefall.
     
    BlueCrimson, toad455 and blissett repped this.
  9. toad455

    toad455 Member+

    Nov 28, 2005
    When do the next rankings come out?
     
  10. GoodHands

    GoodHands Member

    AC Milan
    Italy
    Jul 17, 2024
    My main question for this thread is: Why does it matter what some useless poll/rating system says about any team?

    The "ratings" are simple a way for uninvolved people to rank things (which people seem to REALLY like to do) and make thinking lazier. The ratings are, unfortunately, used for things like seedings but even that does not matter much. (To win a tournament you must beat the second best team or beat whoever beats them and it does not matter what their ranking are.
    If they had any real meaning then they would be used by teams to judge their strength but they give a view of the position of a team that keeps moving up/down/sideways and loses all meaning in the real world very fast. Matches that are way too old are counted too much and team rankings matter not on the current strength but on strength exhibited in the past with largely different teams in most cases.

    The bottom line is that international ratings are of roughly the same value as the teats on a boar hog.
     
  11. cpthomas

    cpthomas BigSoccer Supporter

    Portland Thorns
    United States
    Jan 10, 2008
    Portland, Oregon
    Nat'l Team:
    United States
    My experience comparing how well rating systems do at predicting future outcomes in relation to how well humans do is that the ratings do better than most humans. There will, however, be some humans that do better. I attribute this two one of two factors: (1) luck or (2) true expertise combined with detailed knowledge.

    When it comes to exact accuracy at predicting outcomes, on the other hand, what good ratings systems do is tell you likelihoods, not exact outcomes. This is the case with the system FIFA uses. So, when two teams play, a good rating system can tell you Team A has an X% win likelihood, a Y% tie likelihood, and a Z% loss likelihood. These likelihoods are quite reliable, as they are based on very large data bases.

    This is why bookmakers use both rating systems and detailed current information.

    My experience also is that a good number of people do not like rating systems. They prefer their own judgments and think they are objective.
     
    JanBalk, SiberianThunderT and kolabear repped this.
  12. kolabear

    kolabear Member+

    Nov 10, 2006
    los angeles
    Nat'l Team:
    United States
    Friday, I believe
    You'll understand (won't you?) if I don't try to give a comprehensive answer to a topic some of us have discussed for years, but —
    • The FIFA ratings for women is a variation of the Elo-Ratings for chess, a system which has proven popular over decades and which many other games besides chess have adapted
    • An Elo or chess rating system is objective, it's not based on people's subjective opinions on this or that team but is calculated, objectively, based on team's results, primarily win / lose / draw — although the FIFA system also uses game scores, ie how many goals a team wins by, an adaptation which seems to work for very low-scoring games like soccer.
    • It is not merely a ranking system but a rating system. That is the ratings have meaning, not just the rankings. Two teams can be ranked one after the other but have ratings differences which are far apart meaning one team is significantly superior to the other. Conversely, two teams can have ratings so close to each other that the fact one is ranked above another is statistically meaningless
    • The system is probability-based. For example, one team rated 100 points higher than another has an expected win pct. of about (.640) or 64% over the other team. A team rated 200 points higher has an expected win pct of about (.760) or 76%. Of course it's not exact — it's sports and upsets happen. The lower-rated team is expected to upset a higher-rated team, but naturally not as often when it is 200 or 300 or 400 etc points lower than if it was only 50 or 100 points lower

    I'll talk a bit more about how the rating system can be useful in another post.
     
    cpthomas and toad455 repped this.
  13. SiberianThunderT

    Sep 21, 2008
    DC
    Club:
    Saint Louis Athletica
    Nat'l Team:
    Spain
    There's a lot wrong with your post but I'll focus on this one thing for now, as the two posters above me have already covered a lot of ground.
    This quoted claim is just outright false. The nature of an Elo-based system is that the older a data point is, the less it matters. The rankings are designed so that you only need 20-30 matches for the rating of each team to be considered roughly accurate, and for top teams that means you can essentially ignore matches older than 2-3 years. Sure, rosters can change a bit faster than that, but short of an entire team's worth of players retiring (or boycotting) all at once, the core of the team doesn't change that quickly. Continuity matters. So no, you don't understand the ratings and are just coming off either as upset at what they currently look like and/or just generally distrustful on anything quantitative.
    It's true that no model is perfect. But they can still tell you a damn lot of reliable information when they're built well.
     
  14. fire123

    fire123 Member+

    Jul 31, 2009
    All good points have been made above. I just want to add , I am interested in my beloved USWNT to get back to the top rank so the idiotic mass would just shut up and stop attacking them.
     
  15. GoodHands

    GoodHands Member

    AC Milan
    Italy
    Jul 17, 2024
    Any match that gets considered in the rankings that are more than a year old are meaningless. The football ranking systems are pretty much the same as political polls. That is the almost never give even a close approximation to reality. They exist to make jobs for number crunchers and give the masses something to talk about but most of those that use them really delude themselves into thinking they have meaning. They all use statistics and that, in itself, removes most meaning::
    "Most people use statistics like a drunkard uses a lamppost, more for support than illumination." - Mark Twain


    While I respect the posts above I strongly disagree about rankings having any real value.
    I probably should not have posted to this thread but, sometimes, my desire to make a point wipes out my good sense. Sorry.
     
  16. SiberianThunderT

    Sep 21, 2008
    DC
    Club:
    Saint Louis Athletica
    Nat'l Team:
    Spain
    That's.... no. Teams flipping strength on a dime is incredibly rare. If a player manages to become a consistent starter on their NT, that's going to be at the very least several years on the team (assuming no injuries) and that will be a through point for the team overall; consider that you often have large cores of players sticking around, and the results of last year will be highly correlated with results this year - when you look at all teams over all years. Again, there will always be individual exceptions here and there, but those don't invalidate the system overall. Think "the exception that proves the rule".
    Then you really should have read more of the thread before commenting. Every year, the people here will do post-mortems on major tournaments and consistently show that the pre-tournament ratings were highly predictive taken overall. Again, there will always be an upset or two, (and usually a systematic bias for the CAF teams that we take into account in our analyses,) and yes that means that the "favorite" doesn't always win (in fact they rarely do, because that's how tournaments work), but over however many dozens of games there are in each tournament, most go the way you would expect based on the ratings coming in.
     
  17. almango

    almango Member+

    Sydney FC
    Australia
    Nov 29, 2004
    Bulli, Australia
    Club:
    Sydney FC
    Nat'l Team:
    Australia
    In one sense they have no value at all because we do not award trophies or prizes based on them. We also (at least to my limited knowledge) do not use them as a means to decide who plays in a tournament. We use them as talking points but their main use is as a seeding tool when doing draws for qualification to tournaments and for the tournaments themselves. Even before rankings use was widespread as a seeding tool we often used a more simpler system for seeding teams, often based on recent results or from the last tournament. I think the current systems that FIFA use tend to be more accurate than some of the past ones which could be manipulated more easily. Your post stimulated some discussion which contained some good information for someone not hugely knowledgeable about how they work so I'm glad you posted. With some factual information people can form a more considered on the value of rankings.
     
  18. cpthomas

    cpthomas BigSoccer Supporter

    Portland Thorns
    United States
    Jan 10, 2008
    Portland, Oregon
    Nat'l Team:
    United States
    To add to the discussion about rankings, here is some information from my experience with US NCAA Division I women's soccer, for which the NCAA has a rating system and for which there are two other rating systems, one by Kenneth Massey and the other a variation I created of the NCAA system.

    Each year, I assign pre-season ratings to teams. I do this strictly based on their past ratings (and resulting ranks) under my rating system variation. I do this for two reasons: (1) to use when providing tools coaches with NCAA Tournament aspirations can use when scheduling non-conference opponents in future years and (2) to use in projecting where teams will end up in the upcoming season, which is largely a fun exercise that becomes more useful to coaches as the season progresses during which I substitute actual game results for predicted game results.

    One of the things I do, out of curiosity and as a test for the usefulness of predicting strength using only past ratings, is use the ratings and team schedules to predict where teams will end up in their in-conference rankings at the end of the season. There are two others who do predicted in-conference rankings: (1) the coaches in the conference, who have a lot of awareness of the strengths of their conference opponents, who has graduated, who is new to the team, and lots of other details, and (2) a university faculty member who does in-conference rankings based on a number of metrics such as players who have graduated, who is new, coach past success, lost goal scoring due to graduation, and so on. At the end of the season, I compare how close I, the coaches, and the faculty member came, in their in-conference rankings, to what the actual rankings ended up being. My expectation when I started doing this was that the coaches and faculty member would do much better than my purely history-based ratings/ranks.

    As an additional piece of info, I have found that the average of the last 7 years' rankings under my rating system produces the best match with where teams will end up next year.

    With that as the setting, using the 2023 season as an example, which involved 347 teams that were in conferences, divided among 31 conferences, with the numbers of teams in the conferences ranging from 8 to 15, here is how close the three of us came to our pre-season rankings matching the actual end-of-season rankings:

    Coaches: On average, their predicted rankings were within 2.09 positions of the actual rankings

    Faculty member: Within 2.28 positions

    Me: Within 2.15 positions​

    In 2022, the numbers were 2.44, 2.44, and 2.61 respectively. And in 2021, they were 2.0, 2.13, and 2.40 respectively.

    I take two lessons from this. First, it is impossible to predict precisely where teams will end up at the end of the season. Second, using only past history provides for less accurate predictions than using detailed current knowledge and expertise, but only slightly less accurate.

    In other words, from the second lesson, teams in fact tend to be fairly consistent over the years, to the extent that humans with expertise and detailed knowledge of teams can make only marginally better predictions of where teams will end up than one can get looking simply at past performance. Although I had suspected after watching teams' ranks as they evolved over the years that past performance would be a pretty good indicator of upcoming performance, I still am surprised how well past performance does as compared to what humans with expertise and a great deal of current information (and a lot at stake so far as the coaches are concerned) can do.

    Some might find this irritating or depressing, to think that humans cannot improve a lot on the statistics of past history. For me, it simply is interesting.
     
  19. blissett

    blissett Member+

    Aug 20, 2011
    Italy
    Club:
    --other--
    Nat'l Team:
    --other--
    toad455 repped this.
  20. toad455

    toad455 Member+

    Nov 28, 2005
    Liechtenstein is newly added to the rankings , which is now at a new high for the amount of teams listed.
     
    JanBalk and blissett repped this.
  21. lil_one

    lil_one Member+

    Nov 26, 2013
    Nat'l Team:
    United States
    An ESPN article on Liechtenstein joining the rankings with friendlies against Namibia:

     
    Lechus7, Cheetah101, blissett and 2 others repped this.
  22. Cheetah101

    Cheetah101 Member

    Apr 21, 2009
    Arizona
    Nat'l Team:
    United States
    C'est le ooof!

    France with a huge loss in ratings points and position :alien:
     
    toad455 repped this.
  23. lil_one

    lil_one Member+

    Nov 26, 2013
    Nat'l Team:
    United States
    Meanwhile England and Sweden are 4D-chessing it: if you're not in the Olympics, you can't lose points in the Olympics. /s
     
    Lohmann and blissett repped this.
  24. Lechus7

    Lechus7 Member+

    Aug 31, 2011
    Wroclaw
    Phew... North Koreans play the system for years :D;)
     
    JanBalk, Lohmann and blissett repped this.

Share This Page