I would rep this post multiple times if it was possible. It's incredible how, after so many years, we still have people here who basically don't understand the meaning of FIFA World rankings. Do they want some kind of "power rankings" (whatever it means and however they're randomly decided)? Well, this is not the place for that. Luckily, a resonable poster as @sbahnhof just created a thread for that! (oh, by the way, I can't help asking you the meaning of your nickname, @sbahnhof: it's quite some time I wanted to do, but the fact that next month I'll be in Wien and I'll arrive from Italy at Südbahnhof gave me the final push to ask! ).
Power rankings aren't random... but they also aren't based on things that can be directly proven or disproven since they're basically just someone's opinion. Usually it's "expert A" or "media person B" or "discussion group C" putting out their list of what they think the rankings should be at any given time. Power rankings, especially taken in aggregate, are usually very similar to well-constructed quantitative rankings but will have differences based on whether people think a team is "hot" or "cold" or otherwise misrepresented by a numbers-only approach. Basically, they're squishy and people can adjust them on the fly as much as they want. To fair, I don't wholly disapprove of Power Rankings. I've made some in the past and I read both the MLS and NWSL weekly power rankings articles - but, if you've seen comments from past seasons in the NWSL forum, you'll know I (and others) generally think of the NWSL power rankings as being far too shallow. And that's the issue with power rankings - different people will value different things, and that makes power rankings hard to take seriously. They're fun on their own, but don't mean much unless you have a lot of people agreeing on them - and even then, consensus can be swayed by various factors that aren't ultimately realized on the field.
Yes, the Power Rankings I've been mostly exposed to are NWSL ones, so, as you said, it's difficult for me to take them seriously...
@SiberianThunderT, I would rep your long post a billion times if I could. Once will have to do. I'm afraid your words are just falling on deaf ears with him, though.
Correct, and by design. The FIFA woso rankings are deliberately insensitive to these things. Also, the FIFA woso rankings use a 4-year sliding window, so they're relatively insensitive to meteoric 1-year changes. False. Every Elo-based ranking system is about 70% accurate as a predictor. No ranking system in the world is demonstrably more accurate than that. Accuracy as a predictor is not the design goal of a ranking system. That is, we math/econ dudes know (based on ~300 years of cumulative research) that ~70% is the best humans can do, with any system. Ergo, there's zero incentive to seek a "better" system. So the FIFA woso ranking systems don't give two hoots about prediction accuracy. And neither do the fans of the underdog team. (N.B. The mso rankings are doing even worse in predictiong the WC so far.) You think the 30% of mis-predictions is evidence of failure. We don't; we think 30% failure (upsets) is normal. Elo will never do better than that, and neither will any other ranking system.
The current (soon to be ex-)Men's rankings use a 4-year window, but the Women's results do not use any time filtering. An Elo system naturally favors recent results.
I'm realizing I somewhat misinterpreted Gilmoy's post for the part I quoted - the intent was to say that the FIFA Women's rankings don't show fast changes. The way the rankings are calculated, there's no explicit favoring any given time period. While these calculations generally phase out old results, it's also true that the rankings have an inertia to them - in theory, the rating a team has represents their average performance over the last 20-30 or so matches, with some emphasis on more recent matches but not incredibly so. This happens to work out to around 3-5 years of results for moderately active teams. - - - - - Anyway, as for predictive power, in theory Elo-based rankings are still pretty good at those, though obviously not perfect. Assuming the teams' ratings are truly representative of their current strengths, as is the intent, then the higher one team's rating is, the more likely they are to win. And as one very interesting point of comparison from just the last hour or so.... Argentina men 0-3 Croatia men in Russia. The current FIFA men's rankings make it look like a #18 over #5 upset... but by the men's Elo rankings, it's a relatively predictable #8 over #12 win. X-D Yes, I know it's just one data point, but it's fun to look at! A similar thing happened when Senegal beat Poland - it looked like a #28 over #10 upset by FIFA's ranking, but it was just #23 over #22 by Elo.
now, now, I clearly mean't the #1 FIFA ranked team at the time of a major tournament failed over & over again to win it. In this decade, their battling o for 3(2011 & 2015 WC, 2016 Olympics) which would be ah............ZERO percent (dunno who was ranked #1 going into the 2012 Olympics) and another thing FIFA don't cover is JINX, if Ellis wants to win the next WC she better ditch of that #1 ranking!(in other words, Ellis, stop playing freaking tomato cans!)
It probably means that scheduling friendlies across confederations will me easier, since they don’t mean much. But the real rankings will come in competitions like the UEFA cup and the copa América, which may eclipse the Libertadores as the premier cup in the South. Expect there won’t be invitees from concacaf for the Copa América like there is now. CONCACAF is going to figure out how to get the top teams playing each other more regularly in confederation cups.
You are showing a lack of understanding of how Elo systems work. The highest rankings come from breathing top 10 teams like England, France, Germany, Japan and Canada, all of which they have done recently. Those wins carry twice as much weight in the women’s system. Beating tin coated steel does nothing for top ranked teams. If anything, it drags an Elo ranking down.
One other item to add to the mix: Elo-type systems, and other mathematical systems with rating/result correlations in the same ball park as Elo-type systems, will do better at predicting results of upcoming games than most humans. Occasionally, there may be a truly brilliant human who beats a good mathematical system through skill (although I'm not sure I've ever run into one). Most of the time, however, humans who beat the mathematical systems simply are lucky. And the mathematical systems will beat most of the humans virtually all of the time.
....there's so much factual misinterpretation and misdirection in the tiny post, I could write a response nearly twice as long as my other long post detailing everything you've got wrong, but it's clearly worthless at this point. Enjoy your little frustrated bubble!
The US was ranked #1 going into the 2012 Olympics. And you really need to get over your "The #1 team must absolutely win every major tournament or the ranking is wrong!" nonsense. Weird crap happens all the time in major tournaments (the US being unable to finish their billion chances in the 2011 final comes to mind right away). Deal with it.
https://www.fifa.com/fifa-world-ranking/ranking-table/women/index.html https://www.fifa.com/fifa-world-ran...imb-plus-host-of-names-return-to-ranking.html Leader: USA (unchanged) Moves into the top ten: Japan (6th, up 5) Moves out of the top ten: Sweden (11th, down 2) Matches played in total: 202 Most matches played: Thailand (10) Biggest move by points: Japan (6th, up 38) Biggest move by ranks: Slovenia (54th, up 8) Biggest drop by points: Jordan (57th, down 47) Biggest drop by ranks: Bosnia and Herzegovina (66th, down 9) Newly ranked teams (re-entries unless otherwise stated): Ecuador, Haiti, Jamaica, Indonesia, Algeria, Guyana, Cuba, Congo, Dominican Republic, Puerto Rico, Ethiopia, Suriname, St Lucia, St Kitts and Nevis, Bermuda, Barbados (first appearance), Tanzania, St Vincent and the Grenadines, Dominica, Grenada, US Virgin Islands, Uganda, Antigua and Barbuda, Curaçao, Aruba Teams no longer ranked due to prolonged inactivity: Guam, Egypt
Congrats to Japan! Moving that many spots when already that high in the rankings is quite the improvement. Though it's very tight 6-9, so maybe not so surprising. Australia didn't drop nearly as much as I thought they would... Although Spain didn't earn a new rank this time, they are continuing to make a rapid climb in rating points. If they finish their qualifying campaign strong and earn another March title, they could very well crack the Top Ten by the time the WWC rolls around. Other big movers in UEFA: Wales up 5, Poland down 5 Thailand, despite almost upsetting Australia, don't make a huge move - the draw certainly earned them a lot of points, but less than they would have gotten with a win, and none of their other results were particularly point-winning. SO MANY nations back into the ranking! Including a ton from the Caribbean. Despite all of the matches played, though, if you look at the ratings changes, you don't see any shift in order within CONCACAF until you get down to SKN, who were 1pt behind Bermuda previously but are now 5pts ahead of Bermuda (now 14th and 15th in the region). Barbados and SVG also took hits (the latter used to be 13th but is now 17th in C'CAF). With falls from CAN and AUS, the 100pt-gap-from-1st now only extends down to 4th place. (We haven't examined the 200pt-gap-from-1st before, but it's usually been around 11th place in recent history, thanks to a consistent chasm between 11 and 12 - but Spain's rise in rating at 12th place now has them within 200pts of USA too!)
Well, we had dropped very lower than what the team was worth, in my opinion, so I expected a return to the top ten soon, but in fact winning AFC Asian Cup made it way more resounding that I could have figured. I am a little sursprised that, despite having just qualified for WWC 2019, with 7 wins out of 7 games, Italy didn't climb any position. Ok, the wins vs Belgium and Portugal at home were mostly expected, but I had the feeling Italy was having quite decent results in the last months.
The 140 nations ranked is a new record. I'd expect that record to be broken in September as OFC has their qualifiers that will add 10 more plus three from Central America. Hopefully no one drops out of the rankings in September due to inactivity.
Only the preliminary matches for OFC qualifying will be played before the next rankings come out. The main tournament is in November.
You can't really take the Fifa rankings too literally or too seriously--not when it comes to most of the teams in, say, the top 10 or so. The ranking is a lagging indicator. I would say that the U.S. is a clear no. 1, but most of the other teams in the top 10 are very closely grouped. Thanks for this. I think the Dutch lineup is quite different these days--better. I don't think they are quite good enough in the back, or have enough depth of talent, to beat the United States--but I think they are good enough to beat anybody else in the world on a good day, it wouldn't surprise me to see these two teams meet in the knockout round of the World Cup.
It is true that one shouldn't take a few positions' differences as being of great significance. On the other hand, ordinarily a few positions' differences at the top of the rankings have more significance than a similar number of positions' differences farther down in the rankings (until one gets to the bottom of the rankings). It's also true that the rankings have some "lag" to them. The problem with humans' evaluations as compared to the computer's evaluations, however, is that ordinarily humans' evaluations do not have enough "lag" to them. Humans tend to be too susceptible to "flavor of the month" thinking.
Actually, considering how often the teams in the top 10 play games, especially against teams outside of their home confeds, one could easily argue that the rankings are actually most accurate for that group. As cpt pointed out, the further down you go, the less a difference in rank means, since the difference in points between ranks decreases. It's especially meaningless when comparing teams from different confeds because you get the issue that weaker teams often don't play cross-confederation games, and that "pooling" effect can cause some serious points drift the further down you go. The one thing you really have to remember is that rank is less important than rating. You say that most of the teams in the top 10 are very closely grouped - and the official FIFA rankings would agree with you, since the ratings gap between USA and GER is more than 3x bigger than any ratings gap elsewhere in the top 10. That 1-2 gap is 65pts, while all of the other gaps are less than 22pts and the gap from 6-9 is 11pts total. That's tight! Way back, I had said this: You could add a corollary that, inside the top 20, differences of 10pts-20pts are not meaningless but also not incredibly meaningful either. It's also worth remembering that home field advantage in this system accounts for 100pts, so you could say that any pair of top-30 teams within, say, 50pts of each other are going to be pretty evenly matched, and it would be hard to declare anything between them a true "upset". Netherlands beating England? Brazil beating France? Neither of those are going to surprise anyone - though in both cases you would probably say the opposite result is still more likely.
Actually, you can, if you know what they mean. Rankings are like frequent flyer miles; you earn them for past accomplishments, and so they simply pile up as a measure of those. So they literally mean "earned this many ... miles". They don't mean "is a so-called better team right now". Nonetheless, these two interpretations aren't independent; the correlation between them is actually pretty high. If you insist on using rankings to predict, it will always be a weak predictor. OTOH, if you (correctly) use them to evaluate past performance relative to expectation, then it's 100% objective, plug-and-chug. Multiple anonymous blockchain servers could independently run the algorithm in secret, all agree on the same answer, and thereby verify the next FIFA blockchain ratings So when I look at the FIFA woso rankings on any day of the year, I can be 100% confident that they literally tell me who has earned the most miles, and how many of them. That's not "strength", which is fine; I have other ways to judge team strength.
Emphasis added, because the sentence quoted next doesn't really follow from the bolded one, nor is it backed up by results. As we've tried explaining to another previous poster, the rankings have proven to honestly be a pretty good predictor when taken in aggregate. And that's probably the important point - they're better overall (e.g. many games) than they are for predicting any single game.