Given the never ending debates about the relative qualities of the various leagues, some of them 2nd Divisions, here are two Elo based sites that offer their rankings based on a mathematical formulas. Warning - the cross continental rankings, although offered as well, are next to impossible to establish realistically due to a relative paucity of data. However, the intra-national competitions are taken into account. http://clubelo.com/All/Ranking.html http://www.footballdatabase.com/worldrankings.php?Rnk=1 Do note that this allows to also rank leagues outside of the UEFA formulas that favor the "top heavy" leagues where a few clubs dominate not only locally but across the sanctioned UEFA events. FWIW : Clubelo has these YA clubs thusly ranked - Schalke in 15th, Tottenham in 17th, Malaga in 18th, Everton 19th, Hannover 34th, Roma 47th, Stoke 55th and so on.
Football Database UEFA Rankings for Prominent YAs: 18. Tottenham (Dempsey, Friedel) 20. Schalke (Jones) 23. Everton (Howard, Donovan occasionally on loan) 24. Malaga (Onyewu) 41. Roma (Bradley) 55. Anderlecht (Kljestan) 57. Hannover (Cherundolo) 62. Stoke City (Cameron, Edu) 108. Nurnberg (Chandler) 132. AZ Alkmaar (Altidore) 136. Rosenborg (Diskeruud) 144. Aston Villa (Guzan, Lichaj) 164. Molde (Gatt) 165. Birmingham City (Spector) 184. Hoffenheim (Johnson, Williams) 192. Helsingborgs (Bedoya) 199. Bolton (Ream, Holden) 211. Rapid Wien (Boyd) 249. Hertha Berlin (Brooks, Morales) 287. St. Pauli (Gyau) 297. Nordsjaelland (Parkhurst) 346. Racing Santander (Bocanegra) 477. Brondby (Goodson) I'm sure I missed some, but those were the clubs I noticed.
Since Elo's ratings are cumulative, there's usually a time lag - or should I say, a longer time lag - for the big (up or down) movers, which may include freshly promoted or demoted squads or Johnny-come-latelies (Rosenborg being ranked higher than Molde). But what I do find interesting is that, as expected, some 2nd tier squads are rated above the 1st tier clubs of the smaller nations, as is shown in the case of Hertha and St. Pauli being ranked higher than Brondby and Nordjaeland and Bolton higher than Rapid Wien.
You can see that they're both deeply flawed. For example New Saints of Wales are ranked 15 places above Newcastle Utd, when in reality, Newcastle are one of the better Premiership teams and New Saints play in a league that is comparable to the 4th/5th tier of English football.
That's a bit bizarre. New Saints are ranked about 600 places below Newcastle in the first link. Are they both using something like Elo? Did the 2nd one just make some basic error?
It's not odd, just a fault of the system that also effects the CONCACAF teams to some extent. A top Welsh team dominates its opposition and keeps earning points but it doesn't get the negative points by competing on the European turf since it usually goes one and out. Similarly, other Welsh clubs don't get dinged during these events since they barely participate in them. In chess, where Elo formulas were first used, the cross-reference was fairly easy - top players went against each other in 20-30 match tournaments and either earned or lost points off these matches. But a league like Wales is basically isolated from Euro competition, so there is no sufficient amount of data to chart their clubs as the UEFA events are geared toward top clubs and the related point totals. BTW, this was also the problem with the old FIFA ranking system that had the US at #4 - Bruce Arena could earn points by beating up on inferior CONCACAF foes and did not lose enough points by playing a vigorous top opposition schedule. In this instance, one may as well ignore the statistical outliers. PS. Newcastle, of course, had a disastrous 08-09 season, followed by a year in the Championship. At this moment, they are 14th in the EPL and even that is based on yesterday's big win over Wigan. Otherwise, they were 15th-17th.
The problem is that often inter-league rankings are worse predictors of actual results of head-to-head games than opinion. For example New Saints could go out and buy the entire Barca squad (well not really) and they'd instantly be no.1 in the World, but no ranking could ever pick that up quickly. Inter-confederation rankings are even worse as you only get a handful of competitive games every year involving only a small number of the top clubs from each confederation.
The interconfederational rankings are a moot point due to the absence of data - only World Club Championship (hypothetically, one can reset the starting points based on the WCC but then this quits being a purely Elo model). This means the same thing as Wales - a Malaysian champion, for example, could dominate its local tournament, then quickly lose in the Asian Champions League and still retain its lofted numerical status. The other thing is a lag - ManCity's record in its last two years is obviously superior to the one of five years prior, when it was bringing up the bottom of the EPL under Keegan and Pearce.
Footballdatabase does not rank 2nd division sides. The 2nd division clubs in there are ranked based on 1st division results. E.g. Bochum are still ranked because of their results from 2010, I leave it up to you to decide how much sense this makes.
Ranking exactly one club over the other is very difficult and can lead to arbitrary results. And nearly impossible if you try to rank across federations because there are so little competitions apart form the Club World Cup. For Europe, I actually prefer the UEFA coefficients. You don't get a a club- by- club ranking, that is true, but a good overview of the strength of a league in comparision to the other leagues and the possibility to extrapolate the trend in which it is developing. The methodology is clear and transparent. Drawback is of course that the second divisions are not ranked and that top- heavy divisions might have a slight advantage.
No, what I meant was that it's bizarre because the two links you provided come up with completely different results on that matter, despite their both being based onElo. In the second one, New Saints are well above Newcastle. In the first one they're about 600 places below.
I think that would be due to the different time frame and the corresponding results that are being plugged into the computer. Over time, they should converge. But, hey, there's a reason I linked to both.