FIFA rankings - why they are important and how to beat the system

[Kent]

1 hour ago, CanadianSoccerFan said:

Picking up the FIFA ranking discussion here to avoid hijacking the Phonsie thread.

I dug up the article I read a few years ago that dissects the core problem with the new ranking.  Essentially, teams do not move because the points never drop off and the increase/decrease in points resulting from a match are too small to make any material difference. 

The teams that were on a hot run of form right before the ranking change (Wales, Jamaica etc.) will benefit from this forever.  

http://www.weglobalfootball.com/2018/06/12/fifas-new-elo-rankings/

This excerpt drives home the absurdity:

There’s a 750 point gap between #1 and #28. Slovakia would have to beat Germany 20 times in a row just to pass them in the rankings. And that’s just if it’s a World Cup Qualifier. If it’s in friendlies, Slovakia would have to beat Germany 50 times to pass them. Slovakia could beat Germany by 10 goals on the road 40 times in a row and Germany would be ranked higher. 

Yeah, that makes sense. USA is like 200 points ahead of us. When we beat USA we got 17 points. Then we lost to them, which should be an expected result, and we lost 8 points. So that was a net gain of just 9 points.

I wonder if winning the Gold Cup, winning the Oct, then winning the World Cup would be enough for us to become the top ranked team in CONCACAF? Probably not because Mexico is almost 300 points ahead of us.

[maplebanana]

20 hours ago, CanadianSoccerFan said:

Picking up the FIFA ranking discussion here to avoid hijacking the Phonsie thread.

I dug up the article I read a few years ago that dissects the core problem with the new ranking.  Essentially, teams do not move because the points never drop off and the increase/decrease in points resulting from a match are too small to make any material difference. 

The teams that were on a hot run of form right before the ranking change (Wales, Jamaica etc.) will benefit from this forever.  

http://www.weglobalfootball.com/2018/06/12/fifas-new-elo-rankings/

This excerpt drives home the absurdity:

There’s a 750 point gap between #1 and #28. Slovakia would have to beat Germany 20 times in a row just to pass them in the rankings. And that’s just if it’s a World Cup Qualifier. If it’s in friendlies, Slovakia would have to beat Germany 50 times to pass them. Slovakia could beat Germany by 10 goals on the road 40 times in a row and Germany would be ranked higher. 

The new rankings procedure indeed makes it harder to move up and down, but after a few years I've seen more movement than I expected (I read this link a few years back too). A couple of notes on their analysis:

1) Germany/Slovakia: I don't know where they got those original points values. Maybe it was under the old system, but under the new system I believe they took the #1 and #210 teams and evenly spaced them out points-wise. If you look today, the #1 team is Belgium (1783) and #28 is Japan (1509). This is only a 274 pt delta. Since this modified Elo is based on a zero-sum model, if Japan beat Belgium they would gain the points that Belgium loses, in this case about 7.5 points or a 15 pt swing for a friendly. So the 274 pt swing is made up by 18.3 wins in friendlies. Considerably less if it was let's say Canada vs Mexico in official competitions. So, yes the point still stands that it is hard to climb rankings. But it's not as rigidly locked as this Germany/Slovakia example.

2) Simulation: I didn't see any details of the simulation. If I was to write a Monte Carlo simulation of this, the key metric that would define how it goes is how to calculate individual win probability. If the assumption that 'better ranked teams have a proportionately higher chance of winning', then of course any large number of simulations will yield the steady-state result of the original ranking. The more interesting thing would be to analyze the standard deviation of teams throughout those 500 simulations. Does a team like Canada, currently #70, end up as #40 1% of the time? or does it end up as 70 +/- 3 99% of the time? The measure of the variability in a Monte Carlo simulation is the metric to look at, not the single end result of one simulation.

Ultimately, to me the Elo-style ranking makes sense. I just wish that, given the magnitude of the points and differential, that they would change: a) bake home/away and GD into the calculation of We (expected result), and b) increase the multipliers across the board to increase volatility.

[baulderdash77]

A quick update after our outstanding win against Costa Rica- our live ranking is up to date.

Going into the Semi Finals match against Mexico (it’s so awesome to say that) we are on 1399.28 points and our live ranking is 59th overall.

It also places us 5th in CONCACAF for the first time in a long time.  These wins add up and matter.

 

edit- similar to the last Gold Cup- playing non FIFA clubs in the Gold Cup like Martinique “cost” Canada about 15-20 points.  Playing them 2 cycles in a row, plus all the non-FIFA teams we played in CONCACAF League has actually lowered our rating by over 50 points under the new methodology.  That’s a lot of points and is the difference between us being mid 40’s and high 59’s.

Edited July 26, 2021 by baulderdash77

[TOcanadafan]

59 minutes ago, baulderdash77 said:

edit- similar to the last Gold Cup- playing non FIFA clubs in the Gold Cup like Martinique “cost” Canada about 15-20 points.  Playing them 2 cycles in a row, plus all the non-FIFA teams we played in CONCACAF League has actually lowered our rating by over 50 points under the new methodology.  That’s a lot of points and is the difference between us being mid 40’s and high 59’s.

I checked the  Elo ratings, and they have us at 35th now… what a difference.

[CanadianSoccerFan]

20 hours ago, baulderdash77 said:

A quick update after our outstanding win against Costa Rica- our live ranking is up to date.

Going into the Semi Finals match against Mexico (it’s so awesome to say that) we are on 1399.28 points and our live ranking is 59th overall.

It also places us 5th in CONCACAF for the first time in a long time.  These wins add up and matter.

 

edit- similar to the last Gold Cup- playing non FIFA clubs in the Gold Cup like Martinique “cost” Canada about 15-20 points.  Playing them 2 cycles in a row, plus all the non-FIFA teams we played in CONCACAF League has actually lowered our rating by over 50 points under the new methodology.  That’s a lot of points and is the difference between us being mid 40’s and high 59’s.

Although UK work permit rules have changed,  the top 50 is still the minimum to guarantee a work permit without an appeal.  Those lost points are potentially holding back careers.

 

https://worldfootballindex.com/2020/12/work-permits-premier-league-post-brexit-english-clubs-efl-libertadores/

[narduch]

 

[Club Linesman]

Wow the US and Mexico in the top 10?!?.  If it is close to being accurate, it makes our recent results against them pretty damn impressive.  Time to show Concacaf we are truly the third team in the region come September!

[Obinna]

58 minutes ago, Club Linesman said:

Wow the US and Mexico in the top 10?!?.  If it is close to being accurate, it makes our recent results against them pretty damn impressive.  Time to show Concacaf we are truly the third team in the region come September!

So we pass El Salvador and Honduras. Next release we can perhaps pass Jamaica and Costa Rica.

[Floortom]

US and Mexico are overrated by FIFA (ELO has them in the 15-20 range) while Canada is underrated (ELO has us #39)

[Kent]

1 hour ago, Obinna said:

So we pass El Salvador and Honduras. Next release we can perhaps pass Jamaica and Costa Rica.

Just some scenarios.

Currently we are 24 points behind Jamaica, and 51 points behind Costa Rica.

Canada win all 3 in September = 41 points
Canada win both home games, lose to USA = 16 points (beating USA is +18, losing to them is -7)
Costa Rica and Jamaica play each other in September. If Costa Rica wins they swap 12 points, if Jamaica wins they swap 13 points. With a draw Jamaica picks up about half a point from Costa Rica.

[PegCityCam]

Top five mover in FIFA...no complaints here.

[Kent]

29 minutes ago, PegCityCam said:

Top five mover in FIFA...no complaints here.

Top five written differently.

1. UEFA champ
2. CONCACAF champ
3. CONMEBOL champ
4. UEFA runner up
5. Canada!

[PegCityCam]

1 hour ago, Kent said:

Just some scenarios.

Currently we are 24 points behind Jamaica, and 51 points behind Costa Rica.

Canada win all 3 in September = 41 points
Canada win both home games, lose to USA = 16 points (beating USA is +18, losing to them is -7)
Costa Rica and Jamaica play each other in September. If Costa Rica wins they swap 12 points, if Jamaica wins they swap 13 points. With a draw Jamaica picks up about half a point from Costa Rica

So basically a sweep puts us somewhere around 45/46 but losing to the US has us around 54. Wow, one game...but I suppose at this point we clearly have loftier ambitions than boosting our rankings 😅

[jtpc]

Feels great to pass El Salvador and Honduras. Here are the point totals for the current top 5.

Based on our match with Costa Rica at the Gold Cup, I don't think their hold on 3rd place lasts much longer. If we can really shrink the 51-point gap by the end of the October window, we might be able to overtake them when we play them at home in November. The ranking update after the November window could be very interesting. 😉 💪

Edited August 13, 2021 by jtpc

[Acid-Tone]

13 hours ago, Club Linesman said:

Wow the US and Mexico in the top 10?!?.  If it is close to being accurate, it makes our recent results against them pretty damn impressive.  Time to show Concacaf we are truly the third team in the region come September!

I don't know anyone who follows the game closely and pays all that much attention to those rankings.

Obviously, they're important for the purpose of seeding in qualifying for the World Cup.

I mean the rankings leave a lot to be desired when trying to compare teams from different continents.

Edited August 13, 2021 by Acid-Tone

[baulderdash77]

9 hours ago, PegCityCam said:

So basically a sweep puts us somewhere around 45/46 but losing to the US has us around 54. Wow, one game...but I suppose at this point we clearly have loftier ambitions than boosting our rankings 😅

Getting into the top 50 is important because it allows any of our international players to get a work visa in the UK without any hassle.  So there is some importance to a degree.

[Trois Reds]

On 8/12/2021 at 8:27 PM, baulderdash77 said:

Getting into the top 50 is important because it allows any of our international players to get a work visa in the UK without any hassle.  So there is some importance to a degree.

Isn't that the case with the EU in general?  I remember reading somewhere a few years back (when we were somewhere around 90th) about Canadian soccer players having a hard time transferring to Europe, and even within Europe between countries, and it was an EU-wide regulation.  Which may also be why we see mediocre Americans going to Serie A while players like Eustaqio seem to not move even when his abilities are obvious to everyone.

I wouldn't be surprised that one of the issues with Tajon Buchanan going to a European club is that Canada's ranked too low. If he was American, he would have been in Europe by now.

[nfitz]

I prefer the CONCACAF rankings. The August 1 version has us in 3rd place: https://www.concacaf.com/rankings/men-s-national-team/

Edited August 16, 2021 by nfitz

[CanadianSoccerFan]

On 8/14/2021 at 1:43 PM, Trois Reds said:

Isn't that the case with the EU in general?  I remember reading somewhere a few years back (when we were somewhere around 90th) about Canadian soccer players having a hard time transferring to Europe, and even within Europe between countries, and it was an EU-wide regulation.  Which may also be why we see mediocre Americans going to Serie A while players like Eustaqio seem to not move even when his abilities are obvious to everyone.

I wouldn't be surprised that one of the issues with Tajon Buchanan going to a European club is that Canada's ranked too low. If he was American, he would have been in Europe by now.

It's not.  Each country has their own rules and regulations surrounding import players.  For most EU states, the import restriction is usually a straight quota of non-EU passport holders.  Each team is permitted X number of imports and they let the market sort out who is worthy.  Classic case was JDG in Spain where teams were only allowed 3 non-EU passport holders.  Teams assumed he had one because of his brother so they would call up his agent to sign him and then have to drop their interest because they lacked an available import spot in their squad.

The UK was always a weird outlier because instead of a quota, they had these visa requirements that were purely based on National team appearances and ranking.  It created these weird inefficiencies where young stars from Brazil and Colombia couldn't get work permits because they weren't in the national team yet there were players from Honduras like Ramon Nunez getting a work permit to play for Scunthorpe in League One.  

Brexit necessitated the introduction of club-based performance criteria that makes it slightly less ridiculous but the top 50 national team ranking is still the only way to automatically qualify without an appeal.

 

 

[Kent]

Some meaningless info. 7 years ago I got curious and started keeping track of Canada's position in the FIFA rankings month to month, and how many countries of different population sizes were ahead of us. I don't remember exactly, but I think the curiosity came from comments like "Canada are ranked 107th in the World, between Georgia and St. Kitts." Or it might have been comments acting like we are automatically superior to such small countries.

Here are some random factoids from my useless endeavor.

Since I've been compiling, Canada has been ranked as low as 16th in CONCACAF and 122 in the World (October 2014). Right now we are at our peak over that period at 5th in CONCACAF and 59th in the World.

We have had as many as 8 countries ranked ahead of us with a population below 1 million people (at a time, not a total), 30 countries ahead of us with a population between 1 million and 5 million, and 23 countries with between 5 million and 10 million. Currently there is 1 ahead of us smaller than 1 million, 9 between 1 million and 5 million, and 10 between 5 million and 10 million. Over the entire 7 year period, we have had 11 different countries of fewer than 1 million ranked ahead of us at one time or another.

Some of the less glamorous countries that have been ahead of us in the past include Lesotho, Kuwait, St. Vincent and the Grenadines, Faroe Islands, Benin, Sudan, and Oman. The countries ahead of us at the moment aren't nearly as embarrassing to admit to your friends. Nothing scientific in this but now the least glamorous teams ahead of us are more like Bosnia, Northern Ireland, Iceland and Qatar. Not too bad!

[Stryker911]

2 hours ago, Kent said:

Some meaningless info. 7 years ago I got curious and started keeping track of Canada's position in the FIFA rankings month to month, and how many countries of different population sizes were ahead of us. I don't remember exactly, but I think the curiosity came from comments like "Canada are ranked 107th in the World, between Georgia and St. Kitts." Or it might have been comments acting like we are automatically superior to such small countries.

Here are some random factoids from my useless endeavor.

Since I've been compiling, Canada has been ranked as low as 16th in CONCACAF and 122 in the World (October 2014). Right now we are at our peak over that period at 5th in CONCACAF and 59th in the World.

We have had as many as 8 countries ranked ahead of us with a population below 1 million people (at a time, not a total), 30 countries ahead of us with a population between 1 million and 5 million, and 23 countries with between 5 million and 10 million. Currently there is 1 ahead of us smaller than 1 million, 9 between 1 million and 5 million, and 10 between 5 million and 10 million. Over the entire 7 year period, we have had 11 different countries of fewer than 1 million ranked ahead of us at one time or another.

Some of the less glamorous countries that have been ahead of us in the past include Lesotho, Kuwait, St. Vincent and the Grenadines, Faroe Islands, Benin, Sudan, and Oman. The countries ahead of us at the moment aren't nearly as embarrassing to admit to your friends. Nothing scientific in this but now the least glamorous teams ahead of us are more like Bosnia, Northern Ireland, Iceland and Qatar. Not too bad!

Very interesting. I still saw comments on Sportsnet when the last ranking came out making fun of us being between Mali and Iceland. Clearly many people are just clueless.

I think that that ELO may have us a bit higher than we should be at this point, but could achieve in a year or 2 if WCQ goes well. Realistically, we should be late 30s to late 40s with our current talent.

[Bison44]

1 hour ago, Stryker911 said:

Very interesting. I still saw comments on Sportsnet when the last ranking came out making fun of us being between Mali and Iceland. Clearly many people are just clueless.

I think that that ELO may have us a bit higher than we should be at this point, but could achieve in a year or 2 if WCQ goes well. Realistically, we should be late 30s to late 40s with our current talent.

These are the guys obsessing endlessly over the 4th line center for the Leafs, or the 5th starter for the Blue Jays.  

[masster]

With a lot of casual observers starting to pay attention to the CMNT this week, I'm sure many will focus on our world rank. How much of a jump will we get for a draw against the world #10 team in a competitive match?

[maplebanana]

58 minutes ago, masster said:

With a lot of casual observers starting to pay attention to the CMNT this week, I'm sure many will focus on our world rank. How much of a jump will we get for a draw against the world #10 team in a competitive match?

For USA draw we got 5.54 pts. For Honduras draw we lost 0.53 pts. According to the live rankings I believe that puts us up one rank to 58 (above JAM), but of course we should probably wait till end of this window before doing any calculations. A win against El Salvador would get us 11.9 pts.

If you're curious about any pts gained/lost, I posted plots in the previous page of this thread. And current ranking pts are found here: https://www.fifa.com/fifa-world-ranking/men?dateId=id13372

[Yoginess]

I see we're ranked 59 in the Pepsi-FIFA rankings but on the ELO ratings we're up to 37th! Love it.