World Ranking

[pineapple stu]

The ratings go back to the start of time, but Elo ratings by their very nature will only reflect the most recent 100 or so matches. That's just how they work. So while the guys have gone back to the start of time, effectively, a 1910 win by Ireland v England will not be reflected in the 2016 rating. It's simply too far back in time to have an effect any more. The two teams' ratings would today be the exact same regardless of that 1910 result. (In fact, the link you quoted says that ratings will converge on a figure after 30 matches. So you yourself quoted a refutation of your own "all matches" point. I think once you go to 100 matches, there's literally no impact on a rating (to all intents and purposes anyway - to a number of decimal points)

I didn't talk about Ireland's rating? But let's also bear in mind that Ireland also had a better qualifying campaign than Wales in 2014 and qualified for 2012. These results do still count towards both an Elo rating and a FIFA rating - and rightly so. You can't just rate countries based on one summer. However, if Wales continue playing better than Ireland, our rating will fall, and theirs will rise. And this, in fact, is exactly what is happening - Ireland have gained 36 points in the past year, while Wales have gained 88 points in the same time.

I don't think you understand how Elo ratings work.

[pineapple stu]

In fact, you can compare the two countries' ratings over time -


Date Ireland Wales
13/10/2016 1757 1745
30/06/2016 1737 1728
31/12/2015 1748 1650
30/06/2015 1696 1676
31/12/2014 1711 1597
30/06/2014 1665 1569
31/12/2013 1689 1562
30/06/2013 1726 1582
31/12/2012 1704 1555
30/06/2012 1712 1604
So you can see Wales have been catching us in recent years, exactly as I suggested. Indeed, Wales had passed us out before we beat Moldova and they drew with Georgia at the weekend. If Wales can keep up their Euro 2016 form, they will pass us out again quite soon - possibly even by beating us in Cardiff next March (if they do beat us)

You don't pay absolute heed to exact rankings as with any ranking system - Ireland aren't suddenly a better team than Wales because they drew with Georgia on Sunday for example - but Elo ratings are based in statistical mathematics, they do work, and they are used in other areas (e.g. chess and Go world rankings)

[DeLorean]

If Wales can keep up their Euro 2016 form, they will pass us out again quite soon - possibly even by beating us in Cardiff next March

Not that it's relevant to what you're discussing but Wales come to Dublin in March and we visit Cardiff in the final game.

[pineapple stu]

Whoops! You're right.

It's actually slightly relevant as Wales would get more Elo points by beating us away than by beating us at home I think.

[Gather round]

I don't think you understand how Elo ratings work

They don't work very well as an effective measure of recent achievement in football tournaments, as I demonstrated. Basically because they aren't based on the two-year tournament cycle after which every team starts again on zero points. FIFA's system has many faults, but it does reflect this to some extent (by weighting results two or three years ago in a previous tournament at only 20% or 30% of those this season).

Given the above, I see little point in familiarising myself with the details (although I readily accept their application elsewhere, eg in chess). Which needs some way of comparing a huge number of players and tournaments. International football, with only 200-odd teams and a simple tournament structure doesn't. So the system is unlikely to be adopted by FIFA, a good thing in my opinion. If it were, one of the likely effects would be usually unsuccessful teams (Wales, Iceland, NI) not getting appropriate reward (ie high seeding) following a one-off success.

I suspect you don't understand that just because something is a) professionally interesting to you and b) of nominal benefit to your team doesn't make it universally beneficial ;)

[Charlie Darwin]

They don't work very well as an effective measure of recent achievement in football tournaments
I may be going slightly out on a limb here, but I'd hazard a guess the reason for this is the fact it's not supposed to measure recent achievement in football tournaments.

[Eminence Grise]

This information certainly comes out of the blue. Time for us to face the music and ditch this system.

*grabs coat*

It's a great discovery to find a fellow enthusiast. It's not before time (though it was, of course...)

[Gather round]

I may be going slightly out on a limb here, but I'd hazard a guess the reason for this is the fact it's not supposed to measure recent achievement in football tournaments

That was my point. Instead, it acts effectively as a predictor of future results. I prefer to use a bookie for that.

[pineapple stu]

FIFA rankings include data on games over the past four years.

The Elo ratings site says its ratings include data on the past 30 games - which is actually a shorter timeframe. So the Elo system actually goes the way you want it to - it includes fewer results. (30 games ago for Ireland brings you back to Costa Rica 1-1 Ireland in June 2014)

In both cases, older games are weighted much less. It is not the case, as you suggest, that the FIFA system is the only one that weights older games less; Elo does this too. If you go back to that Costa Rica friendly and say we won, we would have gained an extra 20 points. But if you keep all other results since then the same - so the only difference is now that we beat rather than drew with Costa Rica 30 games ago - Ireland's rating today might be 1 or 2 points higher.

In both cases, there will be a bit of a lag when a team improves quite quickly (like Wales have done). In neither case will a team shoot to fourth in the rankings purely by cirtue of reaching a semi-finals, as you seem to be suggesting should happen. This would be daft as you'd then have huge swings, which would kind of go towards invalidating the entire point of rankings. A similar argument would say that it would be daft to have Wales ranked ahead of Ireland when we're ahead of Wales in the qualifying group at the moment.

The only real difference is that FIFA rankings are based on a makey-uppey formula used only by FIFA, while Elo ratings are based on statistical mathematics and are used in many other sports. I know which I'd prefer.

For the record, I have no professional interest in Elo statistics, and I couldn't give a flying ****e where Ireland are ranked in the world.

Seriously, your argument doesn't stack up here, not helped by the fact that you admit you don't understand how Elo works. I'm not sure how that makes you qualified to dismiss it.

[Gather round]

It is not the case, as you suggest, that the FIFA system is the only one that weights older games less; Elo does this too

Er, I didn't suggest that (FIFA's system was unique in that respect). Incidentally aside from Elo, UEFA's does as well.


a team shoot to fourth in the rankings purely by cirtue of reaching a semi-finals, as you seem to be suggesting should happen. This would be daft as you'd then have huge swings, which would kind of go towards invalidating the entire point of rankings

Yes, I think the potential for 'huge swings' as you call them is a good thing. I explained why above: teams who are normally mediocre can get a tangible credit for improvement in the next tournament (ie immediate higher seeding); teams who stiff and fail to qualify from first seeding get a tougher group next time.

The immediately above doesn't invalidate the entire point of rankings, of course. It just makes your preferred system to calculate them less applicable.


*A similar argument would say that it would be daft to have Wales ranked ahead of Ireland when we're ahead of Wales in the qualifying group at the moment

I don't make that argument. My preferred system for ranking would be to

a) publish them only once each year (ie immediately after a tournament and at the end of qualifying for the also rans

b) use that from the end of qualifying to seed the next qualifying tournament starting the following year

I've made these points consistently above on this and similar threads.


The only real difference is that FIFA rankings are based on a makey-uppey formula used only by FIFA

Fine. We're agreed that FIFA's system is flawed.


Seriously, your argument doesn't stack up here, not helped by the fact that you admit you don't understand how Elo works. I'm not sure how that makes you qualified to dismiss it

I didn't dismiss it (specifically accepting its application elsewhere eg in chess). Again, you've ignored or misunderstood what I actually wrote. My argument is quite simple- a ranking system should quickly reward recent past achievement. We're agreed that Elo often doesn't and can't do that. Obviously you think the ranking system should do something else. Fine, I disagree. You haven't put a convincing argument that Elo is self-evidently better suited to football ranking.


For the record, I have no professional interest in Elo statistics, and I couldn't give a flying ****e where Ireland are ranked in the world

I think I mistakenly quoted you when it should have been John 83 above. Apologies.

[pineapple stu]

Er, I didn't suggest that (FIFA's system was unique in that respect). Incidentally aside from Elo, UEFA's does as well.
You suggested that Elo doesn't weight towards the present, and said FIFA's system was better because of this. But both systems weight towards the present, so your argument is invalid.


Yes, I think the potential for 'huge swings' as you call them is a good thing. I explained why above: teams who are normally mediocre can get a tangible credit for improvement in the next tournament (ie immediate higher seeding); teams who stiff and fail to qualify from first seeding get a tougher group next time.
But the draw for the World Cup was by necessity made way before the Euro finals, so Wales' huge swing wouldn't have happened by the time of the draw anyway. So you can't say Wales should be higher than Ireland because of a good Euros, and then say the system should be heavily biased towards this to help them get a higher ranking.

You also want a highly reactive rankings list, but then complain that the current system overrates Ireland based on recent form. These two opinions are mutually exclusive. If you only want rankings once a year, that's fine - just ignore all the ones in between and don't be complaining about them. They're not used for anything anyway.


My preferred system for ranking would be to

a) publish them only once each year (ie immediately after a tournament and at the end of qualifying for the also rans

b) use that from the end of qualifying to seed the next qualifying tournament starting the following year
This effectively happens now anyway as the intermediate rankings published each month are completely irrelevant. Both Elo and FIFA rankings will do this. (An exception was for the current qualifiers, when the draw was made before the Euro qualifiers were completed - this was daft).


Fine. We're agreed that FIFA's system is flawed.
Yup.


My argument is quite simple- a ranking system should quickly reward recent past achievement. We're agreed that Elo often doesn't and can't do that.
My point is there's no reason for the rankings to react quite as quickly as you've suggested.


You haven't put a convincing argument that Elo is self-evidently better suited to football ranking.
I have - I've refuted your two technical arguments (that Elo ratings include all games going back to the beginning of time, and that they don't weight towards the present like FIFA's system - neither of which are true). I've concluded that this means you don't understand how the system works - which is a logical conclusion. As I've said before, Elo rankings are based on statistical mathematics, which alone means they're slightly better suited to the job than FIFA's rankings.

In particular, FIFA rankings are based on an average score per game, and if you play fewer friendlies (which are scored lower), your average score will increase and you can skew the rankings. Switzerland were top seeds for the 2014 World Cup qualifying because of this, and Romania are top seeds for 2018 for the same reason (https://en.wikipedia.org/wiki/FIFA_World_Rankings#Criticism). In this regard, Elo rankings are far superior, as they're based on trying to evaluate your strength rather than attributing a score per match, and are consequently much harder to manipulate.


I think I mistakenly quoted you when it should have been John 83 above. Apologies.
Fair enough. But I don't think John83 particularly cares about Ireland's world ranking either btw. I think in fact most people don't care most of the time.

He's just showing a different - maybe more reliable - ranking system.

[geysir]

Possibly GR´s partial bias against Elo rankings is because...

ELO Uefa ranking
13. FAI 1757
27. IFA 1625

FIFA Uefa ranking

20. IFA 798
21. FAI 774

[swinfordfc]

33rd in latest rankings

[tetsujin1979]

Article on football365 on how Wales and Romania beat the rankings system by not playing friendlies: http://www.football365.com/news/how-wales-beat-the-fifa-rankings-system

[Real ale Madrid]

Article on football365 on how Wales and Romania beat the rankings system by not playing friendlies: http://www.football365.com/news/how-wales-beat-the-fifa-rankings-system

Great stuff thanks for sharing.

Meanwhile John Delaney has us playing Oman.

[tetsujin1979]

Wales and Romania don't have a stadium to pay off

[pineapple stu]

Article on football365 on how Wales and Romania beat the rankings system by not playing friendlies: http://www.football365.com/news/how-wales-beat-the-fifa-rankings-system
Yep; I referenced it (the issue, if not the article) in my post further up the page. Switzerland have done it too.

A big plus for Elo ratings over UEFA rankings.

[Real ale Madrid]

Wales and Romania don't have a stadium to pay off

Getting to the World Cup will help the debt more than the Oman Friendly I would have thought.

[geysir]

I hope it turns out more fortunate for us that Wales managed to beat the rankings system and are our group's first seed.

[BonnieShels]

That's what Switzerland have been doing for years.

[swinfordfc]

With the win over Austria - this should boosts us up to the mid 20's I guess?

[Charlie Darwin]

I saw 23 somewhere this morning, not sure if it was accurate or not.

[dantheman]

Yup 23:


http://www.football-rankings.info/2016/11/fifa-ranking-november-2016-final-preview.html

[swinfordfc]

Nice one! .... onto Wales in March :)

[DeLorean]

Nice one! .... onto Wales in March :)

Yeah, Wales are ranked 11th so a win there would be worth a nice few points I presume?

I see Argentina are fairly comfortable in top spot. Strange considering they're struggling in the South American qualification, winning less than half their matches and only occupying the playoff spot currently. They did reach back to back Copa America finals I suppose as well as the World Cup final, but lost them all.

[tetsujin1979]

Finish the year in 23rd, although there was only a few internationals since the last update and few teams changed their rank: http://www.fifa.com/fifa-world-ranking/ranking-table/men/index.html
Wales 11th, Austria 31st, Serbia 44th, Georgia 118th, Moldova 164th

[BonnieShels]

So here we go...

Us'uns at 26... (how apt)
Them'uns at 26 (Spleugh)

http://www.rte.ie/sport/soccer/2017/0406/865703-republic-and-northern-ireland-level-in-fifa-rankings/

NB. Them'uns are slightly ahead on 823.23 (vs our 822.62). Both rounded up to 823.

[Gather round]

Afternoon all. I'll admit I wasn't making much sense up there re the Elo ranking system.

NI are currently ranked #29 in Europe using it (#51). Given that

a) rankings are claimed to converge on a team's true strength relative to its competitors after about 30 matches, and

b) we've played 26 since September 2014 (19 competitive, 7 friendly)

can we assume that should we make the play-offs there'd be a rise into the top 20?

[Closed Account]

This website is pretty useful:
http://www.football-rankings.info/

Here is the current standings for the Nations League which will be determined after the World Cup qualifiers are over. (5-7 games) (http://www.football-rankings.info/2017/03/uefa-nations-league-update-31-march-2017.html)
1 1 Germany 34346
2 3 France 32508
3 4 Portugal 32415
4 2 Belgium 31782
5 5 Spain 30990
6 6 England 30891
7 8 Italy 30346
8 9 Switzerland 30006
9 7 Russia 29258
10 10 Croatia 28319
11 11 Poland 28022
12 12 Iceland 26215
----------------------------------
13 14 Austria 26198
14 19 Bosnia and Herzegovina 25340
15 15 Wales 25209
16 17 Republic of Ireland 25009
17 13 Netherlands 24925
18 20 Slovakia 24894
19 21 Sweden 24686
20 16 Ukraine 24666
21 23 Czech Republic 24167
22 18 Hungary 24126
23 24 Turkey 23758
24 26 Northern Ireland 23467

And based on simulations, how likely we are we qualify for the World Cup (64.65%) (http://www.football-rankings.info/2017/03/2018-world-cup-simulations-31-march-2017.html)

AFC: Iran, Japan, Korea Republic, Australia, Saudi Arabia

CAF: Egypt, Nigeria, Côte d'Ivoire, Congo DR, Senegal

CONCACAF: Mexico, Costa Rica, USA

CONMEBOL: Brazil, Colombia, Chile, Uruguay, Argentina

UEFA: Germany, Poland, France, England, Spain, Switzerland, Belgium, Croatia, Republic of Ireland, Italy, Portugal, Serbia, Iceland



These would be the seeds: Russia, Brazil, Germany, Argentina, Poland, Chile, Portugal, Colombia.

[geysir]

Earlier it was mentioned that Romania didn't play friendlies in order to boost their first pot seeding potential in the WC qual draw. I see the main benefit of that ruse has fallen to Poland who are cruising at the top after beating the (false) first seeds 3 nil :)

[geysir]

This website is pretty useful:
http://www.football-rankings.info/

This web site (http://www.footballseeding.com/national-ranking-uefa/ranking-2017/) describes in more detail and in as simple way as possible, where a country stands in the Uefa national rankings and how it got there, use of shades and different font colours is a bonus.

[DannyInvincible]

Earlier it was mentioned that Romania didn't play friendlies in order to boost their first pot seeding potential in the WC qual draw. I see the main benefit of that ruse has fallen to Poland who are cruising at the top after beating the (false) first seeds 3 nil :)

joe_denilson said the following after our recent loss against Iceland:


Should have used a 7th sub for ranking reasons

Would having used a seventh sub, beyond the six we actually used, have nullified the loss, as far as ranking points were concerned?

[Charlie Darwin]

joe_denilson said the following after our recent loss against Iceland:



Would having used a seventh sub, beyond the six we actually used, have nullified the loss, as far as ranking points were concerned?
Subs are agreed before the match - the fourth official would have just said no.

[DannyInvincible]

Subs are agreed before the match - the fourth official would have just said no.

You mean the fourth official would have refused us permission to use a seventh sub against Iceland if we'd requested such permission? So can the teams agree before the game to limit themselves to, say, three subs or seven subs - whatever their mutual preference may be - and the fixture would still count towards both teams' ranking either way?

[Closed Account]

http://uk.reuters.com/article/uk-soccer-world-belgium-luxembourg-idUKKBN0EF18T20140604

FIFA's rules limit the number of substitutes allowed in a competitive match to three - and six for friendlies.
Belgium made seven changes during the match, in which Nacer Chadli and Kevin De Bruyne also scored for them and Aurelien Joachim for Luxembourg.
FIFA said in an email that it would not recognise the game as an "A" match.
Maybe the 4th official could have stopped it but worth a try.

[Closed Account]

However,
Spain tried something similiar
http://www.bbc.com/sport/football/25138160

But Fifa ratified the game because Spain lost, so they likely would have ratified our Iceland loss also.

[geysir]

Man up Danny, Ireland were shamed and given a football lesson in their own back yard by a tiny populated arctic island. Just take the beating, the null ranking points and move on :)
As it turned out, Iceland got more ranking points from beating poncy Ireland in a friendly than they did for beating a grit laden Kosovo away in a WC qual, in a duel to the death in Albania (xxx more scary than Transylvania after sunset).

[Charlie Darwin]

I think if either side gave a football lesson that day then it's one everyone in attendance would be better off forgetting. One of the worst games I've seen, and I've seen us play a lot.

[swinfordfc]

In June 1st rankings - Ireland will hold 26th placed by themselves with Northern Ireland slipping to 28th! Don't know why the reason though! Does Ireland v Mexico match have any bearing on these rankings in June 1st?

[geysir]

Will the Fifa ranking position play a part in the seeding for the WC play off draw? It could be interpreted that the top 4 seeded teams will be decided on position in the runner-up table.
WCQ Uefa round 2 (http://www.fifa.com/worldcup/preliminaries/europe/road-to-russia.html)

The ranking position won't matter at all when it comes to the draw for the group stages of the world cup, for those teams outside the top 7.

[zero]

Will the Fifa ranking position play a part in the seeding for the WC play off draw? It could be interpreted that the top 4 seeded teams will be decided on position in the runner-up table.
WCQ Uefa round 2 (http://www.fifa.com/worldcup/preliminaries/europe/road-to-russia.html)

The ranking position won't matter at all when it comes to the draw for the group stages of the world cup, for those teams outside the top 7.

they haven't even said whether or not the playoff draw will be seeded (have they?) but it has been in the past. it was fifa ranking that was used for 2014 though so one would expect similar.

[geysir]

they haven't even said whether or not the playoff draw will be seeded (have they?) but it has been in the past. it was fifa ranking that was used for 2014 though so one would expect similar.
They didn't divulge the play off seeding criteria in the WC 2018 competition regulations but I see they didn't divulge this either in the 2014 competition regulations. It's most probably as you say, seeding by Fifa rank is the (secret) default.

[swinfordfc]

What's the latest on this?

[boysingreen]

What's the latest on this?

We've taken over the world.

[Closed Account]

Likely to be ranked 29th

Selected Others:
England 13th
Wales 20th
Northern Ireland 22nd
Scotland 60th

Austria 38th
Serbia 43rd
Georgia 113th
Moldova 158th

[DannyInvincible]

Likely to be ranked 29th

Yup, we're down three places to 29th despite the draw against Austria and victory over Uruguay: https://www.balls.ie/football/fifa-world-rankings-july-2017-368386

Holland are now in 32nd, which is their joint-worst ever ranking. Peru are a surprise in 14th.

[geysir]

Yup, we're down three places to 29th despite the draw against Austria and victory over Uruguay: https://www.balls.ie/football/fifa-world-rankings-july-2017-368386
If you click on Ireland (http://www.fifa.com/fifa-world-ranking/associations/association=irl/men/index.html) in the Fifa rankings list, you'll not only see the ranking points gained but also the ones that fall off the radar and Ireland had two good wins fall off, reducing the effect of the recent win and a draw.

[Diggs246]

What's clear is that we need to win the group or I fear we will lose the play off, judging by the list of top seeded teams who may come second in their respective groups

http://www.independent.ie/sport/soccer/international-soccer/northern-ireland-leapfrog-martin-oneills-men-in-latest-rankings-and-it-could-be-huge-very-significant-35901614.html

[swinfordfc]

We relied to much here on drawing games - like drawing against Austria was poor and it will cost us later this year! ... Cant afford to that in these competitions!

[DeLorean]

We took four points off Austria, the top seed in the group when the draw was made. I doubt that'll be what ends up costing us. After the first match in Belgrade people were saying those "two points dropped" will cost us later on. Anybody would think we're the only team dropping points in the group. We're in a position we could only have dreamt about really. Any kind of an ugly win in Georgia will set us up for a home match that could all but guarantee top spot.

I'm as frustrated as the next man at some of the conservative tactics and team selections, but we're exactly where we would have wanted to be. It's only results from here on that can't really damage us I think.