Joined: April 7th, 2007, 12:28 pm

January 21st, 2018, 12:36 pm #101

TheRoonBa wrote:In terms of teams taking it more seriously - I think they are limited to how good their domestic league is, and in some ways, punished by having better domestic leagues that export players.
Also the relative poverty of the country.  Even if the league is not particularly good, the best players in a poor country will seek clubs elsewhere if the conditions at home are crap.  Whereas in some relatively wealthier African countries (particularly South Africa), the league can be considered fairly strong but doesn't export as many players because the incentive to go abroad is not there.   A more extreme example of this is the Gulf.
Quote
Like
Share

Joined: August 4th, 2012, 4:06 pm

January 23rd, 2018, 1:40 am #102

TheRoonBa wrote:
Luca wrote:
TheRoonBa wrote:
to give the tournament more prestige.
But to give the tournament more prestige, shouldn't they upgrade the status of the final tournament matches, as now they're regarded just as "friendlies"?
It has more prestige as the matches are considered as A matches, when they would ordinarily not even be listed on FIFA's site, as they are played by restrictive national teams, bound by regulations of player eligibility.

Now all CHAN tournament matches are considered for the rankings and labelled as official FIFA matches - before they weren't.  That's an increase in prestige.

Why would they do this, when does the insanity stop. FIFA are rabid animals.
Quote
Like
Share

Joined: October 31st, 2006, 5:16 pm

January 23rd, 2018, 4:35 am #103

Well that escalated quickly...
Quote
Like
Share

Joined: November 3rd, 2006, 11:49 am

January 23rd, 2018, 6:24 am #104

do not offend animals, please
Quote
Like
Share

SDb
Joined: January 24th, 2013, 5:10 am

July 16th, 2018, 12:04 pm #105

FIFA are about to change the rankings once again!
If they apply it right after the World Cup, Germany will be number 1. Totally ridiculous but hey it's Fifa.

https://resources.fifa.com/image/upload ... fwyunjbkha
www.soccer-db.info - football internationals
Quote
Like
Share

Joined: April 5th, 2012, 10:54 am

August 16th, 2018, 8:23 am #106

Quote
Like
Share

Joined: October 31st, 2006, 5:16 pm

August 16th, 2018, 10:19 am #107

Some interesting considerations:

1) Friendly matches during FIFA windows are weighted more heavily than other friendly matches (10 instead of 5).
2) Tournament progression also affects the weighting - so the final of a World Cup will be weighted more than the quarter-final.
3) No negative points during knockout stage of "championship tournaments".

Not sure about number 3 - this will of course mean it's no longer a zero-sum formula (points are exchanged between teams - 10 added to one team, 10 taken from the opponent, for example).  Also, teams can only lose once during a knockout stage (except if they lose a semi and then the 3rd place match) - so I'm not sure it would have so much of a negative effect.

To take a real example from eloratings.net for the semi-finalists of the World Cup.
France beat Belgium and Croatia (+29 and +30) = +59
Croatia beat England, lost to France (+34 and -30) = +4
Belgium lost to France, beat England (-29 and +36) = +7
England lost to Croatia, lost to Belgium (-34 and -36) = -70

Belgium had a better points increase than Croatia, despite finishing 3rd and Croatia finishing 2nd - but if we look at the match results, they both played the same 2 teams, and Belgium had a better score in each game (2-0 v England instead of 2-1, 0-1 v France instead of 2-4).

England were the worst losers here - losing 70 points, and in fact dropping from 5th to 10th.

Another variation sees winners of penalty shootouts being given an "enhanced draw" in the calculation.
Win = 1
Draw = 0.5
Loss = 0
Win after penalty shootout = 0.75.

However, Loss after penalty shootout is the same as draw (0.5) - which means it will not be a zero sum exchange of points for that game.
Penalty shootouts to decide aggregate winners do not count (team loses 1-0 in away leg, wins home leg 1-0, then wins on penalties - this is treated as 2 normal 1-0 wins.)

In my opinion, if you are going to give benefits to a team who wins a penalty shootout, you have to also give a negative effect to the team that loses. (0.25 would be a more sensible factor for a team losing a penalty shootout, as technically they have achieved less than a draw.

In determining initial ratings (converting from old FIFA ratings to elo-based ratings), the formula was as follows:
For the initial seeding of teams in the new FWR, teams were evenly distributed over a range of approximately 800 to 1600 points. The point difference between two adjacent teams was set at 4 points. 

For me - this originally sounded stupid - a gap of 400 points in the FIFA rankings and a gap of 2 would be treated exactly the same.  Obviously if a team is 400 points ahead of another, the difference in ability is expected to be greater than if they are 2 points apart.  However, the gap becomes 4 regardless of the previous gap between 2 adjacent teams.  It would have been relatively easy to distribute the teams from 800 to 1600 points while also taking into account relative differences between the points totals of each team. However, because FIFA's previous points system did not really reflect differences in strength between teams, with huge gaps between some of the teams in the Top 10, I think this was a good decision.
Quote
Like
Share

Joined: April 7th, 2007, 12:28 pm

August 17th, 2018, 1:56 pm #108

TheRoonBa wrote:For me - this originally sounded stupid - a gap of 400 points in the FIFA rankings and a gap of 2 would be treated exactly the same.  ... However, because FIFA's previous points system did not really reflect differences in strength between teams, with huge gaps between some of the teams in the Top 10, I think this was a good decision.
Two wrongs making a right, there...

Why didn't they just apply the new rankings to old results, over the last N years, to generate the new rankings from scratch?   If the new system is supposed to be better, give it a chance to work, rather than using the old rankings as a poor initial condition.

Or at least, generate the new rankings from scratch and then compare vs the old ones, with some fudgy halfway house interpolation.
Quote
Like
Share

Joined: October 31st, 2006, 5:16 pm

August 19th, 2018, 5:16 pm #109

Yes - they have a kind of rule that rankings cannot be changed once they are published, even if an error is discovered, but I don't think there is any rule that stops them applying a different system over previous rankings (and not changing the previous rankings themselves).  However, doing so would affect continuity, as it would mean that in the first "new" release, some teams may move up 80 places, or down 80 places.  This would thus have the effect of making their previous system look totally shit.  By extending the period of "convalescence", these changes are spread out over time, and the transition is smoother.

I guess an analogy would be being 100% ill, and being told that taking 50ml of medicine would cure you instantly and make you 0% ill.  FIFA have chosen to drip feed the medicine and prolong their illness., so they can give the impression they were never ill in the first place.
Quote
Like
Share

Joined: April 7th, 2007, 12:28 pm

August 19th, 2018, 6:37 pm #110

Which teams stand to benefit most/least from the new system being phased in this fudgy way, instead of computing fresh rankings using the new system on old results?
Quote
Like
Share

Joined: October 31st, 2006, 5:16 pm

August 20th, 2018, 6:37 am #111

I think it should be a straightforward "the teams that are ranked much higher on the eloratings website than they were in the FIFA rankings" would benefit least (as they will have to wait longer to get their "correct" points).  And those benefiting most will be "those teams ranked much lower on the eloratings website than they were in the FIFA rankings", who will take longer to descend the rankings.  But any benefits or negative effects will be temporary, as eventually, all teams should find their "correct" level, 

However, it may take a while for that to happen, by which time teams will no longer remember anything about their previous form.
Quote
Like
Share

Joined: April 7th, 2007, 12:28 pm

August 20th, 2018, 7:01 pm #112

TheRoonBa wrote:I think it should be a straightforward "the teams that are ranked much higher on the eloratings website than they were in the FIFA rankings" would benefit least...
I know but I was asking specifically for some names, as I can't be arsed to compare the lists myself.
Quote
Like
Share

Joined: April 5th, 2012, 10:54 am

August 21st, 2018, 4:14 am #113

nfm24 wrote:
TheRoonBa wrote:I think it should be a straightforward "the teams that are ranked much higher on the eloratings website than they were in the FIFA rankings" would benefit least...
I know but I was asking specifically for some names, as I can't be arsed to compare the lists myself.
Must... resist... temptation... Oh, very well: clicky
Quote
Like
Share

Joined: April 7th, 2007, 12:28 pm

August 21st, 2018, 7:51 pm #114

Hmm interesting.  I suppose we should have predicted most of the bottom 5 based on non-footballing considerations (suspensions/noobage etc).  Also quite Africa heavy towards the bottom, possibly something to do with CHAN being included by FIFA?

Perhaps some weighting of the raw delta by the max of the Elo/FIFA rankings would help to weed out minnow noise, as obviously, a differential of 10 places is more meaningful higher up in the ranking.
Quote
Like
Share

Joined: April 7th, 2007, 12:28 pm

August 21st, 2018, 7:52 pm #115

BTW if you still can't resist temptations, how about a countrywise list of the best/worst results achieved in terms of Elo points....  Possibly this is already generatable on the Elo website.
Quote
Like
Share

Joined: April 5th, 2012, 10:54 am

August 22nd, 2018, 5:18 am #116

As far as I can see there is a list of biggest upsets (top 15 matches with most points exchanged).

A country-by-country list would require looking at each individual list. So I just queried my own dataset (shouldn't differ too much): clicky.

There is a second table at the bottom, ignoring the competition multiplier, so you'll see the largest upsets regardless of competition.
Quote
Like
Share

Joined: December 26th, 2008, 11:44 am

August 22nd, 2018, 5:30 am #117

Quote
Like
Share

Joined: June 7th, 2007, 8:58 am

August 22nd, 2018, 8:05 am #118

There is a Highest Rank / Rating on the Columns submenu, if that's what y'all are looking for...
Quote
Like
Share

Joined: December 26th, 2008, 11:44 am

August 22nd, 2018, 8:15 am #119

How would have been the highest rank of England ever if the matches played by England amateurs in the 1906-24 era are not included? At the time the England amateur team used to be a restrictive team already. Not the A-team.

I remember until a few years ago these matches were not included and England's highest rate ever was just under 2100. That was achieved just after the 1966 World Cup if I am right.
Quote
Like
Share

Joined: June 7th, 2007, 8:58 am

August 22nd, 2018, 11:05 am #120

I thought that topic was discussed extensively elsewhere, tl;dr being that in those times there was no fine line between "amateur" and "full" teams, "full" teams often had many "amateur" players, and "mostly-amateur" and "amateur-only" teams were at about the same level as "full" or "professional" teams if not better in some cases.
Quote
Like
Share

Joined: December 26th, 2008, 11:44 am

August 22nd, 2018, 11:45 am #121

the difference betweeh England A and England amateurs is:

England A (the full England team) is open to all players who are English. Both professionals and amateurs. So for England A there are no restrictions.
England amateurs is open to amateur players only, so a restrictive team as professional players were excluded to play for England amateurs..
Quote
Like
Share

Joined: October 31st, 2006, 5:16 pm

August 23rd, 2018, 8:24 am #122

As Kirill said, at that time, it's debatable whether the England A team was better than the England Amateur team, as many of the top players remained amateur out of principle, despite being better footballers than some professionals. Herein lies the problem of defining which team was the best.  In that particular era, "professional" didn't always mean better than "amateur".
Quote
Like
Share

Joined: December 26th, 2008, 11:44 am

August 23rd, 2018, 8:40 am #123

TheRoonBa wrote: In that particular era, "professional" didn't always mean better than "amateur".
I fully agree with this. The best English player at the time was Vivian Woodward, who was an amateur. My only objection to count matches played by the England amateur team is that it was a restrictive team. The full England team (which certainly was not a full professional team, but a mix of professionals and amateurs) was not.

The last amateur player that won a full England cap was Bernard Joy, on 9 May 1936 against Belgium https://en.wikipedia.org/wiki/Bernard_Joy
Quote
Like
Share

Joined: April 7th, 2007, 12:28 pm

August 28th, 2018, 5:41 am #124

Back to the countrywise best/worst results - it's interesting to compare the actual data (in terms of points exchanged) with the perception of fans.   This may be something to do with the fact that fans remember great/disastrous results as wins/losses, not necessarily with goal margin in mind.  If your team thrashed somebody, you perhaps generally don't think of the opposition as too impressive, compared to the drama of a battling performance with a 1-goal margin, last minute winner.  Likewise if you happened to play a previously great team on the downslide, and they later slid further, beating them perhaps doesn't live in the memory as the achievement that it was on paper.

Obviously, the way to account for that would just be to give two tables, the second one without goal margin  (or I guess it would be better presented as two sub-rows within the same table), but maybe it just gets untidy.

Some interesting examples - Cameroon's 1990 win vs Romania is apparently a better result than their 1-0 win vs Argentina.  North Korea's 4-1 win in Thailand is better than beating Italy.  N.Ireland winning 2-0 in Greece is better than beating England or Spain at home.

Perhaps the circumstances of a heroic win, in a competitive match, are amplified in fan memory if associated with a great goal, or a brilliant individual player performance, which can't be included in the ranking points.  And the exclusion of some competitions can remove a lot of the best results of some teams, e.g. Bermuda.

There are also quite a few teams who got their best and worst results against the same opponent, sometimes within just a couple of years (or in Grenada's case, one month).  This might also be partly due to a lack of diversity of opposition though, and also perhaps that the initial ranking was out of date - or, that the teams are unrankably inconsistent :-)

Similarly, sparse data may be an issue.  A few teams have fairly arbitrary entries, e.g. Dominica with 2 results in the 1970s and 80s - when it is quite likely that many of their pre-1990s results are not included in the ranking anyway.
Quote
Like
Share

Joined: October 31st, 2006, 5:16 pm

August 28th, 2018, 9:23 am #125

In 2014, before the Greece 0-2 Northern Ireland match (55 elo points exchanged), Greece were 331 elo points ahead of Northern Ireland (and 76 places in the ranking)

In 2005, before the Northern Ireland 1-0 England match (36 elo points exchanged), England were 486 elo points ahead of Northern Ireland (changed to 386 with Northern Ireland's home advantage - and 84 places in the ranking).

Northern Ireland beating England at home is indeed a better result than them beating Greece away if margin of victory is not considered.

In my opinion, the large discrepancy in points exchanged (Northern Ireland getting 19 more points for beating Greece than for beating England) is due to the arbitrary choice of K factor (1.5 for a win by 2 goals, 1 for a win by 1 goal) and the massive difference between scoring 1 or 2 goals.  A more gradual increase in K (instead of jumping straight from 1 to 1.5) would produce a different result.

Before applying the K factor for margin of victory, the Greece match would have been worth about the same as the England match for Northern Ireland (36.7 points).
Quote
Like
Share

Joined: April 7th, 2007, 12:28 pm

August 28th, 2018, 2:59 pm #126

Also the fan memory/folklore effect is greater if the result is against a local rival, and/or against a team who has historically been one of the elite (even if not on top form at the moment).  E.g. beating Brazil or Italy was always considered a scalp, even when they were crapper than usual.
Quote
Like
Share