Jump to content

Canada up to 56 in the FIFA Rankings


Guest Can. in UK

Recommended Posts

Some months back, Craig Forrest & Gerry Dobson did a piece on Soccer Central expressing dissatisfaction and befuddlement with the rankings. I started to write a response but got thrown "off the ball" and never finished or submitted it, but I'll post it here as my effort at explaining the problems with the ranking system. BTW I agree with the above comment that the Elo system (used in chess rankings) is a far superior ranking system.

---

Dear Craig & Gerry,

You correctly noted in your show today that the results of the new FIFA ranking system seem a little off, with rankings fluctuating wildly from month to month, and little apparent rhyme or reason to it.

Let me tell you why this is so, and how we can easily fix it.

The procedure for calculating the New FIFA Rankings is clearly explained here (PDF):

http://fifa.com/documents/tournaments/FIFA_WR_revision_E.pdf

Summary Evaluation of New Ranking System:

1. The method for awarding ranking points for a match is an improvement over the previous method. I note in passing though, that the “Strength of opposing team” factor is linear, probably undesirably so. This means, for example, that more points are awarded for beating 100th-ranked Congo twice than for beating 1st-ranked Brazil once (ignoring confederations and importance of match).

2. The Confederation Weightings detract from the elegance of the model. A confederation is strong because the teams therein win their games, not because those teams are awarded a higher weighting for the games due to their neighbourhood. These weightings are needed, due to the aforementioned problem with the linearity of the opponent’s strength factor – they are a clumsy effort to correct the effect of a strong team in a weak confederation appearing too strong. They could be done away with, with more reasonable (non-linear) awards for opponent strength. Still, it is not these Confederation Weightings that produce the “herky-jerky” rankings that you identified.

3. The “Period assessed” weightings are the culprit. It is sensible to count recent results as more important than distant ones, but it is ill-considered to have such large discontinuities (100% - 50% - 30% - 20%) in the time weightings for each year’s results. Why?

The Problem – A Discontinuous Time Weighting Function

The current FIFA ranking system weights the importance of results according to how recent they are – results during the last 12 months count in full (100%), those from the previous year count 50%, those from two to three years count 30%, and those from three to four years count 20%.

Month Old Weighting New Weighting

1 N/A (0%) 100%

13 100% 50%

25 50% 30%

37 30% 20%

49 20% 0%

The problem with this system is that results which occurred last month, 13 months ago, 25 months ago, 37 months ago and 49 months ago cause all (100%) of the change in the monthly rankings. No other results in the last four years impact monthly changes at all. The other 44 months have no effect on monthly rankings changes. Is this bizarre effect what the rankings system designers intended?

In the attached Excel workbook, the discontinuity (or “jaggedness”) of the current time weighting function is shown in the “Data” worksheet and accompanying charts, contrasted with the continuity of two other functions, #1 (non-linear) and #2 (linear) which I put forward as possible replacements.

The Solution – A Continuous Time Weighting Function

If the discontinuous time weighting function is the source of our grief, well then, what to do about it? The answer, clear and simple, is to replace the current nonsensical discontinuous “100-50-30-20” function with a continuous function.

not only as you pointed out because recent results seem to move ratings a lot, but also because 4-year-old results that are dropped out of the calculations can also cause abrupt changes to the rankings.

Similarly, each month results from 1 year ago that last month were valued at 100% are now valued at 50%, 2yr -> 3yr ... (unfinished)

Link to comment
Share on other sites

Guest speedmonk42

Games played in the ELO rankings is kind of interesting.

Sweden comes out on top with 877.

We are almost at 1/3 that.

Link to comment
Share on other sites

quote:Originally posted by trueviking

how can a team's ranking do this in a single 12 month period, when they played only 6 games during that time?...twice ranked in the mid 50's and dropping below one hundred in between?....it seems to based on nothing.....it certainly can not be an indication of the teams quality if such wild variation can occur in such a short time?....is this team really 50 spots better today than it was 8 weeks ago?

untitled1oc6.jpg

As a self-professed math/stats nerd, I have to agree with you on this.

The main problem with this ranking system are that the swings are too wild.

Never mind the fact that their is a Euro-bias in the calculation.

Link to comment
Share on other sites

quote:Originally posted by Massive Attack

As a self-professed math/stats nerd, I have to agree with you on this.

Never mind the fact that their is a Euro-bias in the calculation.

Euro bias indeed. How does Brazil who beat England and Argentina drop as well. makes no sense. Perhaps maybe when the South American 2010 WCQ starts up this fall, we will see the South American teams shoot up from the ground up again.

As for Canada's "Bump" in the FIFA rankings, I am very pleased and excited, only that most likely we will drop like a sack of hammers based on the new FIFA ranking system within a few months if we dont play games.

Bottom line for Canada: Play more friendlies, then we go up like no tommorow. Dont play, and we fall faster in the rankings then a brick falling from a 20 storey apartment.

The main problem with this ranking system are that the swings are too wild.

Link to comment
Share on other sites

The Euro bias is based on a multiplyer FIFA puts on the results of each confed. If Canada were to beat Bulgaria 2-0 they would get less points then if England beat Bulgaria by the same score. That is totally unfair.

I believe FIFA considers UEFA and South America equal. Its all the lower confeds that get penalized by this bias in the math.

One other note, I dont' think playing games is enough. You have to actually win them as well. This ranking system punishes losers greatly.

Link to comment
Share on other sites

quote:Originally posted by Daniel

The new rankings does NOT punish losers, as you simply get 0x X points (a win is 3x and a draw is 1x).

Exactly. If you look at a group of teams playing a closed set of matches (like a regional tournament) you'll notice that a majority of the teams (60-80%) will see a subequent in increase their ranking. But in theory, this should be a zero-sum scenario where half go up and half down.

Inactivity might be bad for team unity and cohesion, but it doesn't *automatically* make for a worse side.

Link to comment
Share on other sites

Guest Jeffery S.
quote:Originally posted by jonovision

Exactly. If you look at a group of teams playing a closed set of matches (like a regional tournament) you'll notice that a majority of the teams (60-80%) will see a subequent in increase their ranking. But in theory, this should be a zero-sum scenario where half go up and half down.

Inactivity might be bad for team unity and cohesion, but it doesn't *automatically* make for a worse side.

I thought you could get points by scoring goals and losing, especially away. Is that no longer the case?

Regardless, what would be the right thing to do, since they have all the data anyways, is to simply allow anyone to look at a country and look at how it picked up points from game to game, with even a basic explanation of why. That would not be at all difficult, I don't see why FIFA could not do that. I would like to know what we got vs. Bermuda, what vs. Venezuela, what C Rica.

Indeed I am sure a simple software would allow you to stick in two teams, indicate the kind of game and who was home, put in their current ranking, stick in a score, push a button and voilà! the ranking points gained by the win.

They are probably just worried about losing the chance for a dramatic press conference and press releases every two months, since the rankings announcement has hefty sponsorship.

Link to comment
Share on other sites

quote:Originally posted by Jeffrey S.

I thought you could get points by scoring goals and losing, especially away. Is that no longer the case?

Regardless, what would be the right thing to do, since they have all the data anyways, is to simply allow anyone to look at a country and look at how it picked up points from game to game, with even a basic explanation of why. That would not be at all difficult, I don't see why FIFA could not do that. I would like to know what we got vs. Bermuda, what vs. Venezuela, what C Rica.

Indeed I am sure a simple software would allow you to stick in two teams, indicate the kind of game and who was home, put in their current ranking, stick in a score, push a button and voilà! the ranking points gained by the win.

They are probably just worried about losing the chance for a dramatic press conference and press releases every two months, since the rankings announcement has hefty sponsorship.

What you are describing is pretty much the old point scheme. It was abandoned because it was deemed too complicated. The current one is much simpler, but much less indicative. Teams have jumped 30+ spots up or down every month!

Check out the 1999 and 2006 revisions: http://en.wikipedia.org/wiki/FIFA_World_Rankings#1999_update

Link to comment
Share on other sites

quote:Originally posted by jonovision

Exactly. If you look at a group of teams playing a closed set of matches (like a regional tournament) you'll notice that a majority of the teams (60-80%) will see a subequent in increase their ranking. But in theory, this should be a zero-sum scenario where half go up and half down.

Inactivity might be bad for team unity and cohesion, but it doesn't *automatically* make for a worse side.

my sentiments exactly...the rankings currently are largely a representation of the quantity of games a team plays and not the quality of the team in the games they do play...however many....this is especially true for the lower ranked teams who play fewer games...the big boys all play a similar number of games so this disparity doesnt affect them as much.

Link to comment
Share on other sites

quote:Originally posted by torontofc

WTF!?!?!?!?

Did Linford bribe FIFA?

lol nope, just Jack Warner.

Fifa sent them both to a disiplinary committee. After carful deliberation, Fifa declared that Jack Warner was not at fault in anyway. Colin Linford is set to hang from the flag pole of Fifa headquarters at dawn. Tobagan rum coocktails will follow at 5.[:o)]

Link to comment
Share on other sites

quote:Originally posted by trueviking

untitled1oc6.jpg

And as soon as the Gold Cup falls out of the 1-year window, African nations will soar pass the CONCACAF nations. And when the African Nations' Cup falls out of the one year window, Euro 2008 nations will soar pass them and CONCACAF nations will fall further still.

mickey%20mouse.jpg

As much as the UEFA coefficient system is ridiculed, they actually got something correct -- the last 5 years should be weighted equally. Although in the case of FIFA rankings, it would make sense to just look at the last 4 years and weigh them all evenly.

Link to comment
Share on other sites

quote:Originally posted by Canucks fan

http://fifa.com/associations/association=can/ranking/gender=m/index.html

according to this it was just the Venezuela match that counted in the last rankings.

Hmmm Im positive the Costa Rica match also counted, since it wouldnt make sense that a tie in a friendly would mean we jump 38 spots.. Also I remember reading that the deadline for matches to be counted is the Thursday before the ranking comes out (which is always a Wednesday) therefore the Costa Rica match should be counted by this logic since it was the Wednesday before, a day before the deadline.

Link to comment
Share on other sites

  • 2 weeks later...

An interesting point about Guadeloupe. If you look on the the FIFA site for International results and fixtures, you will notice that they dont exist. Any matches played against them are omitted, its as if we never played them. They are not on their database of teams. It looks like, whatever results we or our rivals have against Guadeloupe in GC or otherwise wont count for ranking purposes either.

Link to comment
Share on other sites

Archived

This topic is now archived and is closed to further replies.

×
×
  • Create New...