Latest updates

Check the Important info page for latest updates! (31 October 2014)
Twitter Linked In

Friday, June 27, 2014

Confederation weightings update (27 June 2014)

At the end of the group stage.

CONMEBOL still on top, while UEFA have done some damage limitation with their latest wins. CONCACAF drop due to their recent defeats.




CONMEBOL 1
UEFA 0.98
CONCACAF, CAF, AFC, OFC 0.85




2006 2010 2014 2006-2014 av 06-14 weight
UEFA total games 34 34 29 97


win 26 20 16.5 62.5


av per game 0.76 0.59 0.57 1.92 0.64 0.98
CONMEBOL total games 17 24 18 59


win 10.5 16 14 40.5


av per game 0.62 0.67 0.78 2.07 0.69 1.00
CONCACAF total games 13 11 12 36


win 2.5 4 6.5 13


av per game 0.19 0.36 0.54 1.09 0.36 0.85
AFC total games 14 14 12 40


win 3.5 5 1.5 10


av per game 0.25 0.36 0.13 0.74 0.25 0.78
CAF total games 16 20 15 51


win 4.5 6.5 4.5 15.5


av per game 0.28 0.33 0.3 0.91 0.3 0.81
OFC total games 0 3 0 3


win 0 1.5 0 1.5


av per game 0 0.5 0 0.5 0.17 0.70

11 comments:

  1. Edgar,

    I notice a few things in your overview that are not as expected:
    UEFA 2006: should be 25,5 pts from 34 matches (22 wins, 7 draws)
    CONMEBOL 2006: should be 11.0 pts from 17 matches (10 wins, 2 draws)
    CONMEBOL 2010: should be 15.0 pts from 24 matches (12 wins, 6 draws)
    CAF 2010: should be 7.0 pts from 20 matches (4 wins, 6 draws)
    AFC 2010: should be 5.5 pts from 14 matches (4 wins, 3 draws)

    Resulting average per confederation should be:
    UEFA 0.6357
    CONMEBOL 0.6833
    CONCACAF 0.3659
    CAF 0.3104
    AFC 0.2560
    OFC 0.1667

    Resulting weight per confederation should be:
    UEFA 0.9821 => 0.98
    CONMEBOL 1.0000 => 1.00
    CONCACAF 0.8554 => 0.86
    CAF 0.8210 => 0.85
    AFC 0.7823 => 0.85
    OFC 0.7028 => 0.85

    ReplyDelete
    Replies
    1. FIFA consider PSO wins as full wins. See this PDF.

      Delete
    2. Thanks, you're right (of course).

      Strange assumption from FIFA though, as in the ranking calculation a PSO win isn't considered a full win (it isn't exactly considered a draw neither, but okay...)

      Delete
  2. Few interesting notes:
    despite having 5 teams in the knockout round only 1 in the first round can effect CONMEBOL (Argentina). In fact this is true for the first 2 rounds (assuming Argentina wins their first match)

    Also if we play out what Edgar posted as the "probable results":
    CONMEBOL 1
    UEFA 0.98
    CONCACAF, CAF, AFC, OFC 0.85

    So basically no change from the current formula. However if Argentina lose prior to Netherlands, things can change a little bit.

    For example if everything else holds true and the Swiss beat Argentina, then UEFA are up to .99 and can even get to the full one if they can take out Brazil twice.

    ReplyDelete
  3. It seems brutal for CONCACAF to have their best performance, with 3 of 4 teams making the knockouts and Costa Rica making the quarterfinals, only to have their coefficient fall to the basement, level with AFC (4 teams, 0 wins) and OFC.

    ReplyDelete
    Replies
    1. Actually their performance on the WC14 is slightly worse than on the WC02 (0.55 in WC02 and 0.47 in WC14), which will be dropped after this WC is finished. So their average over the past 3 WC's is affected a bit by that. Combined with the extraordinary performance of the CONMEBOL-countries, CONCACAF drops below the 0.85 mark.

      I'm sure that if CONCACAF is able to do as well in WC18 as they did in WC14, their coefficient will rise again above the 0.85 mark.

      Unless FIFA changes the rules again, which always seems a possibility...

      Delete
    2. I understand that the 2014 wins/game calculation is lower, but that's because CONCACAF earned a fourth team into the World Cup (as compared to only 3 in 2002). The current system has a perverse incentive to have marginal teams from a confederation LOSE the inter-continental playoff if that lowest-ranked team would do worse than the average other team from that confederation at the finals. This is not always true (see Uruguay 2010), but CONCACAF would have had a better statistic had T&T 2006 not qualified for the World Cup and then been smashed.
      I can fully believe that the current ranking system is an improvement on previous systems, but it still strikes me as a process of "throw a bunch of numbers against the wall to get a result that makes sense and use that formula for now". I expect within 5 or so years a new formula will be introduced, one which will incorporate margin of victory to some extent and one which doesn't automatically 0 out on a loss (losing 3-4 to the #1 team and losing 0-4 to the #60 team are treated exactly the same, which is ridiculous).

      Delete
  4. I would highly favour the coefficients to play a part in the determination of the number of WorldCup berths for each confederation.

    My suggestion would be that the average points per game should be a factor. You could use a simple formula like this:

    Bt18 = ((Av14/Av10)*Bt14)*31/BtTot18

    Bt18 = numer of berths for WC18
    Av14 = average points per game after WC14 (WC06, WC10 and WC14 included)
    Av10 = average points per game after WC10
    Bt14 = numer of berths for WC14
    31 = numer of teams that would qualify for WC18 (Russia excluded as hosts)
    BtTot18 = total of berths for WC18 for all confederations after initial formula

    If you do this, this results would come up:

    AFC : 3,680
    CAF: 4,383
    CONCACAF: 3,380
    CONMEBOL: 5,115
    OFC: 0,521
    UEFA: 13,921

    If you consider there would be 2 intercontinental play-offs, the rounded numer of berths could be this (between brackets the berths for WC14):

    AFC : 3,5 (4,5)
    CAF: 4,5 (5)
    CONCACAF: 3,5 (3,5)
    CONMEBOL: 5 (4,5)
    OFC: 0,5 (0,5)
    UEFA: 14 (13)

    So OFC and CONCACAF would hold their numer of berths. UEFA would win 1 at the expense of AFC. CONMEBOL would get a certain 5th spot, with CAF having 4 certain spots and 1 in the intercontinental play-offs.

    ReplyDelete
    Replies
    1. Tobcoach, I like this idea.
      No massive changes in number of berths as a result, but past performances do count in fair measure:
      AFC performed really bad this edition and as a result loses one berth, CONMEBOL performed extraordinary well and gains a half berth.

      I only have some problems with the behaviour of this formula with regard to UEFA. Their average for this edition is at the moment 0.64, just a little better than their average for WC10 (0.62). The formula results in (whoa!) 1 whole berth extra. I think it's caused because their starting point is already 13 berths (almost thrice the amount of the next biggest confederation) which acts like some sort of flywheel...
      For next World Cups this effect will (with average to good performances of the UEFA-teams) only become bigger.

      Delete
    2. Ed, you're right ofcourse, but the formula is a W.I.P. (Work in progress). Maybe the formula should be slightly changed to rule out the advantage of already having high number of berths and thus favouring UEFA.

      Or maybe because of the high number of UEFA berths, you should give UEFA 13 berths, no matter what the results are. If you do that in this occasion, you would get:

      AFC: 4
      CAF: 4,5
      CONCACAF: 3,5
      CONMEBOL: 5,5
      OFC: 0,5
      UEFA: 13

      So the big change would be AFC en CAF which would lose 0.5 berth in favour of CONMEBOL.

      Delete
    3. Then UEFA can never grow, even if they fully deserve it. Instead, maybe it's a good idea to scale for the original number of berths:

      Bt18 = ((1 + Av14/Av10) * Bt14) * 31/BtTot18

      This leads to:
      AFC: 4
      CAF: 4.5
      CONCACAF: 3.5
      CONMEBOL: 5
      OFC: 0.5
      UEFA: 13.5

      AFC and CAF would each lose a half berth in favor of the only two confederations with an increase of their average, CONMEBOL and UEFA.

      Delete