Friday, September 20, 2013

FIFA Ranking: October 2013 probable ranking

As already announced by FIFA, the October ranking will be used to seed the European teams in the play-off draw (see the simulations).

The October ranking could also be used to seed the teams in the final World Cup draw (see the simulations).

Probable results computed using the Elo Ratings. See the list of results used for this ranking.

Based on these probable results:

Bulgaria will be the worst runner-up.

Pots for the European play-off draw:

Pot 1: Croatia, Portugal, Greece, Sweden
Pot 2: Ukraine, France, Romania, Iceland

Honduras will finish 3rd in the hex, while Mexico will face New Zealand in the World Cup play-off.

Colombia, Chile and Uruguay will qualify, while Ecuador will meet Jordan in the play-off.

Assuming FIFA use the October ranking to seed the final draw, the seeds will be: Brazil, Spain, Argentina, Colombia, Germany, Italy, Switzerland. The 8th seed will be Croatia (if they make it through the play-offs) or Netherlands (if Croatia fail to qualify).

Probable October 2013 rank - Team - Probable October 2013 points - +/- Ranking - +/- Points


1 Spain 1513 0 -1
2 Argentina 1307 0 44
3 Colombia 1262 2 82
4 Germany 1230 -1 -31
4 Italy 1230 0 31
6 Croatia 1149 4 98
7 Switzerland 1138 7 146
8 Netherlands 1136 1 78
9 Uruguay 1131 -2 5
10 Belgium 1107 -4 -52
11 Portugal 1096 0 67
12 England 1080 5 133
13 Brazil 1078 -5 11
14 Chile 1013 2 46
15 USA 1007 -2 11
16 Greece 983 -4 -33
17 Russia 926 -2 -42
18 Bosnia-Herzegovina 925 0 -9
19 Côte d'Ivoire 917 0 15
20 Sweden 891 2 55
21 Mexico 871 0 34
21 Ukraine 871 5 72
23 France 870 2 58
24 Ghana 860 0 45
25 Denmark 787 -2 -38
26 Ecuador 774 -6 -77
27 Venezuela 771 9 64
28 Romania 767 3 28
29 Serbia 749 14 83
30 Peru 744 4 12
31 Algeria 741 -3 -21
32 Norway 737 7 38
33 Nigeria 724 3 17
34 Costa Rica 723 -1 -10
35 Panama 721 0 -6
36 Czech Republic 715 -4 -23
37 Honduras 685 3 5
38 Slovenia 671 -9 -83
39 Turkey 670 10 57
40 Mali 668 -2 -36
41 Cape Verde Islands 662 3 2
42 Japan 661 0 -10
43 Montenegro 651 -16 -115
44 Hungary 636 -14 -108
45 Tunisia 632 1 -11
46 Iran 613 2 -20
47 Egypt 610 3 -1
48 Burkina Faso 609 3 2
49 Austria 596 -2 -43
50 Iceland 589 4 -10
51 Uzbekistan 582 6 3
52 Wales 573 0 -33
53 Paraguay 572 -12 -101
54 Armenia 570 1 -22
55 Korea Republic 569 3 -5
56 Bulgaria 565 8 16
57 Australia 564 -4 -39
58 Albania 563 -13 -89
59 Scotland 556 4 5
60 Cameroon 554 1 -4
61 South Africa 551 7 23
62 Republic of Ireland 550 -3 -20
63 Libya 543 7 23
64 Finland 538 -8 -53
65 Jordan 537 8 43
66 Senegal 530 0 -4
67 Slovakia 528 -7 -34
68 Israel 524 1 2
69 Zambia 513 2 -2
70 Guinea 512 6 25
71 Poland 503 -6 -36
72 Sierra Leone 493 0 -4
73 Cuba 492 11 54
74 Togo 488 3 2
75 Morocco 487 -1 -5
76 Dominican Republic 474 11 49
77 New Zealand 470 -10 -59
78 United Arab Emirates 468 4 12
79 Haiti 464 -1 -7
80 Bolivia 458 -18 -97
81 Trinidad and Tobago 457 4 25
82 Jamaica 456 -4 -15
83 Gabon 438 0 -5
84 Uganda 431 -3 -26
85 FYR Macedonia 430 -10 -60
86 Congo DR 411 5 19
87 Belarus 405 -7 -53
88 El Salvador 404 5 18
89 Congo 394 3 6
90 China PR 382 9 20
91 Angola 380 -2 -20
92 Benin 378 -2 -20
93 Ethiopia 376 0 -10
94 Oman 370 2 -12
95 Azerbaijan 361 12 46
96 Northern Ireland 358 -10 -73
97 Estonia 351 -9 -54
98 Georgia 350 -1 -14
99 Iraq 334 6 9
100 Zimbabwe 328 1 -5
101 Botswana 324 3 -2
102 Qatar 320 6 11
103 Liberia 312 11 17
104 Central African Republic 310 -1 -21
104 Korea DPR 310 9 13
106 Saudi Arabia 308 3 1
107 Niger 306 -5 -26
108 Kuwait 301 1 -6
109 Latvia 299 6 11
110 Canada 296 -4 -28
111 Antigua and Barbuda 294 0 -9
111 Guatemala 294 -11 -41
113 Guyana 286 17 52
114 Tajikistan 283 3 2
115 Mozambique 282 1 -1
116 Kenya 274 2 3
117 Equatorial Guinea 273 -19 -90
118 St. Vincent and the Grenadines 271 4 8
119 Bahrain 270 2 6
119 Lithuania 270 -7 -29
121 Lebanon 268 -1 3
122 Burundi 267 2 11
123 Malawi 263 -1 0
124 Turkmenistan 254 1 0
125 New Caledonia 249 -30 -134
126 Luxembourg 247 0 0
127 Namibia 246 0 7
128 Rwanda 242 3 9
128 Tanzania 242 -1 3
130 Suriname 237 5 14
131 Grenada 233 -12 -35
132 Moldova 224 -3 -13
133 Afghanistan 223 -1 -4
133 Philippines 223 8 23
135 Cyprus 219 -1 -6
136 Sudan 218 4 14
137 Kazakhstan 216 -5 -11
138 Syria 213 5 21
139 St. Lucia 203 -1 -7
140 Gambia 202 -4 -20
141 Malta 192 1 -7
142 Kyrgyzstan 186 2 0
143 Lesotho 183 4 7
144 Thailand 181 -5 -25
145 Tahiti 179 1 -3
146 Belize 178 -1 -6
147 Palestine 175 2 3
148 St. Kitts and Nevis 172 -11 -40
149 Mauritania 168 1 0
150 Hong Kong 164 -2 -10
151 Myanmar 163 11 39
152 Nicaragua 155 1 11
153 India 151 2 8
154 Chad 148 4 10
155 Maldives 147 -3 1
156 Liechtenstein 141 0 -1
157 Puerto Rico 139 3 10
158 Vietnam 135 -5 -9
159 Bermuda 127 -2 -12
160 Singapore 124 -1 -7
161 Bangladesh 120 5 7
161 Malaysia 120 0 -8
161 Sao Tome e Principe 120 2 0
164 Nepal 119 -1 -1
165 Sri Lanka 108 3 0
166 Laos 105 6 21
167 Pakistan 102 3 0
168 Indonesia 96 2 -6
169 Dominica 89 -2 -23
170 Curacao 88 4 16
171 Guam 86 4 16
171 Solomon Islands 86 -2 -19
173 Aruba 82 -8 -32
173 Barbados 82 -22 -75
175 Faroe Islands 81 7 29
176 Yemen 72 -3 -8
177 Chinese Taipei 67 -2 -3
178 Mauritius 62 -1 0
178 Samoa 62 -1 0
180 Madagascar 57 -1 0
181 Guinea-Bissau 56 -1 0
182 Vanuatu 53 -1 0
183 Mongolia 49 2 0
183 Swaziland 49 3 1
185 Fiji 47 2 0
186 American Samoa 43 2 0
186 Tonga 43 2 0
188 Bahamas 40 3 0
189 Montserrat 33 4 0
190 Comoros 32 3 -1
191 US Virgin Islands 30 -1 -12
192 Cayman Islands 29 0 -7
193 Brunei Darussalam 26 -11 -26
193 Timor-Leste 26 -11 -26
195 Eritrea 24 0 0
196 Seychelles 23 0 0
197 Papua New Guinea 21 0 0
198 Cambodia 20 1 0
199 British Virgin Islands 18 -2 -3
200 Andorra 16 0 0
201 Somalia 14 0 0
202 Cook Islands 11 1 0
202 Djibouti 11 1 0
204 Macau 10 -2 -3
204 South Sudan 10 1 0
206 Anguilla 3 0 1
207 Bhutan 0 0 0
207 San Marino 0 0 0
207 Turks and Caicos Islands 0 0 0
 

About me:

Christian, husband, father x 3, programmer, Romanian. Started the blog in March 2007. Quit in April 2018. You can find me on LinkedIn.

169 comments:

  1. Hi bro the possible ranking points for INDIA are being shown as 161 on FIFA.COM

    But you are showing INDIA will only have 151 points?

    Why is the differecnce?

    ReplyDelete
    Replies
    1. Sagi, you'll also notice that FIFA has Afghanistan with 227 points instead of 223. I suspect they forgot to include in the calculations the final of the SAFF Cup. That's the only explanation I could find.

      Delete
    2. Yes, according to my calculations - I've checked the calculations for both Afghanistan and India.

      Delete
  2. Edgar, I think there's a typo on your table. It says 'September 2013 probable ranking', might confuse someone.

    ReplyDelete
  3. OK Belgium needs a draw in Croatia and win the one vs Wales ... then they only lose 5pts. In this case they'll have 1154pts ... I'd feel that should suffice to earn a seed ... what's the probability of Belgium being seeded when Croatia = a draw + they win vs Wales?

    ReplyDelete
    Replies
    1. 100% - assuming FIFA use the October ranking.

      Delete
    2. Thanks Ed ... can you let the Belgian FA know that they need a draw in Zagreb and a win in Brussels ... their "destiny" is still in their own hands.

      Maybe remind them that during USA1994 a draw vs S.Arabia in the last group match would have meant a second round match against Ireland instead of Germany. Although that Germany - Belgium match was very exciting and the double foul on Weber has given us another piece of trivia to moan about (like Wilmots's disallowed goal vs Brazil at WC2002, referee error in the decisive qualifier vs Holland WC1974, etc.).

      Delete
  4. Hello Edgar

    I think there is a mistake of the points of Libya

    you put Libya 543

    but Libya will be 527

    the fifi .com put it 527

    and i collection it 527

    Hussein Budejaja`(Libya)

    ReplyDelete
    Replies
    1. Hi Hussein,

      Central African Republic friendly (14 August) lost A match status.

      Delete
  5. yes i see it now thanks edgar

    there were more then 6 sub from libya team

    but i collection it now

    Libya 541

    ReplyDelete
    Replies
    1. You're welcome, Hussein.

      It should be 543 though, but close enough.

      Delete
  6. Edgar

    2 Matches lost ( A) match status

    Central African Republic friendly (14 August)

    and

    Sudan (26 August 2012 )!!

    so it is

    Libya 543 :)

    if Sudan match in it will be Libya 541

    Hussein Budejaja

    Best regards

    ReplyDelete
  7. First of all, thanks for a fantastic site. It was refered recently in Swedish media, as our team has a chance of being seeded in the play off. My questions are:
    1. If the probable results for October rankin remains the same but with the only change that
    Sweden - Germany will be German victory instead of draw, would Sweden still remain before both Ukraine and France?
    2. With the same results but with the change that, Sweden draws against both Austria and Germany (ie Swe-Aus a draw instead of a victory for Swe), would Sweden still be before both Ukraine and France?

    ReplyDelete
    Replies
    1. Hello O. 1)If Sweden wins against Austria and loses to Germany, they will amount 850 points for the October ranking. So they won't be seeded if France AND Ukraine obtain their expected results.

      For 2) Sweden only gets 828 points. So, it's even worse. The ranking points are proportional to real match points. So, in most cases (not always), 1 qualifier victory counts more than 2 qualifier draws (3 points per win is more than 2 points per two draws).

      At this point you can play with FIFA's ranking prognosis tool to figure out the possible scenarios yourself. Available at:

      http://www.fifa.com/worldranking/rankingtable/index.html

      Delete
  8. Frontrunners for a seed:

    - Brazil, Spain & Argentina = can afford to lose both matches
    - Germany & Italy = need 1 more draw
    - Colombia & Uruguay = need 1 more win
    - Belgium = needs 1 more draw + win
    - Croatia, Holland & Switzerland = need to win both matches*

    *If Croatia wins both matches then they'll have a seed instead of Belgium (but when Croatia doesn't qualify in the play-offs that seed will go to another frontrunner). Holland & Switzerland can only be seeded when other teams don't get the necessary results for them to be seeded. If Holland & Switzerland win both their matches Switzerland will be ranked higher than Holland.

    Did I make a mistake? Would you add/change something?

    ReplyDelete
    Replies
    1. If Colombia loses the first one and wins the second (1136 pts) they can be overtaken by Switzerland (1138 pts with two wins). So they especially need not to lose the first one (which is btw against Chile). And then 1 win.

      If Spain loses both matches they could be overtaken by France and be condemned to the play-offs. So they can't afford to lose both matches.

      Your other statements are correct. The nearest other candidate would be Chile with two wins (1126 pts).

      Delete
    2. Oops, and Italy and Germany both needs one more win also instead of just one draw. With a loss and a draw Italy could end at a minimum 1110 pts and Germany at a minimum 1125 points and then they could be overtaken by Croatia (1149), Switzerland (1138) and the Netherlands (1136), all with two wins.

      Delete
    3. If France loses against Finland is it possible that they finish as worst 2nd and miss the play-offs?

      Delete
    4. @Ed Thanks!

      @Daniel Burnier - In 10000 sims, France were never worst 2nd. I didn't look at all the possible scenarios, but I think they have secured at least a play-off spot.

      Delete
    5. Thanks Ed. I had tried to ignore less likely scenario's ... so there's a fair bit of "gestimation".

      I had boiled it down to:

      - Brazil, Spain, Argentina, Germany, Colombia & Italy as the most likely candidates to have a seed.
      - a toss up between Croatia & Belgium for one seed (the odds compilers at the bookmakers predict a very tight match ... Belgium only needs a draw vs. Croatia + win vs Wales).
      - another toss up for the other seed between Uruguay, Holland & Switzerland (again looking at bookmakers: Uruguay would spill too many points away vs Ecuador & home vs Argentina while Switzerland will win both games and finish ahead of Holland ... Holland facing the resurgent Turks in a hell-ish atmosphere also favours the Swiss).

      Delete
    6. @Edgar: thanks! I just checked it - Considering that Malta finishes 6th in the group B (what will surely happen), there is no way the 2nd from this group can finish in front of France

      Delete
  9. the list of countries, ranked by FIFA ranking which can still qualify

    Q = qualified
    q = may qualify
    blank = eliminated

    1 Spain q UEFA: at least 2nd
    2 Argentina Q
    3 Germany q UEFA: at least playoff
    4 Italy Q
    5 Colombia q CONMEBOL: at least playoff
    6 Belgium q UEFA: at least playoff
    7 Uruguay q CONMEBOL
    8 Brazil Q
    9 Netherlands Q
    10 Croatia q UEFA: at least 2nd
    11 Portugal q UEFA: can be 1st and 2nd
    12 Greece q UEFA: at least 2nd
    13 USA Q
    14 Switzerland q UEFA: can be 1st and 2nd
    15 Russia q UEFA: can be 1st and 2nd
    16 Chile q CONMEBOL: at least playoff
    17 England q UEFA: can be 1st and 2nd
    18 Bosnia-Herz.q UEFA: at least 2nd
    19 Côte d'Iv. q CAF Playoff
    20 Ecuador q CONMEBOL
    21 Mexico q CONCACAF hex
    22 Sweden q UEFA: can be 1st and 2nd
    23 Denmark q UEFA: can still reach playoffs
    24 Ghana q CAF Playoff
    25 France q UEFA: at least 2nd
    26 Ukraine q UEFA: can be 1st and 2nd
    27 Montenegro q UEFA: can be 1st and 2nd
    28 Algeria q CAF Playoff
    29 Slovenia q UEFA: can be 1st and 2nd
    30 Hungary q UEFA: can still reach playoffs
    31 Romania q UEFA: can still reach playoffs
    32 Czech Rep. q UEFA: can still reach playoffs
    33 Costa Rica Q
    34 Peru
    35 Panama q CONCACAF hex
    36 Venezuela q CONMEBOL
    36 Nigeria q CAF Playoff
    38 Mali
    39 Norway q UEFA: can still reach playoffs
    40 Honduras q CONCACAF hex
    41 Paraguay
    42 Japan Q
    43 Serbia
    44 CV Islands
    45 Albania q UEFA: can still reach playoffs
    46 Tunisia q CAF Playoff
    47 Austria q UEFA: can still reach playoffs
    48 Iran Q
    49 Turkey q UEFA: can still reach playoffs
    50 Egypt q CAF Playoff
    51 Burk.Faso q CAF Playoff
    52 Wales
    53 Australia Q
    54 Iceland q UEFA: can be 1st and 2nd
    55 Armenia q UEFA: can still reach playoffs
    56 Finland
    57 Uzbekistan
    58 Korea Rep. Q
    59 Rep.Ireland q UEFA: can still reach playoffs
    60 Slovakia
    61 Cameroon q CAF Playoff
    62 Bolivia
    63 Scotland
    64 Bulgaria q UEFA: can still reach playoffs
    65 Poland q UEFA: can be 1st and 2nd
    66 Senegal q CAF Playoff
    67 New Zealand q Intercontinental Play-Off vs. CONCACAF
    68 S. Africa
    69 Israel q UEFA: can be 1st and 2nd
    70 Libya
    71 Zambia
    72 Sierra Leone
    73 Jordan q Intercontinental Play-Off vs. CONMEBOL
    74 Morocco
    75 F Macedonia
    76 Guinea
    77 Togo
    78 Haiti
    78 Jamaica
    80 Belarus
    81 Uganda
    82 UA Emirates
    83 Gabon
    84 Cuba
    85 Trin. & Tob.
    86 N. Ireland
    87 Dom.Republic
    88 Estonia
    89 Angola
    90 Benin
    91 Congo DR
    92 Congo
    93 Ethiopia q CAF Playoff
    93 El Salvador
    95 New Caledonia
    96 Oman
    97 Georgia
    98 Equatorial Guinea
    99 China PR
    100 Guatemala

    ReplyDelete
  10. Hi Edgar, I've been a long long time without posting, around 2 years.

    Four comments, considering that there will be 4 CONCACAF and 6 CONMEBOL teams at the Cup:

    1) If finally a team like Switzerland is in top 7 thus seeded, I think they will change the format, what do you think Edgar?

    2) Being european and seeded (I'm spaniard) may be really hard: you're surely gonna play against another non seeded european team (Netherlands, England, Portugal... are the worst chances), and you have probably a 50% of playing against a great southamerican team (Chile or Ecuador, probably). One of these (Spain-Germany-Italy & other european & Chile/Ecuador & other good team from the rest of confederations) could be a "Grupo de la muerte" as they are usually so called.

    3) There may be a group with a CONMEBOL seeded team and 2 non-seeded europeans, as surely there will be no group with 3 european teams, and being 4 seeded teams, there are 9 left. They'll probably pick the 13th european with lowest ranking (as you said before, like Serbia in 2006) to avoid (mainly Brazil) to have 3 giants (i.e.: Brazil/Argentina, Netherlands and England) in the same group. But It would be great if France is the european team with lowest FIFA ranking...

    4) Which would be the logical distribution of pots? Here my bet:

    Pot 1: Seeds: 8 seeded teams (4 CONMEBOL and 4 UEFA)
    Pot 2: European: 9 UEFA (1 in each group + lowest ranking to be in a CONMEBOL-seed group)
    Pot 3: 2 CONMEBOL + 5 CAF; the grup with 1 CONMEBOL seed & 2 UEFA without african team.
    Pot 4: 4 AFC + 4 CONCACAF

    Again, considering that Jordan and New Zealand will be out; I can be wrong.

    Hope your answer, great to follow this web (all this years, even I didn't post)

    Juan GG

    ReplyDelete
    Replies
    1. Hi Edgar, could you tell us your opinion about these questions:

      - If finally a team like Switzerland is in top 7 thus seeded, I think they will change the selection of seeds, what do you think Edgar?

      - Which would be the logical distribution of pots? Here my bet:

      Pot 1: Seeds: 8 seeded teams (4 CONMEBOL and 4 UEFA)
      Pot 2: European: 9 UEFA (1 in each group + lowest ranking to be in a CONMEBOL-seed group)
      Pot 3: 2 CONMEBOL + 5 CAF; the grup with 1 CONMEBOL seed & 2 UEFA without african team.
      Pot 4: 4 AFC + 4 CONCACAF

      Thank you very much, congrats again!

      Juan GG

      Delete
    2. Well, finally it seems that Uruguay will not be seeded, thus Brazil, Spain, Argentina, Colombia, Germany, Italy, Switzerland and... Netherlands or Belgium? Anyway, 5 europeans and 3 americans, thus:

      Pot 1: Seeds: 8 seeded teams (3 CONMEBOL and 5 UEFA)
      Pot 2: European: 8 UEFA
      Pot 3: 3 CONMEBOL + 5 CAF.
      Pot 4: 4 AFC + 4 CONCACAF

      Supossing that México and Uruguat will beat NZ and Jordan.

      Really weird to see Switzerland as seeded. We could have these two groups at the WC:

      Germany-England-Uruguay-USA
      Switzerland-Bosnia-Burkina/Algeria-Iran

      Juan GG

      Delete
    3. Hi Juan,
      What makes you think Uruguay won't be seeded? A home victory against Argentina will be enough for them...
      Also, Belgium is already a seed.

      Confirmed seeds are: Brazil, Argentina, Germany and Belgium.

      Spain will be seeded when they qualify.

      Italy, Uruguay and Colombia will be seeded if they win.

      Switwerland needs to win, and expect one of those three not to win.

      Netherlands needs to win, and hope that 2 out of (Uruguay, Colombia, Italy, Szitzerland) don't win.

      Delete
    4. OK then I was wrong, I thought that it was difficult for Uruguay to be seeded (before the matches already played) and after their defeat I thought they had no chance even if qualifying.

      So Belgium is sure going to be seeded? Ok thanks.

      So we can still have 4 CONMEBOL teams seeded, and a 8+1 european pot.

      According to your calculations, it seems very difficult for Netherlands to be seeded (2 teams failing to win and they play in Turkey, hard game).

      Thank you Juan.

      Juan GG

      Delete
    5. Hi Juan. Yes Belgium is confirmed, I read somwhere in the blog that just by a few decimals they will be seeded even if they lose.

      I think 4 CONMEBOL teams seeded looks like the most likely scenario, but it's football, we never know. That doesn't make much of a difference for European seeds, but it makes a huge difference for Argentina and Brazil: there could be groups like:

      Brazil, Netherlands, France, USA or
      Argentina, England, France, USA

      European seeded teams are better off having Uruguay as a seed, I think they would prefer not to face Uruguay.

      Yes, Netherlands chances are tough, buuuut I guess we'll find out tonight. There are actually other teams with theoretical chances of being seeded, but their chances are more anecdotical than real. You can see this post to have an idea of the complete scenario:

      http://www.football-rankings.info/2013/09/2014-fifa-world-cup-final-seeding_13.html?showComment=1381681266479#c5034114708810251358

      Delete
    6. Well, I would like to correct myself. Actually, if Netherlands is seeded that would make a difference for all the seeded teams, since nobody wants to face them. In that case, the death case scenario could present itself for a European team:

      Germany, England, Uruguay, USA

      For instance.

      Delete
    7. Why would they change it? Quite surprised by FIFA announcing the seeding system in advance. Good for them. And I agree with your pot setup. With Italy, England and most likely Netherlands not seeded we could see many groups of death.

      Delete
  11. Beating Luxembourg is worth more points than drawing Argentina.

    Is this correct? As a draw against Spain, Germany, Brazil or Argentina seems a lot harder than winning vs LUX, this doesn't feel right.

    ReplyDelete
    Replies
    1. It's correct according to the rules, but I agree, it doesn't feel right.

      Delete
    2. It's the 3pt thing. It should be 2pts for a victory, not 3. There's enough incentive in the competitions to win, 3pts if it's a group match, and if you don't win, you go home if it's KO.

      Delete
    3. How does Elo fare? Do you get the same result, e.g. beating Luxembourg is worth more points than drawing Argentina. Would you say that Elo also needs some tinkering to reflect reality better. In chess the result between a master like Argentina and an "amateur" like LUX wouldn't even affect your ranking. UEFA often pits minnows against masters ... so those results should matter but have to be calibrated.

      For FIFA a decent T multiplier in conjunction with improvements for the other multipliers should result in a ranking that is more accurate ... they just feel too arbitrary.

      Delete
    4. Elo works with the exchange of points between the two contestants in a match. The number of points exchanged depends on:
      - the result of the match;
      - the expected result of the match (win expectancy) based on the difference between the elo-ratings of the teams concerned. Furthermore a form of home advantage is included in the calculation of the win expectancy;
      - the importance of the match;
      - the goal difference in the match.

      In general when the rating differs more than 400 points the number of points exchanged is very low. The more the result differs from the expected result the more points are exchanged. And then it helps if the rating difference is small. For detailed info see: http://www.eloratings.net/system.html

      In general beating Luxumbourg gives a top team 0 points extra (regardless if it's a friendly or qualifier) and drawing Argentina is a lot more rewarding. For instance if Peru draws Argentina in the coming WC qualifier they gain 16 points (and Argentina loses 16 points).

      A huge upset like Denmark-Armenia 0-4 (WCQ June 11th this year) delivered Armenia 69 points and cost Denmark the same amount.

      Delete
    5. OK it seems there's something like a ratings floor (beating LUX = no gain for a top team). This also implies that "masters" shouldn't face "amateurs" that often. A minnows trny before ending up in the UEFA qualifiers could be the way to go. At least it helps the minnows develop and they'll get more support when they finally win something (LUX's legacy would render the most optimistic guy to tears ... might explain why they don't care about footie).

      That goal difference can't take into account the quality of goals I suspect, e.g. in a do or die scenario who cares about another counter when you'll be eliminated if you don't score a goal of your own. For most matches I guess that adjusted K value makes sense. Haven't looked at how you can game the ranking by playing mainly against opponents that don't care too much about defense.

      PS somehow I'd hoped that Elo would be more responsive, e.g. a NT that's stringing together wins against tough opponents/running red-hot should be expected to win. Let's call it momentum which assigns more weight to more recent games and accounts for strings of excellent results.

      Delete
    6. Well, it is just a simple formula...
      So no, quality of goals is not taken into account. In general when one team scores a lot against another team, the rating difference will be rather big anyway and the number of points exchanged will thus be small (f.i. Netherlands - San Marino 11-0, no points won by the Netherlands).
      If two teams are of equal strength and still one scores four more than the other then the points are multiplied by 1.875. Then the goal difference has a (fair) impact.

      Elo takes just the current elo-ratings into account and doesn't look at previous matches. Maybe because a previous match could be years ago. That's also the main disadvantage of the system: it is a bit slow in response because all past results keep their value.
      On the other hand (every disadvantage has an implied advantage), it doesn't react heavily on every surprise result and that makes it a stable ranking where you can see the same tendencies as in the FIFA ranking but not so overreacted. So Colombia and Belgium are climbing in the elo ranking too, but they are still both nowhere near the top 5 like in the FIFA ranking. At the moment Colombia is 10th and Belgium is 16th in elo.

      Delete
    7. Ed, have you ever tried to calculate Elo "again" starting f.e. post-Bosman? would it change the ranking a lot or not?

      Maybe you could also calculate a fifa ranking based on 8 years like it used to be (but with current points system)?

      Have you ever tried alternatives?

      Delete
    8. Jeroen,

      Start the Elo calculation again at the Bosman-arrest wouldn't change anything, because the elo-system has the special feature that the rating of a team becomes more or less stable after some 30 matches. So, if you for instance should exclude one match of Belgium played 5 years ago and then take a look at the time needed for the rating of Belgium to become the same again (as if the match wasn't excluded) you will see that that point is reached within 30 matches later. You should do that calculation yourself once, it is a fantastic phenomenon to discover.

      When a team is newly introduced in the elo ranking it is given a certain starting rate (just a common guess of the strength of the team compared to other teams). After 30 matches it will have reached it's 'stable' strength.

      I am quite satisfied with the elo ranking, because it is stable and easy to calculate and can easily be used as a prediction mechanism. Most of the time the ranking of teams is according to what I expect, it gives logical results beyond the 'issues of the day'.
      I know the FIFA ranking is far from perfect, but I haven't tried other ranking methods. Why should I as none of them would be an official one ?

      Delete
    9. A little experiment might (dis)prove elo's reliability as a prediction tool. E.g. use an elo ranking from the past and instead of updating it with the real life results, only use elo predictions. Actually only next game predictions matter, as elo is constantly updated. So to compare its accuracy to other tools, you'd look at the margin of error in these predictions.

      I'd think it would be too stable to reflect reality ... it's only because elo gets a reality check after every game, that it has some fluidity.

      Delete
    10. It is now since February that I use elo predictions of match results to predict the October 2013 ranking. So far the overall prediction success rate is 56%, with friendlies around 52% and non-friendlies at around 60%.

      Compare that to the toto for the Dutch Eredivisie which I organise for a group of friends a couple of years now: the prediction success rate there lies for the number one after 34 matches at around 53% each year. So no bad result for elo.

      Which 'other tools' do you know anyway to use as a prediction tool for match results?
      I have studied the SPI a bit, but that is a far from easy calculation with necessary knowledge of rosters of the teams involved and can only be used for short term predictions when rosters are known.
      I think that the FIFA ranking as a prediction tool will perform significantly less than elo, even if I don't know exactly how to use that ranking for the prediction of a match result.
      Anyone a suggestion ?

      Delete
    11. SPI uses, amongst others, player-based ratings. Compared to baseball, football stats are a bit harder to interpret correctly and this remains a major weakness (can also be found in other models).

      Other suggestions:

      - Bookmakers use a variety of models to essentially price risk and ensure their highest likelihood of making a profit. This biases their probabilities just a bit, but they can serve as a useful reference point.

      - Bloomberg Sports: Utilizing Opta’s player data, Bloomberg has identified what they feel are the key player offensive and defensive inputs that contribute to the likelihood of various match outcomes (goals scored, conceded, win, draw, loss, etc). They then roll the likelihood of each match outcome up into forecasted season results based upon a 10,000 run Monte Carlo simulation of the season. As far as NTs are concerned, you'll have to wait untill they include them or do the work yourself.

      - Transfer Price Index’s mSq£R model (TPI): This model uses venue and squad cost in terms of inflation-adjusted transfer fees to estimate the likelihood of match outcomes, which are then rolled into a likelihood of finishing in each table position based upon a 10,000 run Monte Carlo simulation. Also lacks info on NTs and that's why I opted for bookmakers as a reference point (instead of elo).

      PS this is from MCofA, blogger at SB Nation who feels that he's using a better game projection model than most. He's using: "a sampling from a bivariate poisson distribution to simulate scores, which allows me to break ties in the table using goals scored / goals conceded just like in the real world. I’m also basing it on team strength estimates built from underlying stats like shots on target in the box which seem to correlate well to goal scoring but with less variance." Maybe you can convince him to apply his (PL) model to NTs.

      Delete
    12. Regarding the accuracy of Elo and criticism of it here, it could be the case that Elo - and/or equivalents - is TOO accurate in the sense that it cannot predict upsets. If I go to a bookmaker at any given weekend and put money on Celtic, Man Utd, Chelsea, Man City, Arsenal, Liverpool, Bayern Munich, Barcelona and Real Madrid to win, invariably there is always one of these teams who don't win even though each one would be expected to win prior to the match. Sometimes there are factors such as form and the pressure of going for a title or CL qualification that can give an indicator that an upset may be on the cards but most upsets are unpredictable - that's why they are upsets!! So for an algorithm like Elo, it's accuracy is the very thing that prevents it from being 100% accurate as it would, for example, never predict Armenia to draw 2-2 in Italy, yet this happened. I don't think you would find SPI or an equivalent predicting this outcome. Elo, or any other algorithm cannot account for the vagaries of sport so criticism of a 53% success rate seems unfair and suggesting this could be significantly improved is probably futile.

      Delete
  12. Hey

    Just have to say that your above ranking is at odds with the seeding probabilities you posted a week earlier.

    ReplyDelete
    Replies
    1. Seeding probabilities are based on 10.000 simulations of results for each remaining match. The ranking above is based on probable results, so one (probable according to elo) result for each remaining match.
      You can conclude that the probable results lead to another set of seeds than the simulations, but hey, both methods are just predictions of the future, maybe (or probably !) both are wrong anyway.

      Delete
  13. It clearly says probable, not final and official. I Think this was calculated provided that all of these teams haven't played any games at all during September - early October

    ReplyDelete
  14. To predict these results, have you ever considered using another tool?
    I've noticed a few striking differences between the probable results (computed by using Elo) and what odds compilers, risk analysts, etc. come up with (odds at bookmakers). As far as seeds are concerned the main difference would be the result of Croatia-Belgium. This match is too close to call according to bookmakers and probably it's the home-advantage that's decisive in the Elo predictions. I've put them down as a toss up (seeing that Belgium can be seeded after a draw vs Croatia they wouldn't even be equally likely to get a seed = Belgium has better odds than Croatia who must win both remaining matches).

    Can Elo predictions be "tuned", e.g. calibrated to include success in away games (Belgium hasn't lost on the road nor even conceded a goal in an away game during these qualifiers) + reflect "momentum" more accurately + ...?

    PS a while ago I had read a more or less comprehensive article about who will end up on top in the PL (on Forbesdotcom). If I remember correctly they used 4 different approaches and an analysis of the bookmakers was one of them (more than simply comparing the odds). I felt that this approach had the most merit. I could be mistaken but there seems to be more data, better info and eventually a more sound knowledge when you analyse bookmakers.

    ReplyDelete
  15. can you update the probable ranking? especially to the philippines that is having Philippine peace cup on october 11-15.,

    ReplyDelete
  16. "Regarding the procedure for the Final Draw for the 2014 FIFA World Cup Brazil™, the seeded teams (Pot 1) will include Brazil and the seven top-ranked teams. The remaining pots will be based on geographic and sports criteria. The FIFA/Coca-Cola World Ranking of October 2013 (to be released on 17 Oct. 2013) will be used as opposed to the November ranking as the latter would give the teams involved in the upcoming play-off games an unfair advantage."

    Source: http://www.fifa.com/aboutfifa/organisation/bodies/news/newsid=2190697/index.html

    I guess we can confirm that the october 2013 ranking will give us the seeded teams.

    Mat

    ReplyDelete
  17. The Belgians are very busy with the whole seeding thing. Here's a newspaper article in Dutch stating that FIFA yesterday has decided (sooner than expected) that the October ranking will be the only criterium for seeding the final draw. On FIFA.com no news about that decision, so I have my doubts.

    http://www.nieuwsblad.be/sportwereld/cnt/dmf20131004_061

    ReplyDelete
  18. Most likely you guys already have heard the news:

    Regarding the procedure for the Final Draw for the 2014 FIFA World Cup Brazil™, the seeded teams (Pot 1) will include Brazil and the seven top-ranked teams. The remaining pots will be based on geographic and sports criteria. The FIFA/Coca-Cola World Ranking of October 2013 (to be released on 17 Oct. 2013) will be used as opposed to the November ranking as the latter would give the teams involved in the upcoming play-off games an unfair advantage.

    ReplyDelete
  19. Egypt will win Ghana in Kumasi they have a powerful team with a record of 100% win in the qualifiers.

    ReplyDelete
  20. This comment has been removed by the author.

    ReplyDelete
  21. The news about the decision of seeding procedure are here:
    http://www.fifa.com/aboutfifa/organisation/bodies/news/newsid=2190697/index.html

    ReplyDelete
  22. Very interesting: France against FIFA Ranking as the criteria to set the seeds for the play off (in spanish):

    http://futbol.as.com/futbol/2013/10/13/internacional/1381664742_756841.html

    I think that they are right, they have played less official matches because being in a 5 teams group. I think that points achieved by other play off teams agains 6th placed teams should not be considered; could you make these calculations Edgar?

    ReplyDelete
    Replies
    1. Juan, France should be very careful what they wish for.

      For instance: if the matches of Sweden against Faroer are excluded their points total grows from 889 to 897.
      Same goes for Ukraine: if the one match played so far against San Marino is excluded their points grow from 894 to 922.

      Normally matches/wins against the bottom-placed teams deliver less points than the last time-frame average, so it is advantageous to exclude them.

      Delete
    2. OK thank you very much Ed, anyway I think that they should consider the effect of the 6th team matches on rankings (being positive or negative for France in this case) so all teams have the seam chances of being seeded.

      Juan GG

      Delete
  23. Ed. Very interresting. In newspapers in Sweden this story is told today aswell. France think the ranking is unfair and so on.... It is almost funny that they don´t realize that they have an advantage to not play a 6:th seeded team in the qualifications. But what i don´t understand is: Why did France play the friendly against Australia? without that they would go ahead of Ukraine in the rankings, right?

    ReplyDelete
    Replies
    1. Hi Henrik. This is not correct. Without the Australia match they would have had one point less. They needed to play and win against Mali, Cape Verde Islands or Iran instead of Australia and they would have been seeded.

      Delete
  24. Henrik, not quite.
    The last year time frame average of France was 329 before the friendly agianst Australia. The win delivered them 410 match points so their average climbs. A win against Finland tomorrow ends them at 870 points and without the Australia friendly that would be 869 points. So it did them no harm.

    ReplyDelete
  25. Igraj Bosno ljubavi moja! Bosnia at the world cup ladies and gentleman! Whoever is not happy for this tortured andvraped country in our moment of sporting joy, has no heart!

    ReplyDelete
    Replies
    1. vedad, congratulations. Well deserved and see you in Brazil.

      Delete
    2. Congrats to Bosnia from Portugal! Four years ago i said (at Fifa site) that their time would come sooner or later...They deserve it!

      P.S. Unfortunately we aren´t going to play Bosnia in the play-offs :)


      Delete
  26. Italy are in 2-2 tie in the 80th minute playing Armenia in home and if it will end as a draw they will have 1136 points - the same account as the Netherlands already have after beating Turkey 2-0.

    So who will be seeded in this case? What about the decimates of points for both teams?

    ReplyDelete
  27. Guess what, you could have put your house on it! 3-2 Italy... NO, still 2-2. How many heart-attacks can I survive today? As far as I read, Netherlands would be ahead of Italy on decimal points.

    ReplyDelete
  28. This comment has been removed by the author.

    ReplyDelete
  29. The Netherlands has the advantage: 1135.95 over 1135.61

    ReplyDelete
  30. Italy just blew it ! So still one to go for the Netherlands, a draw for Uruguay or Colombia is enough to get seeded :D

    ReplyDelete
  31. Thankfully my house is safe as I didn't bet it, wow, good job Armenia, respect! Now I have to stay up another 6 hours or so with all my fingers crossed for Paraguay and Argentina, hoping one will take points of their opponents so that Oranje will be seeded. It would be a real shame if we're not seeded having not lost for a year with the record for the most World Cup goals ever in qualifying! But using the FIFA ranking is the correct decision, so I'll have to respect that. Switzerland and Uruguay have been racking up the points in the same period also. I remember Canada giving up their place at the CONCACAF finals to Jamaica, hopefully Colombia can also show such sporting behavior and give up their seeding to us tonight! ;-)

    ReplyDelete
    Replies
    1. hopefully not... really looking forward for glorious victories of Uruguay and Colombia later :-)

      Delete
    2. btw. Austria finished as best 3rd.... I know it means nothing pratically, but they have shown a lot of improvement in my opinion

      Delete
  32. This comment has been removed by the author.

    ReplyDelete
  33. UEFA play-offs:
    pot 1: Portugal, Greece, Croatia, Ukraine
    pot 2: France, Sweden, Romania, Iceland

    Way to go, Iceland !

    ReplyDelete
  34. Ed, Homer.... are you staying up late tonight to watch the CONMEBOL games? One tie and you are there!

    I can't believe Italy blew it like that, now they have it harder than Netherlands! Just to warn you, we will field 9 subs against Uruguay. I'd put more faith on Paraguay...

    Juan(Arg)

    ReplyDelete
    Replies
    1. Juan, I didn't stay up late. I saw this morning that both Colombia and Uruguay had won, so congrats to the SA-teams!
      I think your fear that the European pot will be loaded with some heavyweights has come true with Italy, England and yes Oranje - who, I am glad to say, showed indeed some promising good form the last two matches. A few possible 'groups of death' at the WC lie in front of us, can't wait :)

      Delete
    2. Yes, Ed. That's my fear, don't forget about France as well, and why not Portugal? We know that a south american seed will have higher chances of being in the group of death.

      Argentina seems to have very "good luck" when it comes to the group of death:
      2002-> Arg, Eng, Swe, Nig
      2006-> Arg, Ned, SCG, Civ

      So, with this record, I have a bad feeling. In 2002 we were ready to be champions, and the group of death stopped that: away by a penalty (Eng) and a free kick (Swe)!

      Juan (Arg)

      Delete
  35. Well, I don't think there's any channel that I can watch it on. There's probably websites, but I don't know if my heart will survive that. I'll go to sleep shortly and then set my alarm to wake up near the end of the game I think. Yes, that Italy blew it was a big surprise, but Armenia have had some decent results lately. I thought Switzerland would blow it, and having just seen their game on Eurosport they really got away with it. Slovenia had some good chances. Sure, Argentina will put some subs in, but they'll be fighting for recognition and a chance to get into the squad for next summer, so there's a chance, especially if Uruguay gets frustrated or nervous if they don't get an early goal. It only needs one draw, and with Colombia partying just 4 days ago I think we have a decent chance there. But the most satisfying this is that Oranje have shown some good form in the last two matches, and even on German TV they were asking themselves why we're not mentioned as one of the favorites now, saying we only let in 5 goals in qualifying, and scored 34!

    ReplyDelete
  36. Great! Colombia is seeded and Uruguay just needs to go through the play-offs to be seeded. Probably we will have 4 SA seeded countries! Excellent!!!

    ReplyDelete
    Replies
    1. Yes, well Daniel, If Brazil gets this group:

      Brazil, Ita, Eng, Usa

      Maybe you don't think it's THAT great! :)

      Delete
    2. Juan... I would love such a group!!! I hate when we have weak opponents in the group stage. But I am in fact happy about the whole continent, for having 4 seeded teams, especially because the WC will be in our continent.

      Delete
    3. HI Daniel,

      Yes, I'm happy too. South American WC having 4 South American seeds. This would favor a possible semifinal with 4 conmebol teams. I'm not saying it will happen, but it favors that scenario. And that would also be great for us.

      But then, be careful what you wish for. Sometimes a team is not at its full during the group stage and then they become stronger. Consider France in 2006. They barely made it through the group stage (which had South Korea and Togo!!) and then they left behind Spain, Brazil and Portugal. If they had fallen into the group of death I doubt that they would have made it through, they were not playing well at all during group stage.

      Don't get me wrong, congrats for our continent for arriving so strong at the WC. But wishing for group of death? Yes, only if your team is not in it! haha.

      Delete
  37. There's a fair chance that Sweden or France will be in that "special pot" (cfr 2006) now that, most likely, 4 C-BOL NTs are seeded. Will have to see what happens in the UEFA play-offs.

    1 of the C-BOL seeds will get that special pot NT.

    C'mon France beats Ukraine and/or Sweden beats Greece (if those meet in the play-offs).

    ReplyDelete
  38. So the seeds are Brazil, Spain, Argentina, Germany, Belgium, Switzerland, Colombia and Uruguay (if they can get past Jordan). Are we sure that there will be a special pot for the worst ranked UEFA team (like in 2006) or could we have all nine teams in one put (like in 2002)? If the latter is the case we could have a group of Brazil, Netherlands, Italy and USA. If there is special pot France are likely to be in it (if they qualify) though they have already complained about the play-off seeding are likely to put more pressure on FIFA if they are indeed placed in the special pot.

    ReplyDelete
  39. Given Elo predicted a Croatia win, does Scotland's win over Croatia give any chance of pot 3 for Euro 2016 qualification? I remember a list you published which had Scotland at 99% for pot 4, which I'm sorry I can't seem to find again. Pretty much wishful thinking on my part really!

    ReplyDelete
    Replies
    1. I recalculated the list, and there were two changes.

      Scotland up in pot 3 and Montenegro down in pot 4
      Scweden up in pot 1 and Croatia down in pot 2

      Delete
    2. Thanks Marko. Music to my ears! Seems a bit tough on Montenegro though

      Delete
    3. I found Ed's post and scotland 100% pot 4 notwithstanding any surprises UEFA have in store for us re the qualification format

      Delete
    4. Already replied on another post - pot 4 for Scotland.

      Delete
  40. Amusing to read that almost all media in Holland say that Italy will be better placed than the Netherlands because last month the points for Holland were rounded down and for Italy rounded up. So according to Dutch media Italy has the upper hand and could be seeded when Uruguay fails to qualify.

    I believe they are wrong though.

    ReplyDelete
  41. Yes, I read that too, totally crazy how one person in the country blindly writes it, and almost everyone blindly follows. Mind you, this is how the press works. I wouldn't trust anything you don't witness yourself.

    ReplyDelete
  42. Hi
    Does any one know what the Oct 2013 ranking (for top 20 or so teams) would be if all of the last four years we weighted equally? Also what would it look like if all matches over the last four years were treated as a single period (in other words, in the first case when each year is effectively 25% of the total even though one year may have 20 matches and another year may only have 8 matches -- in the second case all matches would not be grouped by year).

    ReplyDelete
    Replies
    1. Dorian,
      This would be the ranking if all timeframes would be weighed equally:

      1 Spain 3187
      2 Germany 2856
      3 Argentina 2581
      4 Netherlands 2549
      5 Uruguay 2411
      6 England 2282
      7 Italy 2226
      8 Portugal 2203
      9 Brazil 2111
      10 Greece 2049
      11 Switzerland 2024
      12 Colombia 2020
      13 Belgium 1964
      14 Croatia 1943
      15 Russia 1941
      16 Chile 1937
      17 Côte d'Ivoire 1836
      18 France 1789
      19 Denmark 1733
      20 Sweden 1728

      And this if all points were to be divided by all matches:
      1 Spain 816
      2 Germany 738
      3 Netherlands 675
      4 Uruguay 614
      5 Argentina 613
      6 England 579
      7 Portugal 564
      8 Italy 555
      9 Brazil 533
      10 Colombia 526
      11 Greece 514
      12 Belgium 503
      13 Switzerland 498
      14 Russia 495
      15 Croatia 492
      16 Chile 483
      17 Côte d'Ivoire 462
      18 USA 452
      19 France 447
      20 Ghana 440

      Delete
    2. thanks!
      and what if all friendlies are discarded?

      Delete
  43. Well, finally it seems thay the distribution of pots, being Uruguay and Mexico at the world cup, could be something like this:

    Pot 1: Seeds (4 CONMEBOL and 4 UEFA)
    Pot 2: 9 UEFA (1 in each group + lowest ranking to be in a CONMEBOL-seed group)
    Pot 3: 2 CONMEBOL + 5 CAF; the grup with 1 CONMEBOL seed & 2 UEFA without african team.
    Pot 4: 4 AFC + 4 CONCACAF

    So regarding UEFA teams, Spain-Germany-Switzerland-Belgium are seeded, and non-seeded teams better than Ukraine (20) at the ranking (Netherlands, Italy, England, Portugal, Greece, Bosnia, Croatia and Russia) could'nt be the european team with lower ranking thus the 9th team which would join a group with just another european team (thus, one of 4 with CONMEBOL seed), if I'm not wrong. So, this 9th team could only be: Ukraine, France, Sweden, Romania and Iceland, it's OK?

    So we could have these groups:

    Death one: BRA-ITA-FRA-USA or GER-ENG-CHI-USA

    Light one: SWI-Bosnia-Iran-Burkina/ALG or URU-Bosnia-Costa Rica-Burkina/ALG

    Anyway, we couldnt have the worst/best possible death group, with a CONMEBOL seed -BRA/ARG-, two europeans including Netherlands/Italy and France as ninth and a good african team (Cote d'Ivoire).

    Juan GG

    ReplyDelete
    Replies
    1. which ranking will be used to define the worst UEFA team? october?

      Delete
    2. Hi Juan,

      How certain it is that the lowest ranking european will be in a CONMEBOL seed group? I mean, are we sure they will do it like this? I read somewhere that the 9 teams could be distributed at random and a Conmebol seed would have any two european teams (and not necesarilly one of them being the lowest ranked).

      If France goes as the 9th european they will be very upset:
      Not seeded for group qualifiers
      Not seeded for playoffs
      9th team at the WC, possible having to face Brazil or Argentina.

      Juan (Arg).

      Delete
    3. Hi Juan (Arg.)

      Well, It could be as you say, that way we could have Brazil/Argentina, Italy and Netherlands (or England) in the same group. But I think they will avoid that, as they did in 2006. They should say it now, I think, before the play offs are played. In fact, I think that it should be known before the whole qualification.

      Juan GG

      Juan GG

      Delete
    4. This comment has been removed by the author.

      Delete
    5. SInce I believe using the ranking to separate the ninth European makes sense, the pots would be probably like this, if Uruguay qualifies:

      Pot 2: NED, ITA, ENG, BIH, RUS, (plus 3 among [POR], [GRE], [HRV], [UKR], [FRA], [SWE], [ROM]);

      Special pot: (one among [UKR], [FRA], [SWE], [ROM], [ISL]);

      Hence, there could be a group with BRA, ITA, FRA, plus an African team.

      The problem is that, If URU qualifies, there is no way to avoid THREE groups with two Europeans and one South American teams; and ONE of them also including an African nation.

      Delete
    6. it's not possible that a group has two european teams, one south american and one african.

      CONMEBOL SEED 1: 2 europeans, 1 of (AFC/CONCACAF/AFC)
      CONMEBOL SEED 2, 3, 4: 1 european, 1 african, 1 of (AFC/CONCACAF/AFC)

      UEFA SEED 1,2,3,4: 1 european, 1 of (conmebol/caf), 1 of (AFC/CONCACAF/AFC)

      Delete
    7. You are right, my mistake! One team in each group must be from AFC/CONCACAF.

      Pot 1: seeded teams - UEFA (4), CONMEBOL (4)
      Pot 2: UEFA (8)
      Pot 3: AFC (4), CONCACAF (4)
      Pot 4: CAF (5), CONMEBOL (2)
      Spec. Pot: UEFA (1)

      Delete
    8. One funny thing is that, if Jordania qualifies, there is no problem with distributing UEFA teams in the pots. However, either NZL goes to "Pot 4" or HON goes to a Special Pot. This would allow such crazy group as BEL(SUI), BHI, IRN, NZL(HON).

      Delete
  44. I posted this over on BigSoccer. Decided to repost here :-)


    An Open Letter to FIFA re: WC2014 Seedings

    Dear FIFA:

    Many of us liked your recent announcement to use FIFA Rankings as of October 2013 to determine the seven nations to complement the host nation as group seeds for the WC2014 eight groups.

    However, a quick observation shows a number of traditionally (as well as recently) strong teams were just outside of this group (assuming Uruguay qualifies), including Netherlands, Italy, England, and Chile. We believe that these teams could also be seeded as part of an expanded protocol that would achieve both the sporting and geographic objectives of FIFA.

    Seeding of 16 teams with 4 pots of 4 teams

    Of the 32 teams that ultimately qualify for WC2014, the top 16 ranked by the October 2013 FIFA Rankings would be seeded through four pots of four teams, with the host nation and the top 7 teams comprising the top seed in each group as previously announced. However, these seeded 8 teams would be assigned to two pots. As an example (where favored teams in playoffs all qualify), these first two smaller pots (based on ranking) would be:

    Pot 1a: BRA, ESP, GER, ARG

    Pot 1b: COL, BEL, URU, SUI

    For sporting reasons, Pot 1a would be drawn into Groups A, D, E, H to assure the top 4 wouldn't meet before the semi-finals if they all win all matches up to that round. Pot 1b would be drawn into the remaining groups B, C, F, G.

    The next 8 highest ranked teams would comprise the second seeds in groups. They too would be grouped into two pots. Continuing the example, these next two smaller pots (based on ranking) would be:

    Pot 2a: NED, ITA, ENG, CHI

    Pot 2b: USA, POR, GRE, BIH

    For sporting reasons, Pot 2a would be drawn into groups B, C, F, G (which were seeded by Pot 1b). In this example, CHI would be drawn first against either BEL or SUI (to assure no group has more than one CONMEBOL team); then the remaining three UEFA teams would be drawn against COL, URU, and the undrawn BEL/SUI. Note how the teams of Pot 2a are separated from the teams of Pot 1a.

    Pot 2b would be drawn into groups A, D, E, and H (which were seeded by Pot 1a). In this example, no special geographic consideration would impact this pot.

    The remaining pots would need to assembled in a manner that reflected which teams ultimately qualify. In this example, next to be placed into a group would be the one remaining CONMEBOL team ECU to ensure no group has more than one CONMEBOL team, as well as the three remaining UEFA teams to ensure no group has more than two UEFA teams. In this example, four more teams would needed to complete the assignment of eight teams into a third spot, and there are four qualified teams from AFC that would fill out these spots.

    The fourth spot in each group, in this example, would be the three remaining CONCACAF teams and the five CAF teams. The geographic considerations here are only one: the CONCACAF teams must be drawn away from the USA second seed.

    * * * * *

    So, 16 teams seeded through four pots of four teams expands upon the top eight teams being seeded by not only seeding the top sixteen teams, but also by combining the bottom half of the top eight with the top half of the second eight into the same four groups (B, C, F, G). Simple, actually. In essence, this approach would smooth the distinction between a seeded team and a second seed, an attractive prospect for sporting reasons.

    Your truly,

    Dorian Fans


    ReplyDelete
    Replies
    1. They will try to keep it simple. I, personally, like the way the pots are heading to be...

      Delete
  45. Splitting the Top 4 until the semis (assuming they win their groups) is a very good idea, especially as the look like a very elite group. Where it could get messy is the last 16. If all goes to seeding (unlikely I know) then the Pot 1a teams will meet the Pot 2a teams which will mean some heavyweight departures in the 2nd Round. I'd actually go a step further than your suggestion and split all 4 pots on ranking alone rather than geography (including your Pot 1a & 1b idea). Geographic considerations could still be taken into account similar to how they run the Champions League group stage draw. It might be a bit complicated at the draw ceremony but FIFA love all that complexity anyway!

    ReplyDelete
  46. What is the error for Italy in 2010?
    I have is 409.395 (see: http://andreybird.livejournal.com/44785.html)
    FIFA: 405.36
    According to my calculations Italy above Netherlands.
    The FIFA-vice versa.
    Thanks in advance!

    ReplyDelete
    Replies
    1. andreybird,

      the points for the Serbia match are 1335.0 instead of 1387.5. This match was suspended after 7 minutes and later awarded 3:0 to Italy. For the calculation of the match points the position of SRB was taken from the Oct-10 ranking instead of the Sep-10 ranking as you would expect.

      Delete
    2. Thank you, Ed!
      That is, it turns out that for such matches (AWD) is date of decision, and not when the match was to be held.
      Ed, I take this opportunity, do you have any comments about another error in my calculations for France in 2012 (see: http://andreybird.livejournal.com/44809.html)

      Delete
    3. Hi Ed, I was checking some of my numbers against your spreadsheet and I noticed that the ranking positions for 2010-02-03 are wrong. Fixing this will correct some of the fraction issues, for example for Brazil. Also, I noticed some more errors in the 2012-05-09 ranking, for example, Denmark and Argentina being swapped.

      Delete
    4. Homer, I don't understand exactly what you are saying here. Did you notice some errors in my sheet or are there errors in your numbers that you could fix with the assistance of my sheet ?

      In the first case, please be more specific (I couldn't find anything fault in my calculation of the Brazil points for instance); in the second case, you're welcome !

      Delete
    5. The errors are (were?) indeed in your spreadsheet, or maybe I had a version from several months ago. I extracted the ranking positions as a basis for calculating the ranking points and noticed that for the 2010-02-03 ranking the positions were almost all incorrect from fifa.com. For 2012-05-09 there were just a couple of errors. I'm now looking at other dates in 2010 and I forget the date but I think there were more errors. As such I'm now reworking the positions from 2010 in my database hoping that some of the fraction errors that I have for one or two nations in the top 15 will be resolved because of it. As I said, for Brazil correcting these ranking positions resulted in me getting the correct points.

      Delete
    6. Well, that's strange because, as I said, all the 4 time frame averages for each country are checked against the corresponding numbers presented at FIFA.com as soon as a new ranking is published. Each time it matches (with two decimals accuracy) for 100%.

      If match points for certain matches were calculated with wrong ranking positions then there surely would occur differences in detailed time frame averages between my calculation and FIFA's.
      So I am sure the match points for each match in my sheet are correct and are the ones that are used to calculate the FIFA points.

      As you calculate the used ranking positions from my match points and compare the resulting ranking positions with the ranking from Feb-10 could it be that:
      - the ranking positions for certain matches are taken from another ranking than you expected or
      - your Feb-10 ranking is flawed or
      - the Feb-10 ranking is adjusted afterwards by FIFA without adjusting the match points calculated with it.

      I just mention some possible causes but maybe if you could give a specific example of match points which are calculated with 'wrong' ranking positions then I will gladly research the case.

      By the way, if you have used the filedropper-link which I provided a few comments above to extract my sheet then you automatically work with the latest version of it.

      Delete
    7. Homer, I just saw that the ranking published Feb 3 2010 is actually the January 2010 ranking and the ranking published Mar 3 2010 is actually the February 2010 ranking. Could it be that you are confused by that ?

      Delete
    8. Hmmm, confused, me? Do you mean that the FIFA ranking on the FIFA site has date for the wrong month, or do you mean the spreadsheet has the wrong data? I was kind'a assuming the FIFA website was correct, but I'm not sure anymore now. Anyway, it's going to take me a few more days to double-check everything. There's a lot of data and a lot of calculations to check. I'm trying to automate the whole thing with code and a database behind it so that I can create a "live" ranking as games are played.

      Delete
    9. Homer, I meant that maybe you interpreted the ranking months around February 2010 not entirely correct as it could be confusing. As you don't give any numerical insight at what you have found out I can only mention possible causes for the 'errors' you found in my sheet.
      There is no wrong data in my sheet, unless you can provide a conclusive example that proves me wrong. I will wait for your comments in that respect.
      In the meantime good luck with the development of your system.

      Btw I would make the assumption that the FIFA website is normally correct with regard to rankings and time frame averages, just don't trust the prognosis tool.

      Delete
    10. Hi Ed, thanks for your persistence. I think I have solved the "problem", but I'm not sure what the problem really was. now I get the right result but I made several changes, including re-entering the ranking information Indeed it looks like the data in the spreadsheet is correct, at least for the data that I've verified so far. As you can probably appreciate, running multiple calculations for over 200 countries for 4 years worth of rankings has quite a few points for failure, so I'm not sure where the fix was. Don't get me wrong, it was not my intention to prove you wrong, rather it was to try and help you and others, thinking I had found a very small error. Anyway, thanks for your replies!

      Delete
    11. Do you or someone else know where the error is in the calculation for the 17 October 2013 ranking. For Greece, Portugal and others I get slight fraction errors in the calculations. The spreadsheet has the same numbers, but the FIFA website has different numbers:

      Portugal:
      2009-11-14T00:00:00.000+01:00 1185
      2009-11-18T00:00:00.000+01:00 1185
      2010-05-24T00:00:00.000+02:00 76,775
      2010-06-01T00:00:00.000+02:00 502,275
      2010-06-08T00:00:00.000+02:00 319,125
      2010-06-15T00:00:00.000+02:00 640,1
      2010-06-21T00:00:00.000+02:00 1054,5
      2010-06-25T00:00:00.000+02:00 792
      2010-06-29T00:00:00.000+02:00 0
      2010-09-03T00:00:00.000+02:00 342,5
      2010-09-07T00:00:00.000+02:00 0
      2010-10-08T00:00:00.000+02:00 1282,5
      2010-10-12T00:00:00.000+02:00 750
      AVERAGE 625,3673076923

      But FIFA says: 625,58

      Greece:
      2009-11-14T00:00:00.000+01:00 445
      2009-11-18T00:00:00.000+01:00 1335
      2010-03-03T00:00:00.000+01:00 0
      2010-05-25T00:00:00.000+02:00 86,95
      2010-06-02T00:00:00.000+02:00 0
      2010-06-12T00:00:00.000+02:00 0
      2010-06-17T00:00:00.000+02:00 1986,9
      2010-06-22T00:00:00.000+02:00 0
      2010-08-11T00:00:00.000+02:00 561
      2010-09-03T00:00:00.000+02:00 225
      2010-09-07T00:00:00.000+02:00 465
      2010-10-08T00:00:00.000+02:00 990
      2010-10-12T00:00:00.000+02:00 1230
      AVERAGE 563,45

      But FIFA says 563,52

      Delete
    12. For Portugal the friendly against Cape Verde on 24/5/2010 delivered 79,550 matchpoints instead of 76,775.
      For Greece the friendly against North-Korea on 25/5/2010 delivered 87,875 match points instead of 86,95.

      All matches played on 24/5/2010 and 25/5/2010 are calculated with the May-10 ranking. As far as I can see the correct match points for these matches and correct 4th time frame averages are included in my sheet.

      Delete
    13. Homer, you're welcome.

      The right way to use my sheet is as follows:
      1 calculate the match points as best you can, based on your own assumptions.
      2 check your match points against mine. If they are not completely the same then one of your assumptions is not correct. Usually it is the ranking from which the opponents position is taken.

      Here's the list with deadline dates per ranking month:
      30-07-2009 aug-09
      30-08-2009 sep-09
      15-10-2009 oct-09
      19-11-2009 nov-09
      11-12-2009 dec-09
      01-02-2010 jan-10
      28-02-2010 feb-10
      25-03-2010 mar-10
      23-04-2010 apr-10
      24-05-2010 may-10
      12-07-2010 jul-10
      05-08-2010 aug-10
      11-09-2010 sep-10
      16-10-2010 oct-10
      13-11-2010 nov-10
      13-12-2010 dec-10
      09-01-2011 jan-11
      30-01-2011 feb-11
      03-03-2011 mar-11
      08-04-2011 apr-11
      15-05-2011 may-11
      26-06-2011 jun-11
      25-07-2011 jul-11
      19-08-2011 aug-11
      15-09-2011 sep-11
      16-10-2011 oct-11
      17-11-2011 nov-11
      19-12-2011 dec-11
      12-01-2012 jan-12
      13-02-2012 feb-12
      04-03-2012 mar-12
      05-04-2012 apr-12
      03-05-2012 may-12
      04-06-2012 jun-12
      02-07-2012 jul-12
      02-08-2012 aug-12
      31-08-2012 sep-12
      30-09-2012 oct-12
      04-11-2012 nov-12
      16-12-2012 dec-12
      14-01-2013 jan-13
      11-02-2013 feb-13
      11-03-2013 mar-13
      08-04-2013 apr-13
      06-05-2013 may-13
      03-06-2013 jun-13
      01-07-2013 jul-13
      05-08-2013 aug-13
      11-09-2013 sep-13
      16-10-2013 oct-13
      Read and use this list as follows (for example):
      the matches played AFTER OR ON 11-9-2013 and BEFORE 16-10-2013 are calculated with the September 2013 ranking.
      In the following post I will give the single match exceptions on this list.

      Delete
    14. Here are the exceptions on the deadline-list above with the right ranking month used to calculate the match points for the match:
      2013-10-15 TRI NZL oct-13
      2013-06-09 TOG CMR jul-13
      2013-06-08 CPV EQG jul-13
      2013-06-08 BOT ETH jul-13
      2013-03-24 EQG CPV jul-13
      2013-01-08 RSA NOR mar-13
      2013-01-06 KEN BDI jan-13
      2012-12-16 BDI KEN jan-13
      2012-12-16 TOG BFA jan-13
      2012-12-15 COM MRI jan-13
      2012-12-15 SEY MOZ jan-13
      2012-12-15 MTN LBR jan-13
      2012-12-15 SLE GUI jan-13
      2012-12-02 MOZ SEY jan-13
      2012-12-02 GUI SLE jan-13
      2012-12-02 LBR MTN jan-13
      2012-12-01 MRI COM jan-13
      2012-12-01 BFA TOG jan-13
      2012-08-26 LBY SDN oct-12
      2012-08-15 CAN TRI oct-12
      2012-06-03 EST LTU oct-12
      2012-06-03 LVA FIN oct-12
      2012-06-03 TAH NCL jun-12
      2012-06-03 VAN SAM jun-12
      2012-06-03 NIG GAB dec-12
      2012-06-02 FIJ NZL jun-12
      2012-06-02 SOL PNG jun-12
      2012-06-02 SDN ZAM feb-13
      2012-06-02 BFA CGO jan-13
      2012-06-01 EST FIN oct-12
      2012-06-01 LVA LTU oct-12
      2012-06-01 SAM TAH jun-12
      2012-06-01 VAN NCL jun-12
      2012-05-31 FRA SRB oct-12
      2012-05-27 FRA ISL oct-12
      2011-12-16 POL BIH mar-12
      2011-12-14 OMA KUW mar-12
      2011-07-28 TJK SYR aug-11
      2011-07-24 LBR GAM jul-11
      2011-07-23 SYR TJK aug-11
      2011-01-07 QAT UZB jan-11
      2011-01-08 KUW CHN jan-11
      2010-10-12 ITA SRB oct-10
      2010-01-31 JAM CAN jan-10
      2010-01-23 THA DEN jan-10
      2010-01-20 SIN DEN jan-10
      2010-01-17 POL DEN jan-10

      Delete
    15. Hi Ed, fantastic! I already added extra tables to my database to easily be able to deal with such exceptions.

      Delete
    16. Hi Ed, can you confirm that AFG v IND on 11-09-2013 is also calculated using the sep-13 ranking? Otherwise the numbers don't match.

      Also, I've noticed some other oddities for which I don't have a really good explanation. For the matches played on 2010-10-17 (Dominican Republic versus Dominica, Guyana versus Surinam and St. Lucia versus Netherlands Antilles) they are added to the "fourth" (previous) year with a weighting of 0.2, rather than being included in the third year with a weighting of 0.3. But, for the Poland versus England match, which was postponed from 2012-10-16 to 2012-10-17, this match is included in the "last" (next) year, getting a weighting of 1. Thinking logically the last match should be included in the "third" (previous) year...

      Delete
    17. Homer,

      AFG-IND is calculated with the sep-13 ranking as it is played on 11-9-2013 and as stated in my post about the deadline dates per ranking month "the matches played AFTER OR ON 11-9-2013 and BEFORE 16-10-2013 are calculated with the September 2013 ranking."

      The matches played on the publishing date of a ranking (but in previous years) are assigned to time frames as follows (example publishing date 17-10-2013 for the October 2013 ranking):
      matches played on 17-10-2012 are assigned to time frame 1;
      matches played on 17-10-2011 are assigned to time frame 3;
      matches played on 17-10-2010 are assigned to time frame 4;
      matches played on 17-10-2009 are assigned to time frame 4.

      It is not logical and I do not have an explanation for it. But that's how FIFA calculates.

      Delete
    18. Hi Ed, OK, as always thanks!

      Other than for teams with matches on those boundary dates and for Curacao, I now have the correct numbers for all countries, using my own database of results, and using algorithms based on time (i.e. for aggregate results with penalties before and after November 2012). I'll do the same for those boundary dates, but if the "FIFA rule" is not clear and logical we'll have this "problem" for each ranking month. I'll try afterwards to see if I can find a pattern. For Curacao it is yet another "special" FIFA case that they seem to calculate matches played by Netherlands Antilles as if they were played by Curacao. Sigh!

      Delete
    19. Homer,

      with regard to the boundary dates: luckily FIFA uses always the same, illogical, assignment as described above for each ranking month, so program it once and you've got it covered :)

      Indeed, the matches of the Netherlands Antilles before 31-10-2010 are considered as played by Curacao for the calculation. It concerns 2 matches played in October 2009 against Suriname and Guyana for the Suriname Independence Cup 2009 and 3 matches played in October 2010 against Suriname, Guyana and St. Lucia for the qualifying stage of the Caribbean Nations Cup 2010.

      Delete
    20. This looks like a possible pattern, maybe a little far-fetched, but... The theory being that the period for the match date is the period where the match date is between the start date for the ranking date and the ranking date itself, or greater than the ranking date itself.

      (columns: start date for ranking, ranking date, month number, match date in question)

      Period 4
      2009-08-30, 2009-10-16, 1, 2009-10-17 (greater than ranking date)
      2010-09-11, 2010-10-20, 12, 2010-10-17 (between start date and ranking date)

      Period 3
      2010-10-16, 2010-11-17, 13
      2011-09-15, 2011-10-19, 24, 2011-10-17 (between start date and ranking date)

      Period 2
      2011-10-16, 2011-11-23, 25
      2012-08-31, 2012-10-03, 36

      Period 1
      2012-09-30, 2012-11-07, 37, 2012-10-17 (between start date and ranking date)
      2013-09-11, 2013-10-17, 48

      This is easy to program also, and seems logical (kind of), but we’ll see for future rankings.

      Delete
  47. POR - SWE
    UKR - FRA
    GRE - ROM
    ICE - CRO

    Is it possible to get probabilities ... which NT will be in that special pot. Most likely a procedure cfr 2006 is used and I'm guessing the October ranking will be used as basis for the lowest ranked UEFA NT.

    ReplyDelete
  48. 10.000 simulations for the UEFA play-offs based on elo-probabilities:

    POR-SWE
    1st leg: 41% win POR - 27% draw - 32% win SWE
    2nd leg: 35% win SWE - 27% draw - 39% win POR
    qualification: 52% POR - 7% PSO - 41% SWE

    UKR-FRA
    1st leg: 32% win UKR - 27% draw - 41% win FRA
    2nd leg: 45% win FRA - 27% draw - 29% win UKR
    qualification: 55% FRA - 7% PSO - 38% UKR

    GRE-ROM
    1st leg: 48% win GRE - 26% draw - 26% win ROM
    2nd leg: 29% win ROM - 25% draw - 46% win GRE
    qualification: 60% GRE - 7% PSO - 33% ROM

    ISL-CRO
    1st leg: 15% win ISL - 20% draw - 65% win CRO
    2nd leg: 64% win CRO - 20% draw - 16% win ISL
    qualification: 81% CRO - 4% PSO - 15% ISL

    ReplyDelete
    Replies
    1. Interesting to see how heavily favoured Greece and Croatia are over Romania and Iceland respectively.

      Certainly Greece could win but I fancy Romania to qualify from that tie.

      Whilst I do expect Croatia to win, I would not place the probability as high as 80%. Croatia's form in the last 3-4 qualification games was very poor in relation to their ranking. Maybe somewhere nearer 65%

      I guess this is the limitation of raw statistical data, it can't factor in human vagaries such as form/rection to pressure (or lack of such)/ team harmony (or disharmony)/ managerial factors etc.

      Of course I am assuming I am right and the statistical data will be wrong - a big assumption!!

      Delete
    2. Daniel, you know what they say: you have lies, big lies and then statistics :)
      I like statistics/formulas because they are as objective as possible. All the factors you mention, it all becomes easily too subjective for me and then you get those rather 'pointless', but not unamusing, discussions or should I say exchanges of personal opinions.

      The statistical data behind these simulations are at least simple: the elo win expectancy, derived from the difference in elo ratings between the two teams concerned. For instance the ratings of Croatia (1781) and Iceland (1510) at the moment leads to a win expectancy for Iceland at home of just 0.27. No argument about Croatia's poor form in the last qualifiers, they are still a very strong team (with a new coach and elan now) and I think (just my opinion) they will qualify easily.

      Delete
    3. It wasn't my intention to criticise the statistics, I think they are interesting, useful and as accurate as it is possible to be. There is not an automated statistical software programme/algorithm that could predict football outcomes anywhere near 100% accuracy, not even 90%. This is because, as I alluded to, the vagaries of the human dimension of sport - e.g. no algorithm would ever predict Armenia to draw 2-2 with Italy in Milan. And this unpredictability is a very good thing as football would be very boring without it! In a previous thread I was actually defending the ELO from others suggestuing it's % accuracy was not good :)

      Delete
  49. Edgar, when EURO 2016 qualifying draw seeding update will be ready? :)

    ReplyDelete
  50. Interesting!

    Did you perform similar simulations for MEX x NZL and URU x JOR?

    ReplyDelete
    Replies
    1. 10.000 simulations for the interconfederational play-offs based on elo-probabilities:

      MEX-NZL
      1st leg: 52% win MEX - 25% draw - 23% win NZL
      2nd leg: 24% win NZL - 25% draw - 51% win MEX
      qualification: 67% MEX - 6% PSO - 27% NZL

      JOR-URU
      1st leg: 11% win JOR - 17% draw - 72% win URU
      2nd leg: 72% win URU - 18% draw - 11% win JOR
      qualification: 88% URU - 3% PSO - 9% JOR

      Delete
    2. This comment has been removed by the author.

      Delete
    3. So Jordan have as much chance of winning in Montevideo as they do in Amman? Really?

      Delete
    4. This comment has been removed by the author.

      Delete
    5. James,

      while I was composing my reply I realised something was wrong. I hadn't included the home advantage in elo's home team win expectancy. So thank you, sharp reading !

      Here are the correct results (also for the UEFA playoffs again where I made the same error, sorry all !).

      MEX-NZL
      1st leg: 62% win MEX - 21% draw - 17% win NZL
      2nd leg: 33% win NZL - 27% draw - 40% win MEX
      qualification: 65% MEX - 6% PSO - 28% NZL

      JOR-URU
      1st leg: 17% win JOR - 22% draw - 61% win URU
      2nd leg: 79% win URU - 13% draw - 7% win JOR
      qualification: 87% URU - 3% PSO - 10% JOR

      POR-SWE
      1st leg: 50% win POR - 26% draw - 24% win SWE
      2nd leg: 44% win SWE - 27% draw - 29% win POR
      qualification: 51% POR - 7% PSO - 42% SWE

      UKR-FRA
      1st leg: 42% win UKR - 27% draw - 31% win FRA
      2nd leg: 53% win FRA - 24% draw - 23% win UKR
      qualification: 54% FRA - 7% PSO - 39% UKR

      GRE-ROM
      1st leg: 58% win GRE - 23% draw - 18% win ROM
      2nd leg: 37% win ROM - 27% draw - 36% win GRE
      qualification: 61% GRE - 7% PSO - 32% ROM

      ISL-CRO
      1st leg: 23% win ISL - 24% draw - 53% win CRO
      2nd leg: 74% win CRO - 16% draw - 10% win ISL
      qualification: 80% CRO - 5% PSO - 15% ISL

      Gladly the qualification chances remain nearly the same. Both home advantages cancel each other out.

      Delete
    6. Nigeria and Cote d'Ivoire qualified yesterday for Brazil, as expected.
      Here are the qualification chances for each of the teams in the remaining World Cup play-offs, based on elo-probabilities and given the result of the first leg (in parentheses the qualification probabilities before the first leg was played):

      Tunisia: 41% (53%)
      Cameroon: 51% (40%)
      PSO: 8% (7%)

      Ghana: 99% (48%)
      Egypt: 1% (44%)
      PSO: 0% (8%)

      Burkina Faso: 53% (45%)
      Algeria: 45% (47%)
      PSO: 2% (8%)

      Mexico: 99% (65%)
      New Zealand: 1% (28%)
      PSO: 0% (6%)

      Jordan: 0% (10%)
      Uruguay: 100% (87%)
      PSO: 0% (3%)

      Portugal: 69% (51%)
      Sweden: 20% (42%)
      PSO: 11% (7%)

      Ukraine: 81% (39%)
      France: 10% (54%)
      PSO: 9% (7%)

      Greece: 86% (61%)
      Romania: 11% (32%)
      PSO: 3% (7%)

      Iceland: 23% (15%)
      Croatia: 72% (80%)
      PSO: 5% (5%)

      Algeria-Burkina Faso remains the closest tie.
      Based on elo the four seeds in the UEFA play-offs (according to the FIFA-ranking) will all qualify.

      Delete
  51. This comment has been removed by the author.

    ReplyDelete
  52. So at this moment, these would be the most probable teams to be at the WC:

    Pot 1 (seeds): Brazil, Spain, Germany, Argentina, Switzerland, Belgium, Colombia and Uruguay

    Pot 2 (UEFA): Russia, Bosnia, Netherlands, Italy, England, Croatia, Portugal, Greece.

    Special UEFA Pot: France

    Pot 3 (CONMEBOL & CAF): Chile, Ecuador, Ivory Coast, Ghana, Nigeria, Burkina Faso (*), Tunisia.

    Pot 4 (CONCACAF & AFC): Australia, Japan, Iran, South Korea, Honduras, USA, Costa Rica, Mexico.

    So, a possible draw could be:

    Group A: Brazil, Croatia, Nigeria, Japan
    Group B: Switzerland, Portugal, Chile, Australia
    Group C: Belgium, Italy, Ghana, Costa Rica
    Group D: Argentina, England, Ivory Coast, USA
    Group E: Germany, Bosnia, Ecuador, South Korea
    Group F: Uruguay, Russia, Tunisia, Iran
    Group G: Colombia, Greece, France, Honduras
    Group H: Spain, Netherlands, Burkina Faso, Mexico

    (*) Not sure about Burkina vs. Algeria

    Juan GG

    ReplyDelete
  53. Hard draw for american seeds:

    Group A: Brazil, Italy, France, USA
    Group B: Switzerland, Russia, Burkina Faso, Iran
    Group C: Belgium, Bosnia, Ecuador, Costa Rica
    Group D: Argentina, England, Ivory Coast, Mexico
    Group E: Germany, Greece, Chile, Honduras
    Group F: Uruguay, Netherlands, Ghana, Australia
    Group G: Colombia, Portugal, Nigeria, South Korea
    Group H: Spain, Croatia, Tunisia, Japan

    Hard draw for UEFA seeds:

    Group A: Brazil, Russia, Burkina Faso, Australia
    Group B: Switzerland, Netherlands, Chile, Japan
    Group C: Belgium, Portugal, Ecuador, South Korea
    Group D: Argentina, Greece, France, Iran
    Group E: Germany, England, Ivory Coast, Mexico
    Group F: Uruguay, Bosnia, Tunisia, Honduras
    Group G: Colombia, Croatia, Nigeria, Costa Rica.
    Group H: Spain, Italy, Ghana, USA

    Not so-strong groups:

    Switzerland, Greece, Burkina Faso, Iran
    Belgium, Bosnia, Tunisia, Costa Rica
    Colombia, Greece, Burkina Faso, Australia

    Death groups:

    Brazil, Netherlands, France, USA
    Brazil, Italy, Ivory Coast, USA
    Spain, Italy, Chile, Mexico
    Germany, England, Chile, USA
    Argentina, Netherlands, France, Mexico

    ReplyDelete
    Replies
    1. Having these groups and final positions:

      A: 1) Brazil (Bosnia, Ivory Coast, Mexico)
      B: 2) Uruguay (Switzerland, Russia, USA)
      C: 1) Argentina (Portugal, Ghana, Iran)
      D: 2) Spain (Chile, Tunisia, Australia)
      E: 1) Germany (Croatia, Nigeria, South Korea)
      F: 2) France (Colombia, Greece, Japan)
      G) 1) England (Netherlands, Burkina Faso, Costa Rica)
      H) 2) Italy (Belgium, Ecuador, Honduras)

      We would surely have a finalist (after group phase) among Bosnia, Ivory Coast, Mexico, Switzerland, Russia, USA, Portugal, Ghana, Iran, Chile, Tunisia, Australia, Croatia, Nigeria, South Korea, Colombia, Greece, Japan, Netherlands, Burkina Faso, Costa Rica, Belgium, Ecuador, Honduras. Thus, a country who had never won the WC before, being Netherlands the only team to have been at the final (3 times).

      And great matches at the round of 16, the eight previous champions:

      Brazil-Uruguay, Argentina-Spain, Germany-France and Italy-England.

      Juan GG

      Delete
    2. Those groups are not possible!

      Delete
    3. This comment has been removed by the author.

      Delete
  54. It's true, I had some terrible mistakes. I try to remake it:

    Having these groups and final positions:

    A: 1) Brazil (Bosnia, Ivory Coast, Mexico)
    B: 2) Uruguay (Netherlands, Tunisia, USA)
    C: 1) Argentina (Portugal, Ghana, Iran)
    D: 2) Spain (Chile, Russia, Australia)
    E: 1) Germany (Croatia, Nigeria, South Korea)
    F: 2) France (Colombia, Greece, Japan)
    G) 1) England (Switzerland, Burkina Faso, Costa Rica)
    H) 2) Italy (Belgium, Ecuador, Honduras)

    We would surely have a finalist (after group phase) among Bosnia, Ivory Coast, Mexico, Switzerland, Russia, USA, Portugal, Ghana, Iran, Chile, Tunisia, Australia, Croatia, Nigeria, South Korea, Colombia, Greece, Japan, Netherlands, Burkina Faso, Costa Rica, Belgium, Ecuador, Honduras. Thus, a country who had never won the WC before, being Netherlands the only team to have been at the final (3 times).

    And great matches at the round of 16, the eight previous champions:

    Brazil-Uruguay, Argentina-Spain, Germany-France and Italy-England.

    I think it's ok now.

    Juan GG

    ReplyDelete
    Replies
    1. Is the "spell" broken? :-) Since last world cup was the first in history to have neither Brazil, nor Argentina, nor Germany nor Italy in the final, I wonder what kind of surprise is reserved for this time

      Delete
    2. Well, I think that Brazil will surely be at the final (and in fact I think they'll win; I don't think they are the best team, but they are quite good and playing at home... weather, torcida, referees, etc...).

      But I think there can be surprises this time concerning the other finalist; after 2 WC dominated by UEFA teams, I think they won't do very well in Brazil, the weather is one of the key facts (we saw at the confederations cup how difficult was for Italy or Spain), specially due to timetables (they'll be played soon so it's not so late in Europe = even more hot).

      CONMEBOL teams, not only Argentina but Colombia, Chile, Uruguay, will get really far, I think. And maybe it's the time for an african team (Ivory Coast?).

      I expect at the top eight at least 4 or 5 CONMEBOL teams (Brazil, Argentina, Chile, Uruguay, Colombia...), an african team (Ivory Coast, Ghana...) and obviously there should be any UEFA team (2-3): Germany (reaching top 8 since 1954), maybe Spain (but they're not so strong now), Italy, Netherlands, etc. But maybe there will be surprises (Belgium, Japan, etc.).

      Juan GG

      Delete
  55. I know this was used in 2006 WC but I don't see the need for a special pot for one of the European Nations and don't think it is particularly fair.

    If it is to be France as the lowest ranked team then I suppose it is fair on the cold hard data, but the France of Oct 13 is likely to be different from the France of Jun 14, although we can't base ranking/seeding decisions on predicted future performance.

    In 1998, there were 9 teams in the UEFA pot. The procedure was that the 8 teams would be drawn then the ninth team would be allocated to the Brazil or Argentina group. Was a fairly simple procedure that could be replicated for 2014 and avoids condeming one team to the certainty of facing one of the South American teams, although in 2014 this would be less harsh with Colombia and (probably)Uruguay as possibilities, it is still a 50% chance of facing Brazil or Argentina as opposed to 25% for the others - not that Uruguay or Colombia are vastly weaker than these two.

    I remember this clearly because in 1998 my own team Scotland were drawn in Group A and then had the misfortune to end up with Norway rather than a team from the weaker Asia/Central America pot, although it turned out that ironically, Norway was the only team we took a point from (should have won) and lost heavily to Morrocco who would be considered weaker.

    ReplyDelete
    Replies
    1. Yep and in 2002 they actually had 11 UEFA teams in one pot (there were 15 teams back then).

      I actually am dead against grouping pots of different federations. It means for example that we have no chance whatsoever of seeing Australia play Mexico now (let's face it - neither is getting past the group stage whatever draw they get). I understand the need to keep countries of the same federation apart during the group stage but surely that can be done somehow with confederation having their own pot?

      Delete
    2. Yes, I absolutely agree with this. They could simply create 4 pots based solely on world rankings, then teams from the same confederation could still be kept apart by moving them to the next eligible group where there is a clash. The 2-Euro maximum per group could also still be easily maintained. This method should achieve a better balance of quality across the 8 groups i.e. less chance of a Group Of Death, and should also increase the chances of a better continental mix.
      As things stand, there could be a group with Spain, Italy, Chile & USA i.e. four teams ranked in the Top 13 from which 2 will have to be eliminated before the 2nd Round. Meanwhile, you could also have Uruguay, France, Romania, Australia where there is only 1 team in the Top 20. This unnecessary imbalance every World Cup really annoys me. In just about every other sport, teams are properly seeded in major tournaments. Then again, this is FIFA!

      Delete
    3. I just think it was especially unfair on Serbia in 2006 who were almost by default condemned to a "Group of Death" in that they would automatically face Brazil or Argentina and one other UEFA team. Ivory Coast was just the icing on the cake. Whichever UEFA team faced the same scenario this time would be automatically placed in a Group with one CONMEBOL and one UEFA team and would be denied the more attractive prospect (with no disrespect intended) of drawing Switzerland. Arguably, Netherlands and Belgium would be more desirable for a UEFA team than even Uruguay or Colombia in South America. Given the pretty uncomplicated procedure used in 1998 I don't see a logical reason for changing it on 2006. I may be a bit cynical, and mean no disrespect, but France is not Serbia, and FIFA might be likely to reconsider use of the 'special pot' on this occasion. I think this would be fairer even if it was done for cynical reasons.

      Re Australia and Mexico - Australia qualified for second round in 2006 and Mexico have regularly qualified for the 2nd round in 1994, 1998, 2002, 2006 and 2010 QFs in 1970 and 1986, so a bad example to use

      Delete
    4. Yeah, 2006 was a classic example of this. That :"special pot" was a total nonsense and helped lead to this hideously difficult group while there were 1 or 2 fairly weak groups elsewhere.

      Delete
  56. Ed - any chance you understand Russian? :)

    ReplyDelete
  57. No chance, but I've made my contribution to the Netherlands-Russia Friendship Year anyway :)

    ReplyDelete
  58. Ещё раз спасибо!
    Thanks again!

    ReplyDelete