Friday, October 16, 2009

FIFA Business: Seeding Formula, Pots, & 2014 Spots

1) The 2010 seeding formula. Using the previous formula (from 2006), it appears that Portugal's resurgence may see them get seeded after all at the expense of France. They were always in good shape as long as they managed to qualify, and that looks very likely now. The same eight teams have been in the running ever since July 2006, so one good team is going to get let out no matter what.

Now despite what the formula shows us currently, we can never be certain as to what FIFA will actually do. The average person on the street would probably say that the previous World Cup runner up (France) should be seeded ahead of a team that barely qualified (Argentina). Remember, they've changed it slightly the past few World Cups.

BTW, I didn't see this Soccernet article until now. US Soccer president Sunil Gulati doesn't know what the formula was in 2006. He said it included "...50 percent for performance over the last three World Cups" which is incorrect. Acutally, they only used the past two World Cups for the 2006 formula. Also wrong is author Frank Dell'Apa, so I guess that gives you an idea of how many people really follow this stuff closely. I'd really expect Gulati to know better.

Anyway, there's speculation that this time only the current FIFA ranking could be used, maybe by itself or in conjunction with the previous World Cup performance half of the formula. I would approve such a decision. Why? Well, one thing that I always forget about until somebody mentions it is the fact that because the FIFA rankings half of the formula includes snapshots from three separate years (12/2007, 12/2008, 11/2009), matches from some years count more than others.

The current FIFA rankings procedure takes into account the past four years worth of results (before 2006 it was eight years). Those results are weighted 100-50-30-20. If the seeding formula stays the same and uses three years worth of FIFA rankings, what they're doing is this:

Dec-07 Dec-08 Nov-09 Total PCT

100 100 16.7%
100 50 150 25.0%
2007 100 50 30 180 30.0%
2006 50 30 20 100 16.7%
2005 30 20
50 8.3%
2004 20

20 3.3%

So matches from 2007 are counted the most, closely followed by 2008. Compare this to only using Nov-09:


Current Nov-09
2009 17% 50%
2008 25% 25%
2007 30% 15%
2006 17% 10%
2005 8%
2004 3%

It makes more sense to do it this way when determined who should be seeded for next year. The whole idea of using the rankings in the formula is to measure who's the best right now, isn't it? And then since the other half involves previous World Cups, that half is for who was the best in the past, which gives you a formula that involves both past and present.

2) Not only will FIFA have to decide what formula to use, they'll also have to decide the four pots for the World Cup draw. Unless they decide to seed all pots (which IMO is a good idea, but that's another post), the top pot should consist of seeded teams and the second pot of UEFA teams. That leaves the following 16 teams left:

4.5 - AFC
5.0 - CAF
0.5 - OFC

Last time, AFC/CONCACAF/OFC was one pot and CAF/CONMEBOL was another. Before, it ws CAF/CONCACAF. What will they do this time?

If Uruguay wins the playoff with Costa Rica, it makes things simple. CONCACAF and CONMEBOL would each be at three teams, and either could be paired with the five teams of AFC/OFC or CAF. If Costa Rica & New Zealand win, then you would expect the only option to be AFC/CONCACAF (4 + 4) and CAF/CONMEBOL/OFC (5 + 2 + 1).

But if Costa Rica & Bahrain win, things get messy:

5.0 - AFC
5.0 - CAF
0.0 - OFC

Now this is where Edgar has suggested we could see the return of the dreaded "special pot" which was used for Serbia & Montenegro in 2006 when things didn't add up evenly. That's something to look out for.

Or, they could just draw the groups instead of the teams which would be the fairer way to do it. That probably makes too much sense for FIFA, though.

3) I was discussing the possible 2014 World Cup qualifying playoffs on Bigsoccer the other day. No, it's not too early. Will the number of spots for each confederation change? Here's what I was thinking:

The interesting thing for 2014 will be how Brazil hosting affects things. For 2010, they decided to give CAF 5 spots plus the host spot for a total of 6. That never used to be the case; before, they would've given CAF 4+host for a total of 5 (taking the host spot out of the normal number).

So if they continue with the 2010 policy, Brazil as host should mean their spot wouldn't come out of the 4.5 for CONMEBOL. However, 5.5 total spots for that region (out of 10 teams) seems like a lot. Remember, for 2006 they originally reduced it to 4 (giving OFC a full spot) before changing it back. I would be kind of surprised to see CONMEBOL with 5.5.

Given the way AFC didn't lose their spots after 2002, I wonder if FIFA might do something like give CONMEBOL only 4+host and then give the extra 0.5 to CAF.


Labels: , , , ,

Comments on "FIFA Business: Seeding Formula, Pots, & 2014 Spots"


Blogger ERic said ... (11:03 AM, October 16, 2009) : 

Thanks for pointing out the % on the seeding process. I've always had an intuition that there was something screwy in the way it was done in the past. Some sort of double-counting that made the seeding method not right somehow.

But it also seems somehow wrong to me to just use 4 years to determine strength of a team. Yes, there's a lot of player turnover in that time, but it misses picking up the history strength of a country. I wonder what would happen if they generated a special FIFA ranking just for the seeding process that went back 8 years, like the FIFA rankings used to?

Probably end up exactly the same.:)


Blogger Dorian said ... (1:06 PM, October 16, 2009) : 

Thanks for going through the percentages in section (1) regarding FIFA's use of rankings from 3 different time periods. Not only is interesting to see how 2007 would carry the most weight, but also how little 2004 and 2005 would count (just 11% together). Hence, all the more reason why FIFA should use only one ranking as of a single date (even in the unlikely event that they were to do a special ranking with more than 4 years of matches as they have all the data).

(Incidentally, the results from the last World Cup are captured in the 2004 results, and because World Cup matches receive a higher weight in the FIFA Ranking process, an argument could be made that 2004 World Cup results shouldn't be double counted either...)

Given your comments about how few people really follow this stuff closely, I hope that this analysis gets to FIFA in time for them to come up with something better.


post a comment