PDA

View Full Version : Alternate World Cup Draw


Funko
30-06-2014, 09:40:02
I like this approach.

http://www.nytimes.com/2014/06/05/upshot/the-world-cup-draw-is-unfair-heres-a-better-way.html?ref=sports&_r=1

MoSe
30-06-2014, 11:23:51
I'm a long-time advocate of such system...

moreso, you know I told here in the past that I'm against draws altogether
I'm for *integral seeding*, that is, you take the ranking, and you *seed* all groups snake-like, first down for first teams, then up for 2nd teams, then down, then up
This way for instance group A would have seeds 1-16-17-32, while group H would have 8-9-24-25. The seeds sum in all pots would be 66.
You'd only allow for little arrangements, skipping/switching spots when necessary to ensure geographic separation.

I know you objected that both in Champions League, or in Tennis, this would often tend to produce the same matches, and to make early turns less interesting
I think tho that rankings are variable enough to prevent this

I conducted an analisys about integral seeding for this WC, using both the Oct'13 ranking adopted for the official draw, and the last available one (Jun'14) b4 the WC started

will post it when I'll find a moment free from Fantasy burdens ;) (and work:D)

Fistandantilus
30-06-2014, 13:02:47
They look way more balanced but that shouldn't be a surprise when you prioritize rankings over goegraphy. I have no idea why they do the other way around.

Funko
30-06-2014, 13:06:06
We hardly ever have competitive matches cross confederation, they are trying to make us have a many as possible of those.

Venom
30-06-2014, 15:22:10
Except that the FIFA rankings are a rank pile of shit. To follow what is prescribed, they'd need a better way of ranking. Or at least a way that it it updated constantly.

Also, in no way is the group he drew the US into easier to get out of. I think it's even harder. Americans are overly reliant on rankings to decide things.

MoSe
30-06-2014, 15:37:41
true dat, but...
either you use no ranking at all, or you take what you have, until you can devise a better one.

and without a ranking, you should go with a panel or committee "wise" decisions, which could be much worse
___
BTW

IIRC there was a site tracking a boxing-like World Champion for soccer too

having played chess in my youth, I'd of course say that the best ranking is ELO system
wonder if someone bothered to keep track of that too, for soccer

MoSe
30-06-2014, 15:38:45
hey, just needed to google it !

http://www.eloratings.net/

Venom
30-06-2014, 15:39:58
The "wise" committee is what they use here for the college basketball tournament (commonly called March Madness). And while usually accurate there are always some...disagreements about the seedings.

Funko
30-06-2014, 15:45:52
I don't think the FIFA rankings are that bad. The problem is that teams from across confederations don't play each other enough in meaningful tournaments to make the cross confederation rankings meaningful (a problem with all systems) which might be why FIFA doesn't group by world ranking but by confederation.

Funko
30-06-2014, 15:48:43
The problems

1. Not enough cross confederation competitive football
2. Very infrequent games

would be problems for any ranking system. So yeah, given that I'm not sure that the current one is that bad really.

Venom
30-06-2014, 15:50:21
I agree that the way it is, is bad.

It's just that you can't rely on just a mathematical formula. At some point, a human element needs to be inserted in the rankings. So that if these teams don't play, the logic leaps required to estimate the true rankings can still be made. And, as mentioned, that's rife for fuck ups as well, even if done by committee. But, is that better than the random pots? Perhaps. I just don't trust FIFA not to fuck anything up.

MoSe
30-06-2014, 16:00:47
so, you're [Funko] suggesting that within a confederation the ranking is acceptably meaningful
thus, the problem would only be to properly "adjust", like with sliders, the respective confederations ...scales (?) so that they're better... aligned (?)

they do have a confederation coefficient inside the FIFA Ranking, I probably agree that the interconfederation coefficient is not ...efficient enough, due to the above problems you highlight

tha main problem I always saw in FIFA Ranking, even within a single confederation, is that friendlies are weighed less, but the make average for their running year, and the ranking then sums (with decreasing weight) the signle years averages
so, if you play more friendlies, despite winning them all (leave alone losin'em as Italy does) you'll end to *lower* your ranking compared to a team who have your same official matches results but play NO friendlies...

hey, Fance v Nigeria kicks off in <20 minutes! :)

MoSe
30-06-2014, 16:03:32
wow, I agree with Venon :eek: in general
although, in practice, my trust in the human element is so low that I prefere to fall-back on pure math, with all its downsides :D

Funko
30-06-2014, 17:16:29
A human element?! What would we call it, the FIFA Bribery Ranking Component?

Funko
30-06-2014, 17:18:26
I think I'm saying that any replacement system would have flaws based on the problems with the data collection.

Maybe just make the draw completely random, stop favouring the big teams through ranking.

And I think I posted here that analysis that if England had played 2 fewer friendlies they'd have been seeded instead of Switzerland, so :mad:

Fistandantilus
30-06-2014, 17:23:15
Friendlies shouldn't count at all.

Funko
30-06-2014, 21:28:43
Friendlies should have the highest points coefficient so that people don't treat them as Friendlies and it's like Rugby where every match is a competitive international.