Forum

xml	No replies possible in the archive

Author: badgerboy
Date: 06-06-2006, 14:31

OK - so I have been looking at various methods of calculating my own "coefficients" in advance of the World Cup. I posted World Cup only results (for all countries - World Cup Finals matches only) on the predictions thread.

Next I have tried to compare European teams only - based on results in World Cups European Championships - both qualifiers and finals. All figures are based on results since the start of the 1998 World Cup qualifying campaign.

I have used a number of different methods and the results are quite different, so I'm looking for opinion on which (if any!) looks the most accurate in "mathematical terms".

I guess the first two questions I have are more a matter of opinion than mathematics.

1. Should bonus points be awarded for appearances in the finals or should points simply be awarded for matches played? I haven't taken into account the possible weakness of some qualifying round opponents as (aside from teams who directly qualify) the number of such matches should even out over the period.

2. Straight coefficients over 5 years as per UEFA club coefficients or weighted coefficients (100% for latest qualifying campaign, back to 20% for World Cup 98).

3. The third question is the most mathematical (not my strongest point!). A straight average for all matches played over the five tournaments or a sum of the averages for each individual tournament? Without really knowing why I like the latter but it does cause more problems when taking into account missed qualifying campaigns and the like.

Apologies for the extreme length of the thread. There are eight variations and the results are listed below:

1. A simple average - total pts (2 for a win, 1 for a draw)/total matches played.

France 1.563
Spain 1.557
Netherlands 1.536
Portugal 1.518
Italy 1.500
Sweden 1.491
Czech Rep. 1.467
Germany 1.444
England 1.433
Serbia 1.411
Croatia 1.397
Ukraine 1.268
Poland 1.234
Switzerland 1.128

2. As for option 1 but with bonus points added - 3 for reaching a finals tournament and 1 each for every knockout round reached.

France 1.979
Netherlands 1.804
Spain 1.803
Italy 1.790
Germany 1.778
Portugal 1.768
England 1.700
Sweden 1.684
Croatia 1.603
Czech Rep. 1.600
Serbia 1.554
Poland 1.298
Ukraine 1.268
Switzerland 1.191

3. Exactly the same figures as those used in option 1 but this time with the averages for each individual year totalled. In order to give Germany "equality" I have included their results in Euro 96 to compensate for their lack of a qualifying campaign for this tournament.

Netherlands 7.871
Spain 7.769
Italy 7.500
Sweden 7.443
Germany 7.387
Portugal 7.378
England 7.245
Czech Rep. 7.228
Croatia 7.044
France 7.003
Serbia-M. 6.938
Ukraine 6.250
Poland 6.033
Switzerland 5.616

The possible weakness with this and to a lesser extent, the next method is the effect on France's score for 2002 based on their failed finals tournament without a qualifying campaign to offset it.
Portugal's score for 2004 is also relatively low despite their having reached the final.

4. Option 3 with bonuses included.

France 9.711
Netherlands 9.567
Germany 9.052
Spain 8.949
Portugal 8.942
Italy 8.868
England 8.527
Sweden 8.334
Croatia 7.923
Czech Rep. 7.844
Serbia-M. 7.521
Poland 6.263
Ukraine 6.250
Switzerland 5.889

5. Option 1 with results weighted. As for all remaining options Euro 96 results have been included for the Germans and other matches moved forward accordingly. The equivalent of home advantage?

Netherlands 0.957
France 0.938
Czech Rep. 0.930
Sweden 0.912
Portugal 0.900
Spain 0.898
England 0.870
Germany 0.840
Italy 0.835
Croatia 0.821
Poland 0.817
Serbia-M. 0.761
Ukraine 0.754
Switzerland 0.745

6. Option 2 (average points including bonuses) with results weighted.

France 1.121
Netherlands 1.071
Portugal 1.054
Germany 1.034
Sweden 1.032
Spain 1.023
Czech Rep. 1.017
England 1.007
Italy 0.968
Croatia 0.914
Poland 0.855
Serbia-M. 0.804
Switzerland 0.796
Ukraine 0.754

7. Option 3 (totalled individual tournament averages) with weighting.

Netherlands 4.721
Spain 4.608
Sweden 4.538
England 4.528
Czech Rep. 4.512
Italy 4.505
Portugal 4.480
Croatia 4.328
Germany 4.302
Serbia-M. 4.153
France 4.096
Poland 3.910
Ukraine 3.783
Switzerland 3.548

8. Finally! Option 4 (Totalled individual tournament averages with bonus) but weighted.

Portugal 5.552
Netherlands 5.420
France 5.313
Germany 5.276
Spain 5.203
England 5.185
Italy 5.161
Sweden 5.085
Czech Rep. 4.912
Croatia 4.751
Serbia-M. 4.336
Poland 4.048
Ukraine 3.783
Switzerland 3.766

I'll have a go at assessing the overall results on another post since this one is so long already....

Author: ignjat63
Date: 06-06-2006, 14:54

If you are adding up the points for victory or draw and you consider only final tournaments (no continental qualies), then I suppose you should just add the points (no bonus) and not count the average. Because the most successful team is not the one with the highest average but one with most points won.

But here are some thoughts on the method. WC final tournament is half league and half cup. But is not the league part a prolonged cup stage really? Because of 32 teams 16 go through - as if it was a cup stage. Ans so essentially all the teams that are through made the same success - regardless of number of points.

Also, teams that lost semies made equall success. The same for losing all the way up. So I'd rather link poins to stages rather than count points for winning and drawing.

I wonder can a situation arise in which a losing semifinalist actually gets less points than losing quarterfinalists. That would not be right, would it?

Re: World Cup Statistics - best coefficient method?
Author: panda Date: 06-06-2006, 14:58	What is your aim with these stats? If it is predictive in any sense, maybe weighted ranking and averaging within each tournament gets you a better result? If it's 'history' then maybe don't weight it. BPs give you an idea what team performs on the big stage?

Re: World Cup Statistics - best coefficient method?
Author: executor Date: 06-06-2006, 15:46	{i> In order to give Germany "equality" I have included their results in Euro 96 to compensate for their lack of a qualifying campaign for this tournament. {/i> Did you do the same for France, Netherlands and Portugal? They all have hosted a tournament since 1996 (the start of WC98 qualifiers)

Author: badgerboy
Date: 06-06-2006, 16:18

Ignjat63

I think I was looking to show something of the opposite of what your method would do. So a team losing on penalties would not be shown as "inferior" to the one passing them (except in the case of bonus awards) and a team winning three group games and then losing in the quarter-finals could indeed get more credit than one scraping through with a win and a draw and going on to the semis (maybe again only on penalties). Your method probably better allows for aspects like "mental toughness" but I like cold hard figures.

I do agree with your assessment of how best to deal with results when looking purely at final tournaments.

Panda - My aim is partially predictive and partially just "interested". Trying to adapt UEFAs club coefficient system (or variations of it) for countries and seeing if the expected "top" countries are there statistically. To me the actual UEFA system doesn't really work due to the different number of matches played by the teams (qualify or not, qualifiers or not) so something near the old (pre 1979ish) system of working out team coefficients based on matches played is better. I'd probably agree with your second statement and can't decide if I like bonuses or not...

Executor. I didn't include Euro 96 for these countries. I only realled included it for Germany because of the lack of ANY matches (qualifying or final tournament) for Germany for this World Cup. I think the only country to really suffer (at least in the case of totalled averages) is France for World Cup 2002. In the other cases I'm happy to take it that the points picked up in the final tournament alone produced a reasonable "average" and that the home advantage also compensated for the loss of "easy" qualifying points.

Re: World Cup Statistics - best coefficient method?
Author: panda Date: 06-06-2006, 16:38	The problem with 'predictive' is obvious - the time period is long, necessarily so in order to get plenty of data. So how strong a team was in 98 may not have any bearing on how strong they are now....

Author: badgerboy
Date: 06-06-2006, 17:43

Hence the weighted option is better imo.

Interesting comment though as I partly started this exercise because I think a lot of people base their World Cup predictions on the historical performance of teams (a lot older history than 1998 in some cases too) and I wanted to see how recent performance measured up to this.

For example the bookies understandably have Brazil as favourites but next come:

England 7-1
Germany 7-1
Argentina 8-1
Italy 8-1
France 12-1
Holland 14-1
Spain 14-1
Portugal 22-1

These are Corals latest odds. OK England's odds are understandable due to weight of money bet but on all statistical methods the Dutch and the Spanish have a better chance than the Italians, the Dutch always better than the Germans and the Spanish mostly better than the Germans too. The Portuguese are strong too - actually coming out on top in the last table and ahead of Italy on 3 out of 4 of the weighted tables. Probably ahead in the 4th too if you take into account the lack of qualifying games by Portugal in 2004. England unfortunately don't look so good purely on paper either.

What this proves I don't know. Historically the same teams seem to perform well again and again when it comes to World Cups and I guess the stats may not have predicted those successes either. They certainly didn't predict Greece in 2004.

I think whatever the actual results are in the World Cup I find the statistics interesting. The assumption of which SHOULD BE the strongest teams aren't borne out by the statistics. Of course these "strong" teams may still win (they usually do) but why will they win?

Because they are stronger mentally than their opponents when it comes to the big stage? Because they have a psychological hold over many of their opponents because of their perceived strength? Because of penalty kicks or the way "luck" falls on any given day? Because statistics lie and the supposed superior teams really are better but always do "just enough" to get through? - Ignjat63's method of saying - "getting through is everything" would bear this out.

Author: panda
Date: 06-06-2006, 18:03
Edited by: panda
at: 06-06-2006, 18:09

What does 'doing well' mean in this context?

If any of England, Holland, Germany, France, Italy, Spain, probably Portugal does not reach the SF, its supporters think the team has failed.

Suppose you just do the stats with SF and F and WIN only.

Maybe that will give you a closer match between the (small) stats and the reality.

It's no good if you have 100% in the QR and then it goes wrong after that...

You can't judge necessarily on the betting since (apart from moneyon England) I presume it's supposed to reflect present strength.

Aha! Another thing is- do some countries have different performance in EC and WC?

Author: badgerboy
Date: 06-06-2006, 18:40

I'm trying to make some comparison between actual strength and perceived strength.

Italy. Won a qualifying group where the strongest opposition was made up of Norway, Scotland and Slovenia. Didn't win away against any of these teams. Norway went on to lose twice to the Czechs in the play-offs. Euro 2004. Eliminated at group stage by Sweden and Denmark. Only victory involved a 94th minute goal against Bulgaria. 2002 World Cup. Possibly unlucky against Korea (but no more so than Spain) but only won once at the group stage and would've been eliminated then if Ecuador hadn't beaten Croatia in the final round of matches.

But ask pretty much anyone and Italy are among the "favourites" to win the World Cup. I guess I'm searching for some contemporary evidence to support this.

Put another way. I know statistics (by whatever method) are just that - perhaps a guide to future performance. Maybe a lousy guide but certainly no more than a guide. I look at the stats and England come out a little bit worse than I expected but I'd balance that by saying "look at the team they have - if the core of the team gells they could have a hell of a World Cup". They lost to the best team in 2002 and to the hosts in a great game in 2004. In other words my knowledge (or maybe opinion) of the "actual" somewhat supercedes any statistical analysis. But when I look at Italy (again for example) the statistics back up my knowledge of recent football history and I wonder at their position - whether in bookmaker's lists or people's opinions.

Author: panda
Date: 06-06-2006, 18:44

I remember that the UBS projection says Italy wins. So does John Motson......

Maybe this is because they predict game by game head-to-heads.

Maybe what happens in WC is that you play opponents you might not have had much history of playing. So maybe your past record does not help.

I think there are some projections by Lyonnais (maybe forum 2?)where he takes the last fixture between 2 sides in head-to-head and some very odd results ensue, like Brazil come bottom of the group.

Re: World Cup Statistics - best coefficient method?
Author: ignjat63 Date: 06-06-2006, 18:50 Edited by: ignjat63 at: 06-06-2006, 18:54	No matter how Lyonnais does the projection Serbia always comes last. Maybe Lyonnais is on to something Seriously now, with a competition every 4 years how can you hope to get some reliable results?

Author: badgerboy
Date: 06-06-2006, 19:01

Yes, I saw Lyonnais stats on the other forum. There were some amusing results. I guess some of the match results used were either (a) very old or (b) in "competitive" tournaments in name only - like the Federations Cup etc.

Going back to your earlier point. Yes it's true some teams do tend to perform better (or at least consistently better) in World Cups than in the European championships. I posted some stats on the same thread you mentioned (on the other forum) as well as on the predictions thread on here. More like the "usual suspects" at the top. I didn't do those "weighted" yet.

I still think the most interesting thing I discovered while collating the information was the performance of European debutants in Europe.

"1958 in Sweden - Soviet Union (QF), N.Ireland (QF), Wales (QF)
1966 in England - Portugal (3rd place)
1974 in West Germany - East Germany (beat the eventual winners in reaching the final group phase)
1982 in Spain - No European debutants
1990 in Italy - Ireland (QF)
1998 in France - Croatia (3rd place)"

Looking good for Ukraine on paper... (although for some reason I don't see them getting past the group stage)...

Author: panda
Date: 06-06-2006, 19:09

Yes, there are a lot of variables:

1) the strength of groups is very divergent
2) there are 6 debutants, therefore many matches between countries have never happened in WC finals.
3) as ignjat63 says, WC is 4 yrs apart; EC does not necessarily tell you how a country wil perform at WC
4) OK, you are strong team and win your group, but the draw is already made, you could be meeting brazil in QF

If someone were masochist out there, they could take some euro club teams and do the rankings based not on the last 5 yrs, but THE LAST 5 WC YEARS. Take the co-eff for 2005-6, 2001-2, 1997-8, 1993-4, 1989-90. Omit the years in between. Then rank the teams- that should give you anindication of how reliable a stats projection from past WC performance is. (or I suppose one could use 98, 00, 02, 04, 06)

see how different that is from 01-06. That factor of difference is also (rule of thumb) how unreliable a projection is from stats alone.

Author: badgerboy
Date: 06-06-2006, 20:05

I still think there's a conflict here.

Yes I agree with you that going back five World Cups or even five International tournaments (10 years) seems too far. BUT people actually use performance from back that far and beyond when coming up with the list of "favourites" for every World Cup and more often than not this list of "favourites" is somewhere near the truth (more so perhaps at World Cup time than at European Championship time).

Historically, the "same old teams" keep on succeeding at World Cup time. Or at least (again) the same old teams are always expected to succeed and one or two of them (but not all) usually do. But their favouritism (if based on anything at all) seems to be based on the fact they won a World Cup (or reached a final or semi-final) anything from 8 to 20 years ago. The fact they often "justify" this favouritism (by this I mean at least reach the quarters and/or only lose to another team that's won it before) means I must be missing something.

Perhaps I just have the same blinkered view when it comes to the performances of certain national teams (Germany, Italy, Argentina spring to mind) as I think some people have (STK springs to mind - no offence STK) when it comes to "big countries" in UEFA.

Panda - or anyone else? Did you see the programme on Germany in the World Cup Stories series. In 1954 they beat Hungary (much the better footballing side) based on their grit and determination. Oh and the Hungarians missing a hatful of chances and then the referee disallowing a late Hungarian goal were an enormous help. Twenty years later it was Holland who played the better football. German "grit and determination" again won the day oh and an English referee who didn't realise Holzenbein had dived in the penalty area (and I thought Klinsmann started that game) and let Germany back into the game. Again in the second half the other team just couldn't put the ball in the net.

Next week the programme is on Italy. I can't wait! Then again I do remember actually cheering on Italy in 1982. They actually won playing great football (as did Argentina in 1986) and even today if they can manage to do that again I think I could bring myself to applaud it! But if a team is to win defending stoutly and withstanding the better footballing side PLEASE let it be a team like Serbia... Better still let it be the Czechs, or the Dutch or the Spanish or the Portuguese (or the English - I almost forgot to add1) out-Braziling the Brazilians. Sorry I've gone a little off of statistical analysis here but I'm not asking too much - am I?

Re: World Cup Statistics - best coefficient method
Author: panda Date: 07-06-2006, 12:56	I saw the German programme, also the France one. I conclude from them: Winning the world cup is SO BIG it has a significant effect on the whole country - so in a way it does not amtter how you do it. The benefit to the morale of the winner's country is 1000x more significant than the aesthetic pleasure of the neutral spectators.

Re: World Cup Statistics - best coefficient method?
Author: badgerboy Date: 07-06-2006, 13:58	Fair point I suppose. All I will say is that it would have been just as nice (or maybe even nicer?) and possibly more justified for the morale of the Hungarian people to have been boosted in 1954 and the Dutch in 1974

Re: World Cup Statistics - best coefficient method
Author: panda Date: 07-06-2006, 14:02	My statistical knowledge is nil; therefore I don't know how many WC finals one would need to get a sample where the number of examples was statistically significant. (Even completely discounting any problems from the long time gap and the unknown quantity of playing teams one has never encountered before). maybe there are no stats as such, just a record of what happened.

Re: World Cup Statistics - best coefficient method?
Author: ignjat63 Date: 07-06-2006, 14:18	And even boosted moral does not mean anything if you let it go by - do not do anything worthy with all that adrenalin, moral and enthusiasm. If Serbia wins WC would that really bring significant long term changes internally? I doubt it.

Re: World Cup Statistics - best coefficient method?
Author: exile Date: 07-06-2006, 14:52	Winning World Cups has had political effects in many countries - and so has losing! Winning tends to boost the current government and increase nationalist feeling, losing does the opposite. There are exceptions. The USA winning the world cup would have a net effect of zero outside the sport itself.

Author: Lyonnais
Date: 07-06-2006, 15:07

who knows ?
at least it would have a strong effect for the promotion of soccer (which currently tends to be seen as a girlish sport).

Coming back to the stats, I strongly doubt that one is able to find any statistical method that would enable to make projections. Football is not a perfect science and that's the reason why we like it, I guess.
Even for a domestic league where you have a lot of recent data, and where the champion is supposed to be the best team, I hardly believe that you can find such kind of method.
So for a cup, it is even more challenging.

ignjat63 > I sincerely hope the best for Serbia and I would be delighted if they could go through and advance as much as possible. Not only because I like Serbia, but also because it would mean that at least one of 2 hot favourites in your group would be sent back home.
But as explained, it is just a simulation based on the latest head-to-head results and it's not my fault if Serbia lost 6-1 to Netherlands in the Euro 2000. The positive thing is that you are almost 100% sure that your country will do better this time.

Re: World Cup Statistics - best coefficient method?
Author: ignjat63 Date: 07-06-2006, 15:11	Lyonnais, I know, I was joking up there. And of course I am sure we'll go through the GS. I am stupid and blind as the next fan.

Re: World Cup Statistics - best coefficient method?
Author: panda Date: 08-06-2006, 10:51	@ignjat63 To be honest, after looking in detail through the groups as a last pre-WC briefing, I think there are relatively few teams (though Trinidad and Tobago is one) where I really think they can't get through out of GS. SCG has the hardest group, but to be above Spain in QG can't be bad.....

Author: Lyonnais
Date: 08-06-2006, 18:12

Some stats that I find interesting.

Confidence index

(results of the 5 last games regardless the opponent)

15 pts - ENGLAND
13 pts - Australia, Brazil, Iran, Portugal & Czech Rep.
11 pts - Spain & Ukraine
10 pts - Germany, France, Netherlands & Serbia
9 pts - Italy, Mexico & Poland
8 pts - Korea, Japan, Switzerland
7 pts - Croatia, United States & Ghana
6 pts - Angola, Argentina, Cote d'Ivoire & Paraguay
5 pts - Tunisia
4 pts - Ecuador, Sweden & Trinidad
3 pts - Saudi Arabia, Costa Rica & Togo

Performance index

Last 5 games vs. another WC participant

13 pts - PORTUGAL
11 pts - France & Spain
10 pts - England, Neterlands & Ukraine
9 pts - Korea, United States, Italie, Poland & Czech Rep.
8 pts - Brazil
7 pts - Iran & Japan
6 pts - Germany, Ghana, Mexico, Paraguay, Sweden & Tunisia
5 pts - Cote d'Ivoire
4 pts - Saudi Arabia, Australia, Croatia, Ecuador, Switzerland & Serbia
3 pts - Angola, Argentina, Costa Rica, Togo & Trinidad

Re: World Cup Statistics - best coefficient method
Author: panda Date: 08-06-2006, 18:20	@Lyonnais Did you count Brazil v Lucerne? (see also my stats post running on the WC prediction thread, which gives France 2nd best chance)

Re: World Cup Statistics - best coefficient method?
Author: Lyonnais Date: 08-06-2006, 18:42	panda, I just took into account official games as listed in FIFA website. I am currently back-testing this with 2002 just to see if it makes some sense or not. I'll see your predictions as soon as this back-testing is finished.

Re: World Cup Statistics - best coefficient method?
Author: panda Date: 08-06-2006, 19:00	Lyonnais- sadly as a sign of my math incompetence, and forum laziness, they are not my stats, but if you follow the link, your comments on the system used will be greatly valued by me.

Author: Lyonnais
Date: 08-06-2006, 19:06

Just coming back to 2002:

Confidence index (results of the last 5 games): BRAZIL was the best with 13 pts ahead of Denmark & Mexico (12), South Africa (11), Spain, France & Senegal (10)
Germany had 9 pts, South Korea 8 and Turkey 7.

Performance index: ITALY was leading with 11 pts ahead of Spain, France & Brazil (10), Saudi Arabia, Argentina, Denmark, United States & Ireland (9)
South Korea had 6 pts, Germany 3 (!) and Turkey 1 (!!)

Conclusion: Brazil was rated high in both tables, but none of the other semi-finalist were that much rated before the competition. Furthermore, France was rated high too and we know what happened.

If we take the performance index, we would have found 11 of the last 16 teams (69%).
So if you want to have an accurate list for 2006 please remove 5 teams of this following list (based on the confidence index):
Germany-Poland,
England-Paraguay,
Netherlands-Serbia,
Portugal-Iran,
Czech Rep.-Italy,
Brazil-Australia,
France-Switzerland (or Korea),
Spain-Ukraine.

Author: panda
Date: 08-06-2006, 19:14

@Lyonnais- your '16 confidents' is:

Germany-Poland,
England-Paraguay,
Netherlands-Serbia,
Portugal-Iran,
Czech Rep.-Italy,
Brazil-Australia,
France-Switzerland (or Korea),
Spain-Ukraine.

OK- if I swap the 'generally predicted' for the not

IN= Sweden, Argentina, Mexico, Croatia
OUT = Paraguay, Serbia (sorry ignjat63), Iran, Australia

and that is only 4. So a 'no shocks' result would be quite a good prediction for the confidence index.

It's a great pastime, this; because one of the great things in WC is the unexpected performance of a team, often that few people know about; so for maximum pleasure, our predictions must be almost right, but make one or two big omissions. (Often African teams, past performances by Cameroon, Nigeria and Senegal, and yes, who would have thought Turkey, Croatia, S Korea (but Korea at home with dubious ref decisions)

Author: ignjat63
Date: 08-06-2006, 19:30

"OUT = Paraguay, Serbia (sorry ignjat63), Iran, Australia"

Do not worry, I am not going to kill myself. Or have a breakdown or heart attack.

I'll just drown my sorrow in chocolade with almonds.

But I meant to say, do not be overly surprised if Croatia is a positive surprise again. They have a bunch of very good players, they are strong as a team, have great fighting spirit, and also great psychology. Despite the war and all I wish them luck.

Author: badgerboy
Date: 08-06-2006, 19:53

I don't have any stats (and I think I'm too lazy to do any before the games kick off now) but it would be interesting to see some along the lines of those prepared by Lyonnais. Maybe last five "competitive" matches between the top 10 or so teams.

For example Germany getting to the final in 2002 was seen as a surprise but they did so without beating any of the traditional footballing powers. To me that means Germany is overrated (people tend to remember the stage a team reaches - which is after all want counts - rather than who they beat along the way). But then there isn't any REAL evidence to say Germany hasn't improved greatly - at least since 2004 - because - to me - friendlies count for very little.

I think the problem (which Panda mentioned earlier) is that there are actually very few matches within a reasonable period of time on which to base any evaluation.

Author: MalcolmW
Date: 08-06-2006, 22:07

Panda

Don't write Australia off so easily.
They may have the best 'keeper in the competition, and they are certainly a team with Attitude! Did you see the wild colonial boys in their 'friendly' v Holland? Only 3 bad injuries for the Dutchmen. Brazil's team selection v Oz could be very interesting. Some players simply won't fancy it, and it may simply be a question of whether it's the first strings or those in reserve who face a battle. The Croatia match is interesting in its own right, and it could decide who progresses. If so the French could be hampered for the QFs....

Author: Lyonnais
Date: 09-06-2006, 04:16

fully agree with ignjat63 about Croatia and for Serbia too. These 2 teams are strong enough to get through, no doubt about that. They just are simply unlucky enough to have tough groups, both of them, but I wouldn't be surprised if both of them went through.

Argentina has been disappointing for a while now and Cote d'Ivoire might lack of experience. Serbia have their chances, sure.

In the other group, Croatia, Japan and Australia have very different styles. I have the feeling that Brazil might struggle much more than expected in this group. They will hate the physical challenge imposed by Australia, and Croatia is the technical but also strong enough to annoy them. And Brazil looks like so confident that it tends to turn into over-confidence (some might argue that this is always the case and that didn't prevent them to reach the final 3 times in a row, but this year it's more than usual).

We'll see, I have never been good in predictions anyway.