World Cup 2014 - Group G: Preview & Analysis

The SPI is a joke and not sure why it continues to be referenced. Anyone with knowledge of the game knows that Portugal's form in qualification is irrelevant when tournaments come around.

http://espn.go.com/sports/soccer/bl...7063/euro-2012-breaking-groups-poland-ukraine

Taking a look at their "predictions" for Euro 2012 and who'd advance from respective groups

Group A: Got both wrong.
Group B: Got 1 wrong.
Group C: Got 1 wrong (SPI had 49% for Croatia, and 36% for future finalist Italy uh...really?)
Group D: The don't mention each team's percentage but they probably got that wrong too.

http://espnfc.com/world-cup/story/_...aw-goes-us-not-spi-powerhouses?cc=5901&ver=us
World Cup 2010:

Group A: 1 wrong
Group B: 1 wrong
Group C: Both right but in the wrong order.
Group D: 1 wrong
Group E: 1 wrong
Group F: 1 wrong
Group G: 1 wrong
Group H: Both right.

For a statistical index, those are some very, very awful prediction stats wouldn't you say?

 

And only in la-la land can anyone actually believe Portugal and the USA are both even on chances of advancing from that group. Having Ronaldo alone should be a 10/15% boost.

 

That's actually pretty good. Better than the average prediction for sure. He correctly picked 10/16 to advance. The average in the yahoo pick'em was around 8/16 if I remember correctly.

Picking who will advance isn't about picking the top two teams in the group. The top two teams in this group are obviously Portugal and Germany. Statistical analysis is much more nuanced than that. I already listed why Portugal's chance of advancing is low.

1. They play Germany first. Playing the hardest team in the group first makes you less likely to advance
2. Weather/conditions favor Ghana and the USA, who are both used to playing in tropical humid locations.
3. They play the USA in Manaus, where the temperature at game time with be 30 degrees and it will feel like 44 degrees with the near 85% relative humidity. This is a huge advantage for the USA.
4. They'll likely have long term fatigue problems, especially their star players like Ronaldo. Most American players either play in the Bundesliga or in MLS, where there's a sizable winter break.

These are the factors you have to consider, in addition to overall talent and ability. No one's saying the USA or Ghana are anywhere near as good as Portugal, however the conditions and circumstances give the former a leg up.

 

The Netherlands is at a huge disadvantage. Playing the best team in the group first significantly decreases your chances of advancing. Chile, meanwhile, has basically a guaranteed 3 points in their first game.

Also, that group will almost certainly come down to the Chile vs Netherlands game on the last day, and I think Chile will be the favorite in that one. It's in San Paulo, and Chile will basically be the home team.

 

And on what data have they based their coefficients for fatigue and climate? How much is this leg up?

Oh that's right as there isn't really the significant historical data to go on, it's just the opinion of the designer of the score.

It looks like South American teams are given a large advantage from looking at the probabilities. Again what is this based on?

 

No, not if they are independent trials. Very basic probability theory.

If they are modelled as dependent, how are they modelled? What basis is that adjustment made on? What data is there to actually create any good coefficient of adjustment.

There is massive subjectivity going into these probabilities on factors for which there is no data to go on.

 

They're not independent trials. The results of the second games are influenced by the results of the first games. Each game isn't an independent event. Analyzing it as such would be illogical.

I don't know for sure how he did these calculations, and you're right to be skeptical. I guess we'll have to wait and see if the methodology is released or not.

 

Because this is the only national team that plays its continental tournament every two years as the host. The CONCACAF Gold cup since 1991 has taken place nearly exclusively on U.S soil. Even in Africa, you never see twice the same host in a row.
I think it gives them a boost that other teams can't get (including in the FIFA rankings, which are crucial for seeding, since the importance of the matches in this tournament is equivalent to a confederation-level final competition/ Confederation Cup, just like the Euro, AfCoN, Asian Cup, Copa América) and sometimes make them look better than they are which might have played in the OP biased analysis.

This year the USMNT had a 12-game winning streak, streak that is partly due to improvement but that was also greatly help by the Gold cup at home, there's no doubt about that. You just have to look at their record on the road, notably in Central America during the 2014 WCQ to see it. A draw in Guatemala, losses in Honduras & Costa Rica and a difficult win in Panama (thanks to their "superior" fitness allegedly).
It also advantages them greatly against CAF, AFC ( & other CONCACAF) teams since these teams often have to play away their continental tournament and have to deal with a lower confederation weighting .

 

I don't know because the methodology wasn't released. I'm just going off quotes from Nate Silver.

In general though, there's lots of information regarding how climate effects the game. Teams run as much as 25 km less than they normally would depending on climate. That changes the tactical nature of the game.

 

It makes sense to think games are dependent. The third game for a team that has already won 2 games may be one where they rest players. However, how do you make that judgement in a model? Can you prove there is a genuine dependency for the outcome? Even taking 10 world cups with 8 groups, that's only 160 final group stage games. How do you control for other factors and still come up with statistical significance that there is a dependency. I don't see it happening.

Can Nate prove climate is a usable factor and say for sure his coefficient is correct as a point estimate with low variance. Is there an interaction with other factors? What sensitivity does his results have to these factors?

Overall what I'm saying is that there is going to be a large confidence interval around his results.

I work in an industry where n=10,000,000+ and building generalised linear models is a pain. How he's modelling this is beyond me.

 

That's actually pretty awful, particularly for Euro 2012. He was wrong about at least half of his predictions. That's terrible.

In regards to your points:
1. I'm not saying that's untrue but I'd like to see the stats behind that. Portugal had no issue advancing after playing Germany first last year.
2. The weather conditions are becoming over-exaggerated. Portugal may not be as humid, but it's of the hottest countries in Europe. Ronaldo grew up on an island FFS. Virtually our whole team has experience, because of tournaments like the Champions League, of playing everywhere, from Moscow in the winter to Lisbon in the summer.
3. Again, not that big of an advantage.
4. Our players had no fitness issues last summer or any other tournament that I can remember recently. Ronaldo is also arguably the most fit player in the world so I would especially not worry about him.

 

I believe betting exchanges are relatively efficient these days and can't be manipulated. Money also speaks louder than words.

Group G implied probabilities to qualify (from Betfair) are

Germany 1/7 = 88%
Portugal 4/9 = 69%
USA 9/4 = 31%
Ghana 3/1 = 25%

Probabilities add up to greater than 200% because of margin. If you believe Nate to be correct then you short Portugal and go long on the USA.

 

I'd say the Germany/Portugal stats are accurate but disagree with the other two. I think the US is undoubtedly the weakest team in the group and its unfortunate for them as they'd have a good chance in another group. I think people are forgetting that Ghana has a squad that includes the likes of the Ayew brothers, Gyan, Boateng, Asamoah, etc. The US simply does not boast that type of talent.

 

Lisbon really isn't that hot in the summer. If you look at humidex values in July, the average mid day value is around 31. That's less than New York City at 36. The US national team regularly trains in Miami in humidex conditions of 43. Manaus should be around 44 at game time. That's hot. I'm not sure there are many European players used to those types of conditions.

Talent doesn't win games, team play does.

Besides, this Ghana team is aging. Essien has played two league games for Chelsea this year. Boateng is chronically injured and can barely even train at Schalke (trust me, this one's a weekly headache for me). Sulley Muntari is arguably the weakest first 11 player on a middling AC Milan team. Gyan plays in the UAE league. Kwadwo Asamoah and André Ayew are the only Ghanain players that are in legitimately good form on notable teams.

Don't get me wrong, they have some star power and a better first 11 than the USA on talent alone, however the talent needs to be assembled into a competitive team. They haven't been able to do that on a regular basis recently, which is why the rankings are down on them.

 

http://weatherspark.com/averages/31081/New-York-United-States
http://weatherspark.com/averages/32376/Lisbon-Portugal

Not sure where you got your info. but that site has Lisbon at higher temperatures. I'm not discounting that the US players may be more accustomed to that weather, which will be an advantage. I just don't think it'll be that big, particularly since US players, like Altidore, Jones, etc. are currently living and playing in European countries, meaning they'll more likely than not will need to get used to that humidity again.

In terms of aging, I don't really think so. Essien only recently turned 31. Also, the flip side of that coin is that these guys can bring experience, experience that nobody in the US has. Having guys like Essien and Muntari, who've won the Champions League, will give a leg up. I'll take your word on Boateng since you're a Schalke fan but in terms of Gyan I still think he's arguably their best player, regardless of him playing in the UAE. I think the fact that he still starts for Ghana is indicative that he hasn't lost his quality.

Let's not forget, this team absolutely dismantled Bob Bradley's Egypt team. If you didn't know, Egypt was the only team in Africa to win all their group games in qualifying. I think that may be telling.

 

That's too simplistic.
So why has USA not come close to winning a WC yet whilst Maradona's individual brilliance gave Argentina their cup in '86?
These days teams need both. You have to have that individual brilliance sometimes to unlock those difficult games. USA does not have that.

 

Yes, most players that were 20-25 in 2010 will be 24-29 in 2014...

Unless some odd selections are made for the World Cup squad you will have one player over the age of 30 and that player is Essien. The squad will also likely feature 8-9 players over the age of 25, 2 of them are goalkeepers. The central defenders that were 20-23 years old in 2010 will now be 24-27. I don't see how that can be a minus, given the players in question are regulars in their Ligue 1 clubs and Jonathan Mensah has been in beast mode, whenever played as a DC.

Sort of an idiotic argument. Ghana will likely have one of the younger squads at the tournament. This isn't Cote d'Ivoire, where their star player is past the hill... This squad is largely based on the U20 Youth World Cup squad of 2009 with some star players making up the rest of the squad.

Essien didn't even play in 2010. Essien back in the side is a significant strengthening, as the player he replaced in the midfield is Samuel Inkoom, a right back pretending to be a right midfielder in 2010. Essien brings experience and leadership in an otherwise very youthful outfit (that's why the 'Ghana is aging' comment is pretty ridiculous).

BTW: Small difference, but Essien has started 2 league games, but he's featured in three. And you should focus on the games vs Arsenal and Egypt to see what he brings to the table. Even today he was very good for Chelsea, although he doesn't seem to communicate well with Ramires.

Muntari was a bench player in 2010 and until proven otherwise is a bench player now. You may not like Muntari, but he's a big game player, an impact player and he has as much fans as foes at AC Milan. And he is far from the weakest player in the AC Milan team. And BTW they said the same thing about him in 2010 when he was at Inter (worst first team player, waste of space).

Boateng has played 10 minutes for Ghana since 2010. I'm not sure he's even starting for Ghana. Ghana dismantled Egypt 6:1 without Boateng and with a makeshift defence.

Gyan plays in the UAE league by choice (for money). That doesn't magically make him worse. He has a 1,17 goal per goal average, so he's clearly way to good for his league. And he just had a record 9 goals for Ghana in 10 games this year.

Why? Because they couldn't win a Nations Cup in a tournament which is way more competitive than a Gold Cup? Well... they didn't win a Nations Cup in 2006 or 2010, but still advanced from respective Groups of Deaths at those competitions.

Fact is that in 2010 Ghana had a pretty weak squad compared to 2006 and compared to 2014, but they were still just a handball and penalty away from the semis of the World Cup (whilst playing without Andre Ayew, their best player at the tournament). Why? Because their tactically aware, work as a team and are much more than a sum of their parts.

 

I got it from Wikipedia. And NY has slightly higher humidity, which is why their humidex value is higher. And while it's true that American players in Europe aren't exposed to the heat, they still train with the US every summer in the heat.

Maybe it's just me, but I don't rate CAF teams very highly. They always have the talent, but they always under perform. Ghana absolutely has the talent to go gangbusters at the World Cup, but I doubt they actually will. I also don't think Egypt is really that great of a team. Maybe it's because the last time the US played them in an official match we won 3-0.

Sorry, when I said aging, I should have said slowing down. A lot of the Ghanaian players in the 26-29 age range seem like they have slowed down significantly, for whatever reason. Many of the players on that roster seem to have peaked in their early 20s. This is subjective, so it's just an opinion based on observations.

 

Ok. Let's look at the Ghana 26-29 old players that will likely feature in 2014:

I got your take on Gyan (starter), Essien (starter), Boateng (who knows?) and Muntari (bench).The other players from the 26-29 age range that could make the 2014 selection are: two goalies, John Boye (first choice central defender at Rennes), Harrison Afful (full back, didn't feature in 2010 and can't be any slower than Hans Sarpei who played at LB in 2010) + possibly Jerry Akaminko (central defender, didn't feature in 2010) and Albert Adomah (right wing, known speedster, didn't feature in 2010)

So basically you're entire argument that Ghana has slowed down since 2010 is based on Essien, Muntari (if he even features) and supposedly Gyan being slower...

Meanwhile the first team for Ghana feature:
- Waris - speedster (now behind Gyan)
- Opare - roadrunner (replaced static Paintsil)
- Ayew - workhorse running up and down the field (replaced quasi-midfielder Inkoom at RW, while Asamoah slotted in at LW)
- Addy or Afful - first one is a jackrabbit, second one is miles faster than Sarpei from 2010
- Rabiu (about as fast as Annan, who played DM in 2010)

I agree that Essien is slower now than Boateng, who played the CM role back in 2010, but lets not exaggerate. I agree the middle of the field is somewhat less dynamic. But that's why Ghana uses its wings now - for one thing the wings have never been so fast...

 

I never said Ghana, as a team, is slow. I'm just saying that some of their experienced stars in that age range have slowed down and aren't the players they used to be. Ghana's strength is definitely on the perimeter with the younger players and not in the middle of the field. I don't know how things will play out, only time will tell.

One thing I don't get is how Ghana expect to go far when their left back is so Afful.

 

Ghana's LB position has always been awful. Afful is actually an upgrade. Still hope David Addy starts - don't know why the coach doesn't just let him claim the spot? He is one of the better LBs in Portugal. That must count for something...

Or maybe Gideon Baah (guy actually won a future football star reality show back in Ghana)? Best defender in Finland. Like a bulldozer on the left. And this guy is faaasssttt! Plus he can dribble. And he can score an overhead kick goal

 

But you're leaving out an important factor, Germany has advanced out of the group stage in every single world cup they've played in dating back to the Munich Crisis in 1938, that was the last time they crashed out in the initial stage, so the 90% advance angle is something that would be true for them in any group in any World Cup tourney period. They've earned that respect. It shouldn't automatically negate the possibility of a German Seed Group being a group of death, but by your standard it would. Seems bizarre to me.

I also view groups of death as groups featuring either 3 top 10 teams, or 4 teams considered legit knockout round favorites (top 16 teams in the cup), I think B and G encompass these features, but not Uruguay's group, as England and Costa Rica are a bit too weak to me to be considered that. The 3 World Cup Champion holders argument is a soft one considering two of the three teams haven't won a World Cup in nearly 50 years in one case, and in more than 50 in the other.

 

The '06 loss is bogus, third place games are basically scrimmages.

 

No, because even Germany in a group with a side which is rated as potentially upsetting them and with others who are capable of taking points from them, would not enjoy the kind of odds its history might suggest. For instance, if Germany was placed in a group with Holland and Chile along with Australia, instead of Spain being placed in that group. Germany's chances would then drop significantly regardless of the history. It would drop even further if you had that same group with the Germans, Holland and Chile, but the US instead of Australia.

 

Actually, the worse group for the Germans would be if they got the Italians! Say if they ended up in a group with Germany, Italy, Chile and Japan. Now, that would be a group of death.