Introducing Goalimpact: Cameroon’s World Cup XI
With the World Cup approaching, KingFut have decided to give a closer look into the five African teams heading to Brazil. We start with the Indomitable Lions, using Jorg Seidel’s Goalimpact algorithm to see how they should line-up, and how they compare to the other 31 teams.
The most successful African side in World Cup history. Having qualified 7 times, the Indomitable Lions are perhaps most remembered for that infamous Roger Milla-led run in the 1990 World Cup, becoming the first African nation to reach the quarter-finals and forever changing the perception of African teams by stunning defending champions Argentina in their first game. If you take away that 1990 run, however, Cameroon’s only other World Cup match win came in 2002 against Saudi Arabia! Despite being perennial African contenders (with four titles to their name, tied with Ghana for 2nd), they have disappointed often at the world stage since 1990, recently going out with 3 straight defeats in 2010 to Japan, Denmark & the Netherlands.
In fact, Cameroon’s most famous World Cup moment is probably this iconic dance by Mr. Milla:
Current Profile/Recent History:
Volker Finke is their coach, having taken over from caretaker Jean-Paul Akono. The former Freibug coach has done okay, despite struggling in the group (with FIFA giving them a huge helping hand by overturning a defeat into a win), they qualified after handily beating Tunisia.
They were runners up to Egypt in the 2008 African Cup of Nations, and made the quarter-finals in 2010. Since then however, they have not qualified for both 2012 & 2013 versions of the tournament (similar to Egypt).
Despite their squad not quite being at the levels of the early 2000s (when they won 2 African Cup of Nations in a row) with Rigobert Song (Alex Song’s uncle!) and a young Eto’o, the level of talent is still pretty high. A much more senile Eto’o is joined by younger talent all over Europe, from Alex Song at Barcelona, the impressive Joel Matip at Schalke, to Assou-Ekotto at QPR & Stephane Mbia at Sevilla. So do the Indomitable Lions have one more run in them? Or is this squad too low on talent to be able to make a splash in Brazil?
How good are Cameroon, and how to find out:
When trying to find out the level of performance of teams, fans and scouts often use one of two methods:
1) The eye-test:
Watch a team and assess how good they are.
2) Bottom-up player contributions:
Go through the line-up and add up all the players’ levels to reach a conclusion about the team.
These two ways are often combined to try and predict the level the team is at, or their level of performance. Assessing player talent level is done by a mixture of eye-test (i.e Messi is good at football) to numeric value (i.e Xavi has a pass completion of x%/ Suarez scored x amount of goals). Despite individual statistics advancing and evolving, their drawbacks are still crystal clear. Football is such an interconnected game, that it is almost impossible to define a set of individual actions that accurately convey a player’s level. The first issue is one of positions, how to cross-compare between a good defender and a good forward for example? Which is more valuable to the team? Even within the same position, comparing Xavi’s passing numbers in a dominant Barcelona side, who like to dictate possession, with Xabi Alonso’s in a more reactive Madrid side seems to be unfair to Xabi Alonso and misses the point.
What is the point? Goals. Goals are the only statistic in football that eventually end up determining who the winner is. Viewed through that lens, shouldn’t players ‘ratings’ be assigned on who helps his team not only score the most but have the highest goal difference? So instead of using bottom-up models (from players to team), we decided to enlist the help our friend Jorg Seidel, creator of the most advanced top-down model (from team to players) we’ve seen.
Goalimpact, Jorg’s algorithm, “measures the extent to which a player contributes to the goal difference per minute of a team,” according to his explanation.
What does this mean broken down a little more? It means for every player, Goalimpact takes into account the number of minutes you’ve been on the pitch, and you’re team’s goal difference when you’re on the pitch. So if you play 55 minutes and leave when your team is 1-0 up, then watch your team lose, your goal difference is +1 (you’re not penalized for losing the game since you weren’t on the pitch).
So in a way its similar to the NBA’s plus/minus ratings, but a pure plus/minus doesn’t work, and here’s why:
Let’s say Barcelona pick me up tomorrow to play centre-back (they need me..). I suit up, get on the pitch, and Messi scores 5 past whatever team we’re playing while I sit at the back drink some nice iced tea with Barcelona dominating. Our plus/minus would both be +5, since Goalimpact only cares about Goal Difference. So basically, I’m as good as Messi if you use just raw plus/minus then.
So how does Goalimpact fix that? It adjusts!
1) It adjusts for team-mate strength (so if I play with Messi, my score reflects that)
2) It adjusts for opponent strength (5 goals against El-Gouna vs. 5 goals against Arsenal)
3) Home advantage
4) Red cards, fatigue, etc
So it’s a dynamic rating system based on all of these factors, using goal difference as the only underlying statistic. It’s a beautifully crafted top-down system that, while not without its flaws, gives us a results only rating that can be compared cross-team, cross-position, and across time.
You ask any random fan for the top 5 players, you get a list of the top 5 attacking players in the world right now. The beautiful thing about Goalimpact (in my eyes) is that since it’s result-oriented only, it doesn’t discard defenders, goalkeepers, and midfielders. Football is a team game, and Messi + CR7 can’t score all these goals without the people behind them doing the dirty work!
In the Goalimpact top 50 (February ’14 edition), Lahm is at 2(!!!), Dani Alves is at 7, Casillas and Neuer are at 9 & 10 respectively. There are more defenders than strikers! It’s refreshing to see such a top 10 in today’s football, where goals & attacking have become overvalued and are the only thing worth mentioning (ironic since Goalimpact measures goal difference)
Here’s an example of Goalimpact at work:
Thomas Muller has obviously made Bayern a lot better in these matches, therefore his Goalimpact score goes up. This is “Goalimpact at it’s purest form,” according to their website.
Assuming the adjusting is done well (which is a very big and mathematically challenging assumption to get into) the one thing that sticks out for Goalimpact as a weakness for me is its lack of movement or robustness in defining form. Goalimpact is a player’s playing level over the past couple of years, it does not take accurately represent the player’s current form. Another small problem is col-linearity (when two players always play together–> Vidic & Ferdinand ’08 for example). When two players are direct replacements for each other or always play together, it is often difficult
A major strength of Goalimpact, however, (and one we’ll discuss further in other articles) is its ability to tell good players from a very young age. Goalimpact compiled a list of young players in 2007 that it would’ve recommended, and on that list were some gems that were not known to traditional lists (such as transferkmarket) such as Subotic, the Bender brothers, Rakitic (when he was at Basel!), etc. They were put in along with traditional youngsters such as Bale, Walcott, Pato, etc.
To make the rating more easy for normal fans, Jorg has converted the Goalimpact score to a scale of 100. The average player in his database has a score of 100. Anything above 150 is ‘world-class’. Above 120 is ‘Champions League-class’. Here’s a detailed post talking about Goalimpact distribution.
(P.S the top player is Cristiano Ronaldo at 195. Other notables: Messi (175), Zlatan (164), Alonso (164))
So how does Cameroon stack up?
Here’s the optimal Cameroon XI according to GI:
At first glance, it seems that Goalimpact matches up pretty well with people’s general ideas about Cameroon. Cameroon’s two ‘CL-Class’ show up with Alex Song having a rating of 141.3 (in the top 50) and Samuel Eto’o scoring 126 (which is ‘CL-calibre’). The Eto’o rating seems about right, as he is still a CL calibre player but on the downwards (his peak was 154, world-class). The Song rating maybe more controversial, as he has had a couple of bad years at Barcelona and does not seem to be in the top 50 players in the world right now. Despite having a down year at Barcelona, GI really likes Song (who’s W-L record with Barca is quite impressive, but perhaps misleading).
Despite Assembe being the highest rated goalkeeper, it seems that Finke actually prefers Itjande more. Regardless, it does seem a position of weakness for Cameroon. Carlos Kameni of Malaga is also an option.
The big surprise here is the absence of Nicholas N’koulou the young Marseille defender who did not make the GI starting 4. Joel Matip (just 22) takes his place, a player who GI has having a good career with a projected peak of above 120. On balance, I think it’s a good decision to play Matip and not N’koulou, but N’koulou seems to be popular with Finke and the odds are of him starting in the World Cup. The defence is also deep at CB with the GI of subs such as Nounkeu at 108. Other than that, the rest of the three positions are pretty much nailed down. The only other decision to make is whether to play Mbia at RB or in midfield. GI suggests that at RB would be better, since it is a weakness of the team. Marcel Ndjeng used to be a very useful player (132 in his peak) but is now down to 105.
Here’s Stephen Mbia’s career chart according to GI:
As you can tell, his decline has started around 2012, and he should no longer be first choice in midfield for Cameroon. However, he is still useful at RB due to Ndjeng declining more.
The strength of this Cameroon team. Alex Song may have a lot of detractors (perhaps for good reason), but the fact that Barcelona spent 19M Euros on him tells you a lot about his talent level. The interesting part will be looking at his form since he has not played much for Barcelona this year (although he did look impressive in the title decider against Atletico). Song will be paired with Enoh and Makoun, both above average players according to GI, and both favored very much by Finke. Makoun in particular had a peak GI of 132, and although it is down now to 118, his experience may prove crucial to a relatively young Cameroon side.
This is where GI starts to disagree with Finke and reality, a lot. Spearheading the attack obviously falls to Eto’o, rated highly by both GI and his coach. The other 2 forwards however, Oumaro & Priso have not even been included by Finke in the 28-man provisional squad. Moting has a GI of 102 and seems to be favored by Finke, and so will probably start. Interesting to see if Cameroon’s forwards don’t do a good job whether Finke could’ve used the services of these 2 highly rated GI players.
Overall, Cameroon have an average GI of 115 when including Oumaro & Priso, and a peak GI of 127 (showing that a high number of their world class players are past their peak). GoalImpact seems to agree with Finke in defence and midfield, yet not very much in goalkeeping and in attack. Moukandjo (unknown rating) and Moting (102) will probably take the place of Oumaro & Priso, lowering Cameroon’s GI.
Since this is part of GoalImpact’s World Cup preview, all 32 teams GI will be released and previewed on different blogs. Cameroon’s group members have not all been published yet (only Croatia), so we cannot compare the entire group yet. Brazil will most probably top the group, with Croatia heavy favorites for 2nd. Croatia’s GI recommended squad is much higher than Cameroon’s, which confirms their status as favorites to qualify. Here is Mario Maric’s round up on Croatia:
Srna (142), Modric (128, probably should be higher but time at RM hasn’t been fully incorporated), Mandzukic and Rakitic both above Eto’o! In fact, almost all of Croatia’s squad are higher than Cameroon’s. Read the full analysis on Croatia’s squad here.
Compared to the Ivory Coast’s Goalimpact, Cameroon are deeper, yet do not have as many world class stars. Ivory Coast have Yaya Toure (136), Kolo (136), Bony (128) , Kalou (140!), Gervinho (117), etc., all who play at a higher level than most of the Cameroon squad. Both Ivory Coast and Cameroon are hurt a lot by age, with Drogba, Yaya, and others having very high peak GI but have fallen off recently. This coincides with the rise of Ghana & Nigeria as the dominant forces in African football, and this maybe Cameroon’s (and Ivory Coast’s) last chance with this generation to make a run for Africa!
Can Eto’o and Drogba, two of the best players in African history, lead their nation to one last run at glory in the World Cup? Goalimpact thinks Ivory Coast have a better chance, but Eto’o is still the more influential player between them! It also puts Cameroon as heavy underdogs to progress past the group stages. What do our readers think?
Check out all the rest of the teams GI here. Follow GoalImpact on twitter here to ask for your favorite player’s ratings! To understand more about the algorithm, visit www.goalimpact.com and check the FAQ or look at articles Jorg has written.
/ 12 hours ago
Liverpool boss Jurgen Klopp has confirmed that Mohamed Salah has fully recovered from his...
/ 17 hours ago
Ivory Coast’s youngster Jeremie Boga completed his move from London giants Chelsea to Serie...