Anda di halaman 1dari 3

World Cup 2010 · Focus Focus · World Cup 2010

And the World Champion is… 2


Past performance and experience teams had never met in the 16 previous
are indicators of future World Cups. However, even more astonish-
Back by popular demand, we once The moral of this story is that one needs performance ing than this oddity was the fact that,
to be humble about the predictive power of according to their Elo ratings (see page 28),
again open our econometric toolbox
one’s models. Successful forecasting can Unlike in the case of the European Champi- the matchup was between the weakest
and use quantitative models to assess often depend as much on luck as on skill, onship, past performance and experience Brazilian and the second-weakest German
the likely winner of the upcoming which is a lesson that is too often forgotten matter considerably when assessing the fate team ever to participate in a World Cup.
World Cup. However, drawing on our when it comes to quantifying the future. This of a team at the World
should also serve as a warning to our read- Cup. Of the 18 World
past successes and especially our past
ers: take the following with caution and a Cups, five have been
failures, this time we are making a pinch of good humor. After all, football is won by Brazil, four by
more cautious prediction of the tour- only a game (in most countries). Italy, three by Germany,
nament’s outcome. Given our experience, we have decided two each by Argentina
not to go down the perilous path of predict- and by Uruguay, and
ing a likely outcome for the whole World one each by England
Forecasting football: Cup. Rather, we assess the probability of and France. Hence, it
1 more art than science being among the final 16 (i.e., surviving the should not be a surprise
first round), being among the final four and, that picking the winner
During the 2006 World Cup, our model’s of course, winning the whole thing. out of this small group is
predictions were very close to perfect. Our a rather safe bet. More-
forecast champion, Italy, which had only an over, Brazil is the only
outsider status among experts, managed to team that has partici-
Explanatory factors
win its fourth title. In addition, our model pated in all the World
As in our previous studies, we rely exclu- Source: www.dreamstime.com
correctly predicted 50% of the semi-finalists, Cups; Germany and
sively on three factors to estimate the dif-
75% of the final eight and 81% of the final Italy each have missed two. All three teams Eight teams had better Elo ratings than Brazil
ferent winning probabilities:
16. This astonishing result led to a rather were present in the last 12 World Cups. and Germany before the 2002 World Cup
1) past performance;
large amount of press coverage worldwide The number of times those three teams started.
2) whether or not a team is a host nation;
and earned the inventor of the model his 15 appeared in the finals is also impressive: At least one surprise guest
and
minutes of fame on CNN.
3) an objective quantitative measure that
seven times for Brazil and six times each for 4 in semi-finals
We came under tremendous pressure to Germany and Italy. It is even more astonish-
assesses the strength of each team three
use the same model to predict the outcome ing that only six teams reached the last 10 The final four of a World Cup is slightly more
months before the start of the World
of the 2008 European Championship, and World Cup finals: the three aforementioned, open than the finalists. Twenty-three teams
Cup. Socioeconomic factors like popula-
we relented – despite reminders that it has Argentina, France and the Netherlands. have made it to the final four. But here again
tion size or GDP growth have been proven
yielded unpredictable results several times in five teams (Argentina, Brazil, France, Ger-
to have no explanatory power when it
the past. Who can forget the fabulous run 2002 World Cup final many and Italy) dominate – accounting for
comes to forecasting the performance of
of the Danes in 1992 or the unbelievable win
a specific team. 3 (Brazil vs. Germany) was a 53% of all final four places. Nevertheless, it
of the Greeks in 2004? Our forecast winner, premiere, not a classic seems that at every World Cup there is at
the Czech Republic, did not even make it into least one surprise participant in the semi-
the second round. Even worse, although our Especially worth noting in this context is the finals. In 2006, many saw Portugal in this
model correctly predicted 63% of the sec- 2002 World Cup final between Brazil and role, although it had been a semi-finalist in
ond-round teams, it forecast none of the Germany. What might sound like a classic 1966. Here again, 2002 is especially worth
final four. World Cup game was actually a first as both noticing: Two rather unlikely participants (at

26 UBS investor’s guide special edition April 2010 UBS investor’s guide special edition April 2010 27
World Cup 2010 · Focus Focus · World Cup 2010

least at first glance), South Korea and Turkey, the continent advantage. So far, out of the which, taking into account the former Soviet 2010 the weakest World Cup
reached the semi-finals. However, one needs 18 World Cups, 10 took place in Europe, six Union, has almost always made it into the 7 since 1994
to acknowledge that South Korea had accu- in Latin America and two on “neutral final 16 (when it was present). But can we
mulated a great deal of experience (five pre- ground” in the US in 1994, and in Japan and objectively assess the current strength of a The 2010 World Cup, which has an average
vious World Cup appearances), and was also South Korea in 2002. The score for Europe national team? One can, of course, take the Elo score for all teams involved of 1,785, is
one of two host nations (the other being versus Latin America is nine to nine. Latin evaluation of bookmakers as a measure of the weakest World Cup since 1994 (despite
Japan). American teams have won all the World subjective strength. Another alternative the fact that three teams – Brazil, Spain and
Cups held on their turf. European teams would be to consult the official FIFA rankings. the Netherlands – have an Elo score above
Being host is an advantage have won nine out of the 10 held on theirs. What we prefer to do (as we have in our two 2,000. The last World Cup with three such
5 Brazil is the only Latin American team that previous football studies), is to rely on the Elo strong teams was in 1978 (Brazil, Germany
Being the host nation of the World Cup is has won in Europe. It has also won twice on ratings (see box) of the teams in the month and the Netherlands). This World Cup will
obviously an advantage. One-third of all neutral ground. of March preceding the World Cup. also host the strongest Spanish team ever to
World Cups organized so far have been won Since 1950 and the go to a World Cup,
by the host. Twelve hosts (63%) have surprising win by the strongest Eng-
Historically long odds for the
reached the final four and all hosts have 6 strongest team
Uruguay, no team with lish team since
reached the final 16 (i.e., survived the first an Elo rating below 1970, the strongest
round, when the World Cup had less teams All the historical data in the world cannot 1,820 has won the Dutch team since
than the current 32). The biggest exceptions turn an uncompetitive team into a winner. World Cup. As men- 1978, the strongest
so far was Spain in 1982, which though hav- Countries that failed to qualify for the 2010 tioned above, the Brazilian and Ger-
ing a rather strong team did not manage to World Cup include some traditional football weakest team (as man teams since
reach the final four, and the upset of mighty countries like Hungary (two times World Cup measured by its Elo rat- 1998, and an Ital-
host Brazil by tiny Uruguay in the 1950 final. finalist), and Sweden (once World Cup final- ing), to win the World ian team of compa-
In the same vein as the host advantage falls ist and two times third place), or Russia, Cup was Brazil in rable strength with
2002. The strongest Source: www.dreamstime.com
the one that cap-
team to win the World tured the World
Elo Ratings
Cup was Germany in 1974, followed by Brazil Cup in 2006. Hence, almost all historical
Elo ratings were developed by the Hun- improve a team’s Elo ranking much more
in 1962 and Brazil in 1970. Interestingly favorites are traveling to South Africa with
garian-American Physicist Arpad Elo than beating a non-powerhouse like
enough, with the exception of Germany in very strong teams, making it very unlikely
(1903-1992) to measure and rank the Malta or Andorra. Moreover, for example,
1974, the strongest team going into a World that, with the noticeable exception of Spain,
strength of chess players. The ratings winning abroad gains more points than
Cup has never won it. we will see a new, fancy World Cup winner
have been used in other sports like ten- winning at home, winning a World Cup
The most disappointing teams in World in 2010.
nis but especially in football, where they qualifier gets more points than winning a
Cup history, i.e., the ones with very strong
have proven to be a better indicator and friendly, and winning with a score of 5 to
teams that did not even make it into the sec-
forecasting tool for determining the 0 will give more points than winning by 2
ond round were France, Portugal and
objective strength of teams than the to 1. A nice feature of Elo ratings is that
Argentina in 2002, Spain in 1998 and Italy
FIFA ranking system. The Elo method for you can compare the strength of teams
in 1950. The most surprising teams in World
ranking football teams not only takes across times. For example, the Brazilian
Cup history, i.e., the ones with rather low Elo
into account the number of wins, losses team, which won the World Cup in 2002,
ratings that made it into the semi-finals,
and draws of each team, it also exam- was significantly weaker than the one,
were South Korea and Turkey in 2002
ines the conditions under which those which won it in 1962. You can find the rat-
(already mentioned above), Uruguay in 1950
events occurred. As a result, beating a ings at
and Argentina in 1990.
strong team like Brazil or Spain will www.eloratings.net.

28 UBS investor’s guide special edition April 2010 UBS investor’s guide special edition April 2010 29
World Cup 2010 · Focus Focus · World Cup 2010

Round of 16: African teams Out of the top eight contenders to win
8 – except South Africa – cannot the World Cup, six are former winners and
exploit host advantage the top three are the usual suspects. This
Table 1: Round of 16 Table 2: Semis
might sound boring but, as stated above,
Country Likelihood to reach the round of 16 Country Likelihood to reach the semi-finals
Looking now at the likely outcome of the past performance is a very strong predictor
World Cup according to our model, tables 1, when it comes to forecasting the likely win-
South Africa 78% Brazil 49% 2 and 3 list the 20, 12 and 8 most likely ner of a World Cup.
Brazil 74% Germany 38% teams to emerge from the round of 16,
Spain 73% Netherlands 34% reach the semi-finals and win the World Cup. And the winner of the 2010 World
Netherlands 68%
Italy 32% Despite starting in its own World Cup with a 10 Cup may be Brazil
England 63% much weaker team than the two it fielded in
Spain 28%
Germany 59% its previous World Cup appearances (1998 Many readers might feel that European
France 22% and 2002), South Africa’s status as the host Champion and secret favorite Spain is under-
Italy 59%
Argentina 55% England 21% nation boosts the likelihood that it will make rated according to this model. Our English
Mexico 52% Argentina 20% it into the round of 16. So far, every host readers will certainly feel the same about
Chile 49% South Africa 17% nation has at least moved one round beyond England, though here there might be a home
France 49% Uruguay 14% the first one. Any other outcome would be bias. The truth is, however, that both teams
Portugal 47% a huge surprise – even though South Africa have tended to be underachievers when it
Portugal 14%
enters this World Cup with the weakest Elo comes to the World Cup. In the last three
Serbia 42% Chile 13%
ranking of any team, even below North World Cups, Spain and England were ranked
Uruguay 42% Source: UBS Wealth Management Research Korea and New Zealand. Besides South among the strongest teams but did not
USA 33%
Africa, there is no other African team on this make it beyond the quarter-finals. Spain’s
Australia 33% top 20 list, but teams with a likelihood of best World Cup result was fourth place in
Denmark 32% reaching the second round of more than 1950. England’s performance at World Cups
Table 3: Winner
Switzerland 30% 25% are Didier Drogba’s Ivory Coast and has been better, having won it once in 1966
South Korea 29% Samuel Eto’o’s Cameroon. and finishing fourth in 1990. Given the
Country Likelihood to win the World Cup 2010
Paraguay 29% objective and subjective strengths of both
Brazil 22% teams and of their respective national cham-
Source: UBS Wealth Management Research
Germany 18% Semi-finals: the usual suspects pionships, the stardom of many of their play-
Italy 13% 9 plus two surprises ers and the football frenzy in both nations,
Netherlands 8% this can truly be seen as an underachieve-
France 6% This home bias could help South Africa reach ment.
Argentina 5% the semi-finals, though we view such an out- For the sake of keeping the World Cup
Spain 4% come as very unlikely. We are more likely to interesting and making forecasting its out-
England 4% see some of the traditional teams in the come even more challenging, it would be
semis. Surprises could come from Chile and great if one of those two teams can engineer
Source: UBS Wealth Management Research
Portugal and (going beyond this top 12 list), a big surprise in Johannesburg on 11 July.
South Korea, the US or Australia, which have andreas.hoefert@ubs.com
rather strong teams, according to their Elo Chief economist, UBS AG
rankings.

30 UBS investor’s guide special edition April 2010 UBS investor’s guide special edition April 2010 31