Is historical performance useful for predicting the World Cup winner?

Is historical performance useful for predicting the World Cup winner?

By Mirio Mella Dec 18, 2013

Tweet

With the draw for the 2014 World Cup in Brazil now complete, all discussion is focused on who’ll lift the trophy on July 13th. Given the tournament only takes place every four years, is historical performance useful for predicting the eventual winner? And if not, what is?

For any bettor looking to predict the 2014 World Cup winner, historical form would appear to be an obvious place to start. There have been 19 FIFA World Cups since the inaugural tournament in Uruguay in 1930, so analysing nations’ historical performances could be considered a reasonable way to arrive at a rough idea of their chances.

Year

Host

Host confederation

Winner

Winner confederation

Runner-up

No. of participants

1930

Uruguay

CONMEBOL

Uruguay

CONMEBOL

Argentina

13

1934

Italy

UEFA

Italy

UEFA

Czechoslovakia

16

1938

France

UEFA

Italy

UEFA

Hungary

16

1950

Brazil

CONMEBOL

Uruguay

CONMEBOL

Brazil

13

1954

Switzerland

UEFA

W. Germany

UEFA

Hungary

16

1958

Sweden

UEFA

Brazil

CONMEBOL

Sweden

16

1962

Chile

CONMEBOL

Brazil

CONMEBOL

Czechoslovakia

16

1966

England

UEFA

England

UEFA

W. Germany

16

1970

Mexico

CONMEBOL

Brazil

CONMEBOL

Italy

16

1974

W. Germany

UEFA

W. Germany

UEFA

Netherlands

16

1978

Argentina

CONMEBOL

Argentina

CONMEBOL

Netherlands

16

1982

Spain

UEFA

Italy

UEFA

W. Germany

24

1986

Mexico

CONMEBOL

Argentina

CONMEBOL

W. Germany

24

1990

Italy

UEFA

W. Germany

UEFA

Argentina

24

1994

United States

CONCACAF

Brazil

CONMEBOL

Italy

24

1998

France

UEFA

France

UEFA

Brazil

32

2002

South Korea & Japan

AFC

Brazil

CONMEBOL

Germany

32

2006

Germany

UEFA

Italy

UEFA

France

32

2010

South Africa

CAF

Spain

UEFA

Netherlands

32

The historical World Cup winners’ data (table 1) shows that the Confederation of the winner and of the host appears to correlate. But does more detailed analysis of those 19 tournaments support this theory?

Continental shift

The winners of the first five World Cups came from the host Confederation, but this should come as no surprise. Take the first tournament in 1930, hosted and won by Uruguay. Only four of the 13 competing teams were from Europe, the reason being that it took them three weeks to make the Trans-Atlantic journey.

Aside from there being more South American teams, the implications on performances for the travelling Europeans would have been significant – not much chance of training on board a ship – compounded by the fact that few European players would have had any experience playing overseas.

Travel remained a key issue at the 1934 & 1938 events in Italy and France, which respectively saw only 4/16 and 3/16 nations represented from outside Europe. Both tournaments started in the knockout stage, giving teams no chance to settle into the event. The Italians won both.

Even by 1950, holders Italy travelled by boat to Brazil, where many teams withdraw due to financial constraints, or the difficulties of travel. Only 13 teams participated (one Group had only two teams), and India instructed that they couldn’t play in bare feet, withdrew. It is clear few countries participated on merit. The final saw Uruguay defeat Brazil.

Home Continent Advantage

In these early World Cups, the skewing of participation towards home continent teams, the serious challenges of travel and the haphazard nature of qualification, all contributed to Home Field Advantage.

Though the difficulty of getting to World Cups receded as air travel became more prevalent, HFA was still a significant issue up to the era of the modern format (considered to be 1986). Very few players were based overseas, and tournament organisation remained amateurish by modern standards. The hosts won in 1966, 1974 and 1978.

In the modern game HFA has been widely researched, with studies such as Harvard’s Research Assistant Ryan Boyko who suggesting a 0.1 goal advantage for every 10,000 fans present. It is believed HFA provides an influence on refereeing decisions and psychological benefits to home players.

Beyond the influence of these universal HFA measures, World Cups have seen inferred examples of bias, particularly in 19741, and 19782, as well as explicit imbalances. In both 1978 & 1982, final group games weren’t played in parallel, leading to the infamous ‘Schande von Gijn’ – ‘the Disgrace of Gijon’ – where West Germany and Austria openly contrived a 1-0 result which saw both progress to the second round of the ’82 event, at the expense of Algeria.

Given that since 1978 the host has won just one of the eight tournaments held, we might infer that the game has become more global, organisation more professional and scrutiny has increased. For example, the rules changed following the open collusion in Spain, with all final group games played in tangent.

Equally important is the fact that the choosing of hosts has become more political and moved away from established nations. Only two of the last eight hosts have previously reached a final.

Power Shifts & UEFA/CONMEBOL Bias

The relationship between hosting continent and eventual winner was first broken in 1958 when Brazil and a teenage Pele won in Sweden. This was the dawn of the Golden era of Brazilian soccer, with the national team winning again in 1962 in Chile and 1970 in Mexico. The success of the Selecao illustrates another key factor for bettors to consider when predicting World Cup winners – power shifts.

In 1977 Pele made a now infamous prediction – ‘an African nation will win the World Cup before the year 2000’. Though Pele’s playing ability didn’t qualify him as a pundit, the failure of his prediction is relevant as it underestimated the enduring strength of Euro-South American axis of power in soccer.

FIFA places more value on games featuring European and South American teams than otherwise. This explains why only three teams from outside Europe and South America are ranked in FIFA’s top 30 – the USA, the Ivory Coast and Ghana. If all continents were treated equally, then in theory the Ivory Coast and the USA would be seeds for the World Cup, impacting their chances of success.

Though the power of Europe and South America has withstood, there have been significant power shifts within that duopoly. This was seen at France 1998 and South Africa 2010, with two first time winners – France & Spain – going on to win European Championships.

The power dynamics of soccer reflect its geographic origins, and other random factors that influenced its adoption around the world (e.g. English public schools in Argentina). There isn’t a direct relationship, for example, between population size and success, but socio-economic factors are relevant to improvement at grass roots, as well as influencing migration, but these factors are relevant over decades not years.

Any changes, and influences of the shape of international soccer are easy to divine in retrospect, but Spain had flattered to deceive on many occasions before winning in 2010, while the French went into the 1998 event 18th in the FIFA World rankings.

Luck – The best laid predictions

Brazil’s trio of wins from 1958-70 was interrupted by England’s solitary victory in 1966. The Canaries failure highlights another key factor that World Cup bettors must consider – luck.

Pinnacle Sports has previously written about how success=skill + luck. In a sport where luck plays a part, and in the context of a short format event, it should come as no surprise that the winner may not necessarily be the most talented team, or even the team that has played the best football.

Every modern tournament begins with a qualifying tournament, the basis for which is a random draw, then the finals themselves are preceded by the glamour and often confusion of the draw (again random), where national coaches nervously wait to see what hand fate will deal them.

Any bettors who bet on the World Cup before the draw would have been equally nervous as this random process has a significant impact as witnessed by the drift in Spain’s outright odds since being drawn in a tough Group B which could see them meet Brazil as early as the 2nd round.

Once the tournament is underway fate can lurk in many places, not just the pitch. The team bases are scouted many months in advance to try to ensure the best environment, but the draw in Brazil has put thousands of miles between base camps and game venues (e.g England). Even when preparations go smoothly it doesn’t mean this translates onto the pitch.

Up to the 1966 World Cup Garrincha & Pele had played 40 international games together without defeat. In their opening 2-0 win over Bulgaria at the tournament in England they both scored, but Pele picked up an injury after being targeted by the Bulgarians, which meant he missed the game against Hungary and they lost 3-1, then the unthinkable happened as Brazil – without Garrincha – lost to Portugal , and were eliminated; their worst performance at a World Cup.

The pair would never play together again; what would have happened if Pele and Garrincha had avoided injury? We will never know, but this kind of circumstance and countless other unforeseeable events in other World Cup tournaments will have had huge bearings on the outcomes.

What constitutes luck is hard to define; poor refereeing decisions are a good example. HFA suggests that the referee is influenced by home support, which could account towards South Korea’s unlikely passage to the last four in 2002. However most games at the World Cup do not involve the host, so infamous episodes such as ‘the Hand of God’ (1986 quarter-final); the Schumacher/Battiston3 incident (1982 semi-final) or the mystery surrounding Ronaldo shortly before the final in Paris (1998) were all extremely important, but impossible to predict.

Luck operates on another level by for example bringing together an unusual concentration of talented players within a generation, which could be said of the current Belgian side.

Unpredictability – Strange things will happen

Sometimes, luck isn’t the issue but sheer unpredictability. Who would have predicted that a 38-year-old African (Roger Millar) would be a hero at the 1990 World Cup, or that a largely unknown Italian striker, Toto Schillaci would top-score for the hosts. In 1994 the Golden Boot was won by a Russian who’s six goals were his only ever at international level, with five in one game against Cameroon. (Read more about betting on the Golden Boot winner here)

It is now clear that there were huge differences across the previous 19 World Cups in terms of participation, format, regulation, biases and luck but outside of these tournament specific elements, the socio-economic and political context were also entirely different which is very relevant.

World Cup comparability – Out of Sample

The challenge of making predictions where a large number of important contributing factors and variables persist, but there is a comparatively small sample size, is known as being ‘out of sample’. The World Cup provides a potent example of this.

Given the difficulty in direct comparison against previous World Cups and the role that luck and wider context play, any tipsters who proclaim certainty about predicting winners should be treated with caution. With such difficulties in the data bettors should instead focus a probabilistic approach to picking a winner, employing something like a Bayesian method (to be looked at separately), and seeking value in odds, rather than trying to draw clear conclusions based on historical tournament trends.

Things to consider:

Given the format of the tournament has been consistent since 1986, barriers to travel removed and soccer become more globalised, results since then should be given far more weight. For example Uruguay’s World Cup wins in 1930 & 1950 are less relevant than their 2010 Copa America success.

Early tournaments provided extreme HFA but this has waned in modern times – since 1978 the host has won just once.

Though Pele was wrong with his prediction about an African Nation winning by 2000, past success doesn’t guarantee future success; the same teams won’t keep winning but power dynamics within World soccer change relatively slowly, and are the result of a multitude of factors, some entirely outside of the game.

Click here to see the latest 2014 World Cup winner odds

Notes

1. In the run-up to the 1974 World Cup final between Germany and The Netherlands, German tabloid Bild were involved in a “setup”. After bribing security guards the paper paid for a group of escorts – in various states of undress – to jump into the pool the Dutch team were relaxing by as paparazzi lay in wait. As the story circulated, Cruyff’s wife was furious and the star spent the night convincing her that nothing happened. The damage was done however, and Germany won 2-1.

2. Controversy surrounded the 1978 World Cup. None more so than the second round game between Argentina and Peru. The hosts needed to win by four goals to proceed to the final and did so by 6-0. Conspiracy theorems ranged from interference from the Argentine military dictatorship, to the Peruvian goalkeeper – who was born in Argentina – throwing the game. Argentina went on to win the World Cup, while no conspiracies were ever proved.

3. Battiston was knocked unconscious, and later slipped into a coma. Michel Platini later said that he thought that Battiston had died, because “he had no pulse and looked pale”. The Dutch referee Charles Corver did not even award a free kick for the incident.

If you have feedback, comments or questions regarding this article, please email the author or send us a tweet on Twitter.