- World Cup 2018 has been one of the most unpredictable world cup, after World Cup 2002.
- Ranking systems for teams, work through out year and considered to be most robust measure of performance and hence, potential for winning the cup.
- Both FIFA and ELO ranking of teams were considered, to review how teams have fared from 1994 to 2018.
World Cup 2018
World Cup as it always had been, came up with loads of surprises and some of the World Cup tournaments like recent ‘Russia 2018’, has been enthralling. Betting Market had their odds before World Cup started. It turned out to be, France (with odds 6.5) pitted against Croatia (with odds 33) in the final. France and Croatia had little more than 19% chances of winning the Cup between before WC 2018 started .
Before tournament started all prediction pointed out towards two teams – Brazil (with odds 4) and Germany (with odds 4.5), who were actual favourites in the tournament. Many, who predicted Germany as winner before tournament shifted their bet on Brazil as possible winner, after early departure of Germans. Even Brazil could not last beyond Quarter-finals, with Germany bowing out of the tournament in group stage itself, ranking last in Group F, after Sweden, Mexico and South Korea. In other words two countries which had 47% chances of winning the Cup between them eliminated early. In effect, a huge number of people lost their money!!!
FIFA and ELO Ranking for Football
FIFA and ELO rankings are two ranking systems of Men’s football team. Prior to introduction of FIFA ranking in 1993, ELO ranking was used to rank football teams around the world. Though major factors for both FIFA and ELO ranking are same, they differ in some areas. FIFA ranking procedure includes factors match results, Importance of the match, strength of opposing team and strength of confederation. ELO ranking method also include the goal difference in a match.
Prediction for Winners
Before the tournament started, we had all kind of predictions from Soccer Experts and Legends, Astrologers, Investment Banks and from every possible sources. When proved wrong, they all came with new updated predictions based on recent match results, eliminations and team performances. They used different methodologies and factors for prediction purpose.
Performance of top Eight teams and their Rankings
World Cup quarter-finalists from 1994 to 2018, were considered to compare the estimated ranking (prior to World Cup) to actual World Cup Performance. For predictability purpose following indicators are to be used to compare predictability performance across ranking methods and across
Unpredictability Index:
Measure of unpredictability in outcome of Quarter-finalists and first four position-holders. Bigger number indicates higher unpredictability. If teams rank in World Cup as per FIFA/ ELO ranking, Unpredictability Index would be zero.
Avg. Rank of Quarter-finalists:
If first eight teams as per ranking, feature in the quarter-finals, Avg. Rank of Quarter-finalists, would be 4.5. Bigger number indicates, presence of one or more outlier/s.
Unpredictability Index, which consider variation, is more robust indicator of unpredictability than, Avg. Rank of Quarter-finalists.
As it happened…
To explain the above table, the quarter-finalists are mentioned in a row, with first four position holders are put sequentially as Winner, 2nd, 3rd and 4th spot. Loser of Quarter-finals, are put as per FIFA ranking afterwards. The FIFA and ELO ranking of each team are mentioned below, with Red and Blue respectively.
At right-hand side of the table, Unpredictability Index and Avg. Rank of Quarter-finalists for FIFA and ELO ranking are mentioned as per
The later table, perhaps provides preferable representation of rankings as it considers relative ranking of participating teams in any given World Cup. From both the ranking (and both the tables), it is evident that World Cup 2002 and 2018 have had the most surprising line-ups in quarter-finals and subsequent matches. Elo ranking has lower -Unpredictability Index and Avg. Rank of Quarter-finalists from second table, only exception being Avg. Rank of Quarter-finalists in 2002. So ELO ranking can be considered to have comparatively robust predictability. No wonder, FIFA ranking was reviewed and would follow ELO ranking methods from June 2018.
One important observation from the table, team with FIFA ranking 1 failed to win the world cup in last seven editions of World Cup and same is applicable for ELO ranking with only exception being 2010 edition of World Cup. If percentage of quarter-finalists who actually rank in top eight teams are considered, ELO ranking fares better there for last seven editions. As the table shows, ELO ranking had 57% of the first eight of the teams played in quarter-finals for World Cup since 1994, whereas 48% of the quarter-finalists are from top eight teams as per FIFA ranking during the same period.
The intention of article is not to discuss the comparison for establishing better ranking methodology, but questioning prudence or audacity, to predict the outcome, that would take its own time to unfold in July, before the tournament actually starts in June. If the most rigorous ranking systems cannot predict it for you, you are sure to fail; and this is the most beautiful thing about football World Cup. By the way, next time it would be in October…