Category: Statistics

  • Predicting the world cup winner

    Solely for academic purposes. There are no blessings in betting.

    For convenience this study uses fifa world ranking scores1 of mid-december of the year in which the world cup2 was held.3

    YearFifa ranking no. 1Fifa ranking no. 2Fifa ranking no. 3Fifa ranking no. 4Winner world cup
    1994BrazilSpainSwedenItalyBrazil (final won from Italy)
    1998BrazilFranceGermanyCroatiaFrance (finalist Brazil)
    2002BrazilFranceSpainGermanyBrazil (finalist Germany)
    2006BrazilItalyArgentinaFranceItaly (finalist France)
    2010SpainNetherlandsGermanyBrazilSpain (finalist Netherlands)
    2014GermanyArgentinaColombiaBelgiumGermany (finalist Argentina)
    2018BelgiumFranceBrazilCroatiaFrance (finalist Croatia)
    2022BrazilArgentinaFranceBelgiumArgentina (finalist France)
    2026 (pred.)4ArgentinaSpainFranceEnglandsee text

    As can be noted from the table the winner is always from the first two positions. Furthermore it is split exactly 50/50 between the first spot and the second spot. This would indicate that both Spain and Argentina have a chance of 0.5 to win the world cup.

    Can we improve on this? Believe it or not but the historical results since 1952 are absolutely equal: Argentina has won six times, drawn two times and lost six times. So the only indication that we have is the “resultaat” or “=goals scored – goals conceded” which yields “-1” on a total of 37 goals. So this favours Spain over Argentina with 1/37 = 0.027 (p).

    1. https://inside.fifa.com/fifa-world-ranking/men?dateId=id56 ↩︎
    2. https://en.wikipedia.org/wiki/List_of_FIFA_World_Cup_finals ↩︎
    3. This is possible if one assumes a point of view of cyclial time rather than linear time. ↩︎
    4. Fifa world cup rankings data as of 11th of june ↩︎