This is a second post in the scoring first in football project. The first reported on six European leagues in the 2017-2018 season.
The methodology discussed in the first post was replicated in my analysis of the 2010 and 2014 FIFA World Cups. My data came from the Wikipedia pages for the tournaments.
Here I share probabilities for:
- score first win (SFW)
- score first draw (SFD)
- score first lose (SFL)
- 0 v 0 game (0goals)
- scoring first and not losing (SFNL)
I wondered if these measures could inform a naive Bayes approach to probabilistic behaviours in the upcoming World Cup.
For the 2010 World Cup my median probabilities are:
- SFW 0.67 (Group 0.67, Knockout 0.69)
- SFD 0.17 (Group 0.17, Knockout 0.13)
- SFL 0.00 (Group 0.17, Knockout 0.06)
- 0goals 0.17 (Group 0.17, Knockout 0.13)
- SFNL 0.83 (Group 0.83, Knockout 0.82)
For the 2014 World Cup my median probabilities are:
SFW 0.67 (Group 0.67, Knockout 0.69)
SFD 0.00 (Group 0.00, Knockout 0.13)
SFL 0.17 (Group 0.17, Knockout 0.06)
0goals 0.17 (Group 0.17, Knockout 0.13)
SFNL 0.82 (Group 0.75, Knockout 0.82)
Note: for the purposes of this analysis, I have used the score at the end of regulation time. The data do not include any games won in extra time or in penalty shoot outs. I have recorded the 2010 World Cup final, for example, as a 0 v 0 (0goals) game. Spain scored the winning goal in the second period of extra time (Iniesta 116 minutes).
Photo Credit
Soccer City (Shine 2010, CC BY 2.0)
