R - FIFA football world cup 2018
2018 FIFA World cup football is reaching semi finals tonight. It is time to have a review and to draw the lessons learned. As a reminder, know that I am just interested about right pronostic of the end of the game, I mean Win, Draw or Loss, not about the score.
Pool predictions
Here is the summary table of my pronostics … not really good prediction in fact
| Pool | Result | ||
|---|---|---|---|
| name | number of match | number of right pronostics | ratio % |
| A | 6 | 4 | 66.0 |
| B | 6 | 1 | 16.7 |
| C | 6 | 3 | 50.0 |
| D | 6 | 2 | 33.0 |
| E | 6 | 3 | 50.0 |
| F | 6 | 3 | 50.0 |
| G | 6 | 1 | 16.7 |
| H | 6 | 6 | 100.0 |
| All | 48 | 23 | 48.0 |
I achieved a 48% right pronostic score. Not really good. I had no clear strategy and was insufficiently knowledged about some teams. I bet on famous teams that really counter performed e.g. Germany, Argentina and under estimated many teams that performed really well Australia, Iran, Morroco, ….
8th of final
On second blog about FIFA WC, I provided win probabilities for teams based on their pool performance. If you had followed this mathematical approach, you would have got 7 right results on 8. Only exception, is the organizer Russia, defeater of Spain. 7 on 8 means 87.5% of success in prediction. This appears to be a reliable prediction system for football game
4th of final
Here, if you would have applied same reasoning as the one for 8th of final, you would end up with a 75% success score of predictions.
| Team A | Team B | Result | |||
|---|---|---|---|---|---|
| country | pool performance place | country | pool performance place | winner | logical |
| Uruguay | 3 | France | 5 | France | illogical |
| Brazil | 4 | Belgium | 1 | Belgium | logical |
| Sweden | 8 | England | 6 | England | logical |
| Russia | 7 | Croatia | 2 | Croatia | logical |
France n°5 defeated Uruguay n°3 is the only counter intuitive result.
Semi final
I predicted on first blog about FIFA WC, that France will reach semi final. It happened. Now my predictions for those semi finals.
| Team A | Team B | Result | |||
|---|---|---|---|---|---|
| country | pool performance place | country | pool performance place | winner | logical |
| France | 5 | Belgium | 1 | France | illogical |
| Croatia | 8 | England | 6 | Croatia | logical |
Final
Again, a good prediction, although contrary to the logic.
| Team A | Team B | Result | |||
|---|---|---|---|---|---|
| country | pool performance place | country | pool performance place | winner | logical |
| France | 5 | Croatia | 2 | France | illogical |
Conclusion
No easy pattern to foresee game result. To determine which team will pass, I would suggest to have two different strategies. One for the pool, one for the direct elimination match.
For the pools, stick to team international ranking, as basic strategy, and ponder it, according to inverse of average age of the team. This world cup has shown a great advantage to young teams over more experimented ones.
For the direct elimination match, reusing global pool results ranking, weighted by the ratio total number of minutes spent being running after score on total number of minutes spent being ahead on score. This is very factual of the strength of the team at the considered time.