It's coming home! Actually it's probably not
Who needs Ronaldo and Mbappé when you have the big lad Data Science in your team?
After a delay of what seems like forever but is actually only a year, Euro 2020 finally kicks off.
Who will win? Will Mbappé inspire the French to add the Euros to the World Cup they won in 2018? Will Ronaldo once again lead holders Portugal to glory? Will Wales taste the semi-finals as they did in 2016? Will Foden do a Gazza and light up the tournament or will Scotland come home too soon?
To find out the answers to these questions you can spend hours poring over the form book, analysing previous results, goals scored and goals conceded but you don't need to do that because we've done it for you. We've done the data science taking all these factors into account and predicting the result for every group match. From this we have worked out probabilities for who is likely to progress from the group stage into the knockouts by virtue of finishing first or second in each of the six groups and which four teams will qualify as the best third placed teams.
After each round of matches we will feed the results into the model to update it and amend our predictions. Check back in or follow us on Twitter for the very latest Euro 2020 (or is Euro 2021) predictions.
Predictions for the group stage
For each team we show the percentage probability of progressing to the knockout stage once our model has predicted the outcome of every match played.
We predict mixed news for the home nations. England will qualify as Group D winners but Scotland and Wales will finish bottom of their groups. If Scotland and Wales are to progress then they will need to upset the odds and become one of the four best teams in 3rd pace who qualify. Scotland v Czech Republic and Wales v Turkey look to be crunch matches in deciding their fates.
In Group F, known as the 'Group of Death', our model calculates that Germany has a slightly higher probability than France of getting to the qualifying round. Germany play their group games in Munich and home advantage is likely to make the difference in who tops the group. We know from our previous research that home supporters make a huge difference to teams. Portugal is predicted to be third. If Hungary is to qualify then the Magyars will need the shock of all Euro shocks to progress as our model only gives them a 20% chance of progressing.
Group C looks as if it will be very tight between Ukraine and Austria for 2nd and 3rd with Ukraine expected to edge it into 2nd place in the group. Similarly, in Group E, Sweden and Poland have more or less the same probability of progressing.
Group stage predictions : The probability each team has of progressing into the knockout stages
The race for 3rd place
Four of the six teams finishing in 3rd place will qualify for the knockout stage. The big question here is how many points is likely to be enough to get through as one of the best 3rd placed teams?
Data science can help answer this one. We ran a model for the 6 teams predicted to finish in 3rd by analysing all the different results permutations and averaged out how many teams would finish with one point, two points, three points and so on. We then worked out the percentage probability of teams actually qualifying in 3rd with one point, two points, three points and so on.
For example, looking at all the possible permutations of results it is possible for one of the six 3rd placed teams to finish with two points. It is really unlikely, however, that two points will be enough as our model suggests that the probability of qualifying with two points is less than 14% on average.
Running the different points permutations shows that four points or more will see you through - win one, draw one and you should make it through as one of the best 3rd placed teams. On average, between 2 and 3 teams are predicted to finish 3rd on three points. Goal difference could be important in deciding who goes through and our model suggests that 67% of the time a team in 3rd place with 3 points and a goal difference of no worse than -1 should be enough to qualify.
Bringing this altogether we predict that the four best 3rd place teams will be Portugal, Austria, Poland and Russia.
Who will we get to the knockout stage?
If the model is right it is probable that England will play France in the next round. In fact, finishing 2nd might be better because England would then face Poland in the round of 16. Losing the opening match to Croatia may not necessarily be a bad thing!
If Scotland and Wales do manage to make third they are rewarded with fixtures against some of the big teams. There's a lot of either/ors involved but we think Scotland would play either Spain, Netherlands or Belgium and Wales would play either Spain, Belgium or Germany.
So, is football coming home? The probability doesn't look good but anything can happen once the whistle blows.
For more data science at the Euros go here.