So, to sum the last two posts up. The lower the delta, the fewer rounds to predict – the better the model is at projecting the table positions and how many points a team will get. The easiest way of visualizing would be to look at how well the model performs in terms of correctly guessed table positions or near table positions.
Allsvenskan and Superettan has 8 rounds to go so the model should be able to guess at least 30% of all table posistions right and be right within 1 position up or down at about 70%. And within 2 positions up or down we should look at an accuracy of at least 80%. To visualize the model I have chosen box plots. The box has 4 values each where the two inner bounds stand for a 50% chance of the team taking points within that range. The outer bounds are more or less the absolute max and min theoretical points. The chance of a team getting points outside of that bound is lower than 1%. So without further due – here are my predictions. And I will not edit this post in order to try to cheat. So forgive me for any spelling mistakes.
Allsvenskan is far from over. I can see three teams still having a realistic shot at the title. IFK Norrköping may have to hope that all three teams ahead of them under perform which seems unlikely. Then there is a clear gap between place 6 and 7 that just won’t change. In the bottom of the table, Åtvidaberg is the only team that seem destined to be relegated. For them to overtake Örebro is down to a possibility of 6.25%
Another observation is that Hammarby are expected to gain a few more points than GIF Sundsvall.
For Superettan, the top three spots seem very clear. Sirius has a 6.25% chance to overtake Östersund to claim the second direct place to Allsvenskan. Remember that the model does consider the difficulty of remaining games to play for all teams. There is a small overlap between ÖFK and Jönköping within the inner box but for ÖFK to actually win Superettan it is quite obvious that they will have to win against both Sirius at home and against Jönköping away.
What may be more interesting is the bottom of the table where my model puts Brommapojkarna ahead of IK Frej. That will be interesting to follow since it is one of few significant changes of the tables that my model suggests.
The teams that have secured next years contract (at a possibility of 99%) are:
For Allsvenskan: Djurgårdens IF and upwards
For Superettan: Varbergs BOIS and upwards
Now, one could argue that the model is wrong and that the predictions won’t be correct at the end of the season. And I am the first to agree – the model will be wrong, at the percentages stated above! So: Malmö FF will not win Allsvenskan. And I will be right about that 87 times out of 100!
Delivered by Everysport
Thanks to Football Data