THis would be the right way of doing it, but we should remember thatFEPG wrote: ↑Wed Sep 01, 2021 01:56If I were to do it, I'd probably use ClubElo & simulate each group independently using the Monte Carlo method to find the mean number of meaningless games for each permutation (4! = 24 perms per group). Then I would find the arrangement (1 perm from each group) with the lowest weight (sum of means) satisfying geographical constraints (i.e. teams from the same city cannot play at home on the same week, etc).elkjiaer is back wrote: ↑Tue Aug 31, 2021 11:16In other words: Given 4 teams where 1>2>3>4, what is the best allocation of teams into calendar order A, B, C, D to ensure maximum excitment of the competition?
But then again, there's a reason why I don't work at UEFA.
1) Uefa does not care about Elo ratings or any other rating that is not the Uefa coefficients
2) They should define a rule which is generally valid for every competition and every season for the sake of transparency
Besides all the constrains, what Uefa should be aiming for is to minimize the number of dead/half-dead matches especially on the last 2 matchdays. This is a bit of a problem in competitions like UCL and UEL where at least 3 final positions out of 4 are relevant, but in a competition like ECL (and formerly also UEL), only finishing 1st or 2nd matters so there the issue is even greater.
Let´s just use intution for now and forget for a moment about simulations.
You have 4 teams where A>B>C>D (using either pots or their actual coefficient)
You are aiming to maximize the uncertainty so that the highest possible number of teams has still something to play for on the last matchdays (none is eliminated nor already qualified for example)
So basically you just to make sure that A and B play against C and D as late as possible so they do not gain too many points.
In the new configuration 1-2-3-3-2-1 it means that
MD1: D-A, C-B so that on MD6 A-D and B-C
MD2: A-C, B-D so that on MD5 C-A and D-B
MD3: A-B and C-D and MD4 B-A and D-C