@texterstexte Magst du auf Bluesky dem Account @ap.brid.gy folgen? Dann könnte man deine Posts auch auf Mastodon sehen. Siehe hier: https://t.co/dkFP45FXSA
@niggi Zumindest die bisher genannten >24000 getöteten Kinder & Frauen sind nun nicht mehr plausibel. Um diese Zahl zu erreichen dürfte unter den noch nicht identifizierten Todesfällen nur ein sehr niedriger Anteil an Männern sein.
@angelopq29 I give an outline of the algorithm in the GitHub project: https://t.co/8CpCYS6Uia
The actual algorithm is available here: https://t.co/WHugH7otBQ (you can ignore the matrix sorting and the caching of probabilities - that is just optimization to speed up the computation).
@mo_abdulkerim There is only a very low number of pairings that are not allowed which leads to a high symmetry. I.e., teams that have the the same allowed opponents will also have the same probabilities.
Draw probabilities for the 22/23 knockout rounds of UEFA Europa League (left) and Conference League (right).
Interactive version: https://t.co/LrnncZQsgo #UEL#UELdraw#UECL#UECLdraw
@3LeggedSledDog If UEFA requires that associations with 4 teams must have 2 teams playing on each day, the probabilities shown in the table are incorrect. Do you have any official documentation on that restriction? If so, I might try to update my calculator to reflect it until Monday.
@3LeggedSledDog In your example, it would still be possible to have at least one team from ENG, GER, and ITA on each day - but not a 2:2 distribution for the teams from ENG and GER. 2/3
@miri_munich @paulchenpeter Ja, das ist berücksichtigt. Wenn z.B. der Binomialkoeffizient "19 über 4" verwendet wird, sind die 4 Teile nicht unterscheidbar. Es gibt also für das Legen der ersten 4 gleichen Teile nur 19!/(15!4!) = 3876 und nicht 19*18*17*16 = 93024 Möglichkeiten.
@miri_munich @paulchenpeter Im Prinzip multipliziert man nur ein paar Binomialkoeffizienten, dann kommt man auf seine (korrekte) Lösung. Erst werden 19 Plätze vom 4 Karten belegt, dann 15 von 4, usw. Macht dann 244.432.188.000 Möglichkeiten.
@DaCinal Now it makes use of the symmetries of different draws and thus avoids many calculation steps. Could be optimized further by implementing an actual graph canonization algorithm but I didn't bother doing this after UEFA announced the new UEL mode without an R32. 2/2
Some more probabilities: UEFA Europa League 21/22 Knockout Round Draw. Teams that finished second in #UEL face teams that finished third in #UCL. Teams from the same association can't be matched.
Interactive version here: https://t.co/IjVoBCqHme
#UELDraw
@DaCinal Bonjour Cinal! My first 8x8 wasn't efficient enough for 16x16. The overall algorithm stayed the same but I had to do the storing of intermediate results (i.e., the memoization) in a more sophisticated way. 1/2
Now that Villarreal has made it to the round of 16 of the #UCL, these are the final draw probabilities.
Interactive version here: https://t.co/LrnncZQsgo #UCLdraw