The probabilities are taken from that general simulation (linked) of qualifiers, before draw - therefore the question about how likely it is for Germany to fail and neither win the group nor be runners-up was taken as 0.6% for approximation (this was just an example for reasonable chances). Yes, specific groups have somewhat different probabilities... as mentionned right at the beginning of the text. Whatever approach, that can be fixed easily, thanks.amirbachar wrote: There is a major bug... The events are not independent, for example for both Germany and The Netherlands not to qualify means not only Northern Ireland pass both of them, but also another team from that group, which is very very unlikely.
One major bug is this:
I have only the possibility for Netherlands/Germany to end up in UNL play-offs with Croatia/Poland specifically, dependant incidents of course. Here the other independant incidents of having Netherlands/Germany with every other possible combination of 2 division A teams should be added (same with only 1 or zero other division A team joining them).
I clearly 'over-simplified' it a bit too much there with the ~0.05% (summary of these independant possibilities still leaving Netherlands/Germany in play-offs is somewhat higher).
The funny thing - though - is this:
According to your table I was 'reasonably close'