Depends entirely on the circumstances. I’d argue having the extreme surplus of players makes this analogy more likely:
While I wrote that for a different suggestion, it can also be applied to this one.
To clarify, since you seem to base your protests on that both of the suggestions would be implemented in the first post you quoted of mine, is that the probability estimations can also include things like size on matchmaking pool to determine the likelihood of it being a non-randomized pattern.
It doesn’t have to be as simplistic as you seem to imagine.