Thank you for the explanation, but I think (unfortunately) it's much more complicated than that.

Agreed. I think my second table only considered the case of the first block per round. I am surprised to see gaps up to 28, though. I guess it makes sense, given that 42-28 is 14, which is one less than the threshold - but I haven't manually traced through the steps, yet.

Thank you for the code. I got it running today, and if I find time, I might play with it by reducing the number of witnesses to 3 (top-2 plus random) and increasing by one from there in order to understand how the patterns change as the number of witnesses grows. It might also be interesting to reintroduce random shuffling and simulate thousands of blocks in order to see the expected distribution over time. I don't think I'll have time this week, though.