I was wondering about issues you mentioned and I think it works as follows. Elastic effect works even if you have 1 related skill above the one you train.
I'm somehow tempted to believe this. But what to do when one skill is higher and one lower than the skill we're looking at? or is it that only the highest skill is slowed down, and all the other ones are accelerated? even if it is something like 2-10-11? the 10-skill would be accelerated? hard to believe somehow, but maybe.
I will just speculate right now, but seems to me strange that we have here different training types (2 position minimum on rebounding - two related skills (IS, ID) and 2 position minimum on DV - three related skills and so on). The question is if it is this way because developers wanted to offer wider opportunity for elastic effect or the opposite.
This can be figured out by the character of these skills. Imagine how is DV underrated, it doesnt affect match ratings and it seems that it is only helping to the scoring abilitiy which has to be on high level already otherwise does not have any effect.
Rebounding seems to be way more important, maybe thats why there is no opportunity to have 1 position training for that and you have to have a good IS and ID to make elastic effect happen.
Imagine the future when all of this managers know, how it would influence theirs behavior?
Logically it would mean that DV would become most effective training, however it could be only until the day managers figure out that the best way to stop that is the shotblocking. For RB such a tight elastic option can be necessary for managers to realise that theirs guards can use ID at decend level...
So I see this issue like a twisted way of motivational factors, while easiest or necessary for elastic training are the skills which have to be most usuall for a competitive game. Therefre it means that you need only 1 related skill above, if Im right.