Okay, let's start from the top:
Potential for abuse is an extremely poor developers excuse.
Feasibility, however, is a good excuse. And in this case, the feasibility of fixing what you call "potential for abuse" and others might call "completely obliterating pretty much any integrity the ranked system had up to this point" is a very good excuse.
You have to take risks
If you jumped in front of a train and managed to survive, you'd become insanely famous. That's no reason to jump in front of a train. "You have to take risks" is not the same as "You have to take every risk and ignore any dangers".
That's a job for testing to ensure the design works as intended.
Again, feasibility comes first. If you can't figure out a way to make it work before implementing it, guess what: You don't implement it. Having developers implement and testers test features that were doomed from the start is a waste of two departments' time.
It hasn't worked with some experiments in the past
Exactly. Weren't you the one that mentioned testing whether the design is working as intended? This one didn't work.
I also don't quite understand how only getting your rank back is more of an incentive than getting your entire account (which obviously includes the rank) back.
And finally:
I refuse to believe as Ulanopo does that if permabanned RIOT "simply doesn't want you playing here". Because that's what you use IP bans to accomplish.
No. No you don't use IP bans to accomplish that. Partially because they don't really accomplish that, but mainly because you simply don't use IP bans. At all. Ever.
They're extremely easy to circumvent if you know what you are doing. But even if you don't know what you are doing, you might circumvent them without actually even intending to do so (my IP address automatically changes every 24 hours) and they have a huge risk of collateral damage if multiple people happen to play on the same IP address.
tl;dr: Risks and feasibility of a project need to be considered and can absolutely be valid reasons to dismiss it.
