Totally agree!

by current standards, immortal would come out of beta at most at 3 kyu. Which says a lot about the current standards...

(1) and (3) are overrated, and (1) should not have been translated to Python ( or at least certain imports should have been forbidden ).

I don't have an opinion on (0) or (2) because I've never been able to solve them.

Is it unreasonable to rank a kata on its simplest possible solution

I think so, yes. There are solutions to Become Immortal which make the kata look like it could have been 6kyu. There are many other kata whose solutions can be much simpler than their rank might imply.

Having said that, even the naive approach to this kata should be able to solve it with just a a few loops and simple variables. No advanced algorithms are needed, there's no hidden information. There are tools that make it easier if you know about them, but even without them I think this should be a 6kyu at best, and I think 7kyu is fine.

Is it unreasonable to rank a kata on its simplest possible solution, instead of ranking it according to how difficult some people make it for themselves?

There is a simple ( but elegant! ) pattern here, which can also be expressed simply ( but elegantly :) if you have the tools.

Availability of tools is also a reason why a task can have different difficulties in different languages. In Python, I had to massage the available tools a little. In JS, I had to partially write my own. In Haskell, this is an absolute no-brainer because the tools exist out of the box. Kata can't have different rankings in different languages though.

Anyone can partake in beta testing (ranking and voting on beta kata's), so why do I only see 1 kyu (+) users there ...

sure this could be 6 imo, but average vote during beta was in clear favour of 7

5 kyu for a simple algorithm? Should flooding a maze (5 kyu) then be purple?

Fair 6 kyu, or an easy 5kyu.

I estimated 4 kyu 🤷‍♂️