Sorry, there are some cases where TCO functions are absolutely critical. Kill your system and make it worthless if you don’t do it critical. You just have to get it right.
The examples you show are all calls to functions which (eventually) return and in those cases TCO is a nice but not need. It may not save you heap but it will save you stack, which can actually save you GC as well[*].
No, the main case where is TCO is critical is in process top-loops. These functions never return (unless the process dies) so they will build up stack never to release it. Here you have to get it right. There are no alternatives. The same applies if you top-loop is actually composed of a set of mutually calling functions. There there are no alternatives. Sorry for pushing this again, and again, but it is critical. 
In other cases my way of choosing is the one which gives me the most intelligible code. And that varies depending on what the function is supposed to do and how I choose to do it.
Robert
[*] This is because the heap and stack share the same memory area and when it is full you get a gc. So keeping the stack small can delay gc.
