Perhaps instead of spell versatility, it should just be one versatile spell known.
I could see tables doing something like that if they think this option would be a problem.
Personally
I don't think this is necessary. Versatility from being able to swap spells known is definitely going to suffer from diminishing returns. In other words, there are going to be 1-3 spells at most spell levels that most players will never want to sacrifice. That's why Wizards don't really complain, right? Everybody wants
fireball or
healing word or
charm person or
misty step or
shield, etc. You always pick the most valuable spells to know first, right? So there's not a lot of
de facto gameplay difference in allowing a PC to spend weeks to completely change their spell list. It's just not going to happen in most games. They're very, very unlikely to want to change the lion's share of their spells. Few actual players will want to do that, and if they do than something so catastrophic happened to that PC or in that campaign that that character was going to be useless unless they spent that week learning a new set of spells.
Think about your Clerics, Druids, and Wizards. How often do they exchange spells? How many do they exchange? IMX, it's 1 or 2 tops that change. Sure, you might then argue that, "one slot is too many then because it's maximum versatility gained and one is all you'll ever need," but I have to think that this optional rule just wouldn't be applied at those kinds of tables.
You would have your regular spells known, which are static except through leveled retraining. Then, separately from this, an extra "known spell" that can be changed daily. This would prevent the "change the entire repertoire" issue that some have with the UA-proposed feature. It could be changed to a different level slot as you gain more levels but is always just the one variable known spell. Number of known spells could be modified a bit but since it floats at any level, not sure how much to modify it, if at all.
Thinking about it, I like this idea less.
I think it conversely provides
more versatility for highest level spells because it allows you to chose a spell of any level essentially at the cost of a 1st level spell known. Normally, when you hit level 5, you learn 1 new spell (certain to be 3rd level) and you can lose a 1st or 2nd level spell to learn a second 3rd level spell. So you're stuck with two 3rd level spells max. With as-written spell versatility, you never get to a third 3rd level spell known until level 6. If you set aside one spell slot as "versatile", then you can get to a third 3rd level spell slot: One 3rd level spell from level, one from swapping, and one from versatility.
Unless you introduce a table to control spells known by spell level you can't prevent this from happening. Worse, if you just stop the highest known spells from being versatile, then you subvert the primary purpose of the option: allowing players to correct mistakes or experiment with their spell selection. If you can't experiment with 3rd level spells until level 6 or 7, you're really failing to accomplish the goal your rule is intended to accomplish.
I suppose you could eliminate the core swap-spell-on-level-up, but now, again, you're eliminating the ability to correct mistakes.
I think the as-written rule in UA is the overall best version of this rule I've seen so far. Indeed, the more I think about it, the fewer problems I have with it.
I also really like the suggestion in the other thread to keep it spell level 5/6 or lower, as a more simple solution, alongside allowing daily preparers to prepare their cantrips daily the same way they prepare spells of 1st level or higher.
Eh. I don't really care about 7th to 9th level spells. First, they don't come up in 90% of games. Second, I would never implement this kind of limitation without seeing it in actual play first to determine if it was actually necessary. Spell power in 5e is a
lot lower than it was in prior editions. And again, it's not a problem for Wizards, Clerics, or Druids to do it. It will take some convincing to get me to agree out-of-hand that it's automatically a problem for Sorcerers or Bards.