Do you have evidence of choice paralysis? Did you run play tests this?
Let us assume you did test it, and found this result...
I will not suggest to hide characters. That can also lead to some other predatory practices (loot-boxes are not the only bad thing out there). In fact, many people enjoy shopping. If the decision is hard, it could spark a community of character reviews in social media, that is all good.
Furthermore, I would like to encourage the philosophy that players should enjoy buying. You want them to like to buying things in your game. So spice up the market!
Besides pure user experience, you can address that choice paralysis with non-premium version that can be acquired by other means. Sure, they would try to get the non-premium first, but it also means they will be buying something they know they like. That is good customer satisfaction. Perhaps you can limit the number of non-premium character they can have at a time as a side incentive.
And of course the premium version would be extra fancy... they may even have slightly better stats, never to the point where the non-premium version are not usable. You want them to enjoy the non-premium so they know they will enjoy the premium one.
Addendum: to make sure you are not making a pay to win. The improvement should not be in power, just in convenience.
Which remind me. You want differences in kind, the characters should differ in mechanics, not just stats. There should be no clear better character, getting a character should feel game changing, and you should offer characters that lead well to different character styles (ranged, melee, stealth, etc...). That last point should also help the player decide what to buy.
If you are thinking in offering multiple variants with different stats... Instead provide ways to alter the stats of the characters (leveling up, power up items, collectible cards, whatever).
Also do not forget to offer alternative skins for sale. Let the player express themselves with the customization. In fact, in multiplayer, that is advertisement! If a player can see another player with a cool cosmetic, they might want to get it too.
Addendum: I understand the argument against the premium version having better stats. It easy to make it so that player with premium versions dominate. I suppose it works better in some games, for instance if the game is cooperative instead of competitive. Another things that helps is having things that can only be acquired by playing, and let those be tradable with other players, plus allowing trading premium currency among players. That way, if you put the time you can still get the premium content despite not using real money by selling things to those who do.
However, please understand the argument against hiding characters and rotations. It is easy to design rotations in such way that it prays on the fear of missing out. Not to mention making it so that everything in the rotation is not an easy combination of the amounts of premium currency that can be acquired... making it so that you always need just a little more premium currency, but you got to buy a bundle bigger than what you actually want to expend... leaving you with little extra that sets you to need just a little more premium currency on the next rotation.
pack
. That would bump the average sale price if they want a few of the 5 characters in each pack, for example. The user still gets to choose without randomness. \$\endgroup\$