If the goal is to avoid save-scumming and reloading to get only successful rolls, then it is not sufficient to only hide the DC. You need to hide the DC, and ALSO make it so that succeeding the roll can be the "bad" option, while failing the roll can be the "good" option, say, 15-25% of the time.
Huh? What's a difference if the success and failure will be switched around? You rolled, you get bad outcome you try again. I don't think it is relevant if you got bad result because you succeeded, or you got bad result because you failed.
The problem at the heart is very simple - player's get one chance to attempt to succeed at something, and if they don't they have easy (but unenjoyable) way to remedy it. Player's in majority of cases are invested in succeeding as otherwise they would be unlikely to pick that option to begin with. Even if there is an elaborate (and enjoyable) option B for when a player fails, player is still unlikely to be satisfying because that's not what they opted to do.
Set back or failure can lead in a long run to a more enjoyable experience, but initial feedback will always be one of disappointment as simply the player didn't get what they wanted to get when they pressed the button. I don't think they are many options, beyond:
1) devs trusting players to choose the experience that will be most fun for the players (reload or stick to the choices)
2) devs deciding that sticking with choices is a superior experience and enforce the choice - either by making it a flat, unmovable check or saving RNG seed like in FiraxCOM1 and I think Solasta - where after reload the roll is still the same.
