I have the feeling Lucky feat in 5e is overpowered.
I have a modification that I want to propose my players with says you have to decide to use your luck point BEFORE rolling the dice. That's the only modification so the feat will be:
You have 3 luck points. Whenever you make an attack roll, an ability check, or a saving throw, you can spend one luck point to roll an additional d20. You can choose to spend one of your luck points only before you roll the die and before the outcome is determined. You choose which of the d20 is used for the attack roll, ability check or saving throw.
You can also spend one luck point before an attack roll is made against you. Roll a d20, and then choose whether the attacker uses the attacker roll or yours. You choose the dice before the attack roll is made and, therefore, before knowing the attacker roll result.
The rest of the feat description (i.e. cancelling rules) remains unchanged.
I feel that makes the feat less powerful by turning a "I don't really like this result" to "I really want to make sure I get this right". That way the player needs to carefully decide what matters and what doesn't before rolling, afterwards is just too late.
Is this modification, however, underpowered to the point that it makes the feat useless (meaning the feat will not be picked up by players since its not worth it)?
Note: I'll accept frame challenges if you think the original lucky feat is not overpowered, although I think that may end up going to opinion based quickly.
Reasons I feel it is overpowered There are numerous discussions already on the internet about the Lucky feat. My take on it is that it allows way too many rerolls for any roll, after knowing the die roll (which means you at least have a feeling of whether you'll fail or not).
In my view this means that the players are able to, on demand, reroll at will any critical situation they find themselves on and those usually do not happen more than 3 times per day and when they happen, it is simply not that difficult to spot them. I don't usually put my players on such tight corners that they have more than 3 situations a day where the outcome of a single roll is of vital importance and the fact that you can decide after looking at the dice means that out of those critical situations, some of them would be naturally saved, allowing you to simply "save" the luck point. Then the next adventuring day, boom, you have your points again to start over with your safety net of, "if everything goes wrong I can count on my luck".
Making the players decide FIRST means that luck points are precious and valuable and you really have to think first, is this worth making absolutely sure I have the best chances to succeed or is it better to save it for later? Is this a really critical situation?
With deciding after they can instead go like: oh, it seems like I'm going to fail this thing... ok, I don't want to fail so I'm going to spend a luck point. It's a fallback, not a carefully thought tradeoff.
Even though the players do perform dozens of rolls per adventuring day, most of them do not matter so much. Attack roll, you fail, fine,next. Attack roll, you succeed, great. But then there's this occasional thing that really matters, "saving throw" against a fireball, that lockpick to enter the throne room, that deception check to escape from a contrieved situation, that last attack you've just failed with 5 HP left which you feel, if succeded may have just killed the monster... those are the ones that do change the course of the adventure. Yes, players roll a lot, but critical, potentially changing situations are far and between events that do not happen that often (at least not on my adventures).
I don't want to quote all the internet here but Lucky is consitently considered either broken or one of the most useful feats on the game, usually being quoted as THE most useful feat for all the above reasons.