Lucky: Underrated for Wizards who copy spells from scrolls

[Merudo]

The "Lucky" feat is criminally underrated by Wizard guides.

Treantmonk (https://docs.google.com/document/d/1IeOXWvbkmQ3nEyM2P3lS8TU4rsK6QJP0oH7HE_v67QY) doesn't even list it in his guide, and RPGBOT (https://rpgbot.net/dnd5/characters/classes/wizard/) lists it as orange ("OK").

Lucky should be rated as amazing for all Wizards though. Why? Because it helps tremendously at copying spells successfully from scrolls.

Indeed, according to RAW (DMG p.200), copying scrolls is highly likely to fail:


When a spell is copied from a spell scroll, the copier must succeed on an Intelligence (Arcana) check with a DC equal to 10 + the spell's level. If the check succeeds, the spell is successfully copied. Whether the check succeeds or fails, the spell scroll is destroyed.

This means:

- A level 5 Wizard with 18 intelligence has a 25% chance of failing to copy a level 3 spell, destroying the scroll for nothing.
- A level 9 Wizard with 20 intelligence has a 25% chance of failing to copy a level 5 spell.
- A level 15 Wizard with 20 intelligence has a 35% chance of failing to copy a level 8 spell.
- A level 20 Wizard with 20 intelligence has a 35% chance of failing to copy a level 9 spell.

Pretty bleak odds if you ask me.

If the Wizard takes the Lucky feat and is prepared to reroll the Arcana roll, the 25% chance of a failed copy turns into a 6.25% chance. That's much easier to deal with.

But wait, there is more! The Wizard could also pull an all nighter before coping the spell, earning an exhaustion level. Now they have disadvantage on all checks, including the Arcana check to copy the spell. When they fail the copy, they can use Lucky to turn the disadvantage into triple advantage. The chance of a failed copy goes from 6.25% to 1.56%. Awesome!

[ProsecutorGodot]

Consider it a side effect of this clause being found in the description of Spell Scrolls rather than in the Wizard's Spellbook sidebar like it should be. To be honest, for as long as I've been playing 5E now I wasn't even aware of this rule because why would I look in the description of Spell Scroll for my Spellbook rules?

The rule is just a bit ridiculous to me as well, not because I find it inherently unbalanced or anything but because for some reason it's specific to Wizard spells. Other classes are somehow able to copy ritual scrolls that aren't found on the wizard spell list with no difficulty into their Ritual Caster book (ditto for Pact of Tome Warlocks) and this strikes me as an oversight.

[LudicSavant]

The rule is just a bit ridiculous to me as well

It seems to be a result of them basically copy-pasting a line about scrolls from a previous edition without respect for the fact that the context has changed (it made more sense in the original context).

[MaxWilson]

The "Lucky" feat is criminally underrated by Wizard guides.

Treantmonk (https://docs.google.com/document/d/1IeOXWvbkmQ3nEyM2P3lS8TU4rsK6QJP0oH7HE_v67QY) doesn't even list it in his guide, and RPGBOT (https://rpgbot.net/dnd5/characters/classes/wizard/) lists it as orange ("OK").

Lucky should be rated as amazing for all Wizards though. Why? Because it helps tremendously at copying spells successfully from scrolls.

Indeed, according to RAW (DMG p.200), copying scrolls is highly likely to fail:

This means:

- A level 5 Wizard with 18 intelligence has a 25% chance of failing to copy a level 3 spell, destroying the scroll for nothing.
- A level 9 Wizard with 20 intelligence has a 25% chance of failing to copy a level 5 spell.
- A level 15 Wizard with 20 intelligence has a 35% chance of failing to copy a level 8 spell.
- A level 20 Wizard with 20 intelligence has a 35% chance of failing to copy a level 9 spell.

Pretty bleak odds if you ask me.

If the Wizard takes the Lucky feat and is prepared to reroll the transcribe roll, the 25% chance of a failed copy turns into a 6.25% chance. That's much easier to deal with.

But wait, there is more! The Wizard could also pull an all nighter before coping the spell, earning an exhaustion level. Now they have disadvantage on all checks, including the Arcana check to copy the spell. When they fail the copy, they can use Lucky to turn the disadvantage into triple advantage. The chance of a failed copy goes from 6.25% to 1.56%. Awesome!

Eh.

(1) It's highly DM-dependent whether you're going to find spells in treasure as scrolls or spellbooks. Logically spellbooks should be more common because they are orders of magnitude cheaper to produce, especially for higher-level spells.

(2) You can gain advantage from other sources including friendly Enhance Ability spells.

(3) Disadvantage ==> tri-vantage is stupid and should not be used. WotC may agree because the latest PHB errata (https://media.wizards.com/2018/dnd/downloads/PH-Errata.pdf) for page 173 arguably make this trick illegal: Lucky lets you reroll or replace one of the disadvantage d20s, not both of them.

But the fundamental issue with this trick is that reverse-engineering spells from scrolls instead of spellbooks is expensive, inefficient, and not something that logically should be happening in the game world, so not something a wizard build should be assuming will happen unless you know your DM is different.

[ZorroGames]

Interesting. Never seen that in AL play. I have a Gnome Wizard/Tempest Cleric who seems to get the NPC spell books because other Wizards were never present when we defeated NPC arcane casters.

[MaxWilson]

Interesting. Never seen that in AL play. I have a Gnome Wizard/Tempest Cleric who seems to get the NPC spell books because other Wizards were never present when we defeated NPC arcane casters.

It's not needed for spellbooks, only for reverse-engineering from spell scrolls.

[Lunali]

Consider it a side effect of this clause being found in the description of Spell Scrolls rather than in the Wizard's Spellbook sidebar like it should be. To be honest, for as long as I've been playing 5E now I wasn't even aware of this rule because why would I look in the description of Spell Scroll for my Spellbook rules?

The rule is just a bit ridiculous to me as well, not because I find it inherently unbalanced or anything but because for some reason it's specific to Wizard spells. Other classes are somehow able to copy ritual scrolls that aren't found on the wizard spell list with no difficulty into their Ritual Caster book (ditto for Pact of Tome Warlocks) and this strikes me as an oversight.

An additional effect of this rule, spell scrolls are consumed when copied into wizard books, but not when copied into ritual or warlock books.

[Merudo]

Eh.

(1) It's highly DM-dependent whether you're going to find spells in treasure as scrolls or spellbooks. Logically spellbooks should be more common because they are orders of magnitude cheaper to produce, especially for higher-level spells.


Pretty much all the published adventures have some spell scrolls to be found.



(2) You can gain advantage from other sources including friendly Enhance Ability spells.


Enhance Ability lasts 1 hour. It can only work for copying level 1 spells from the school from which you are a savant.



(3) Disadvantage ==> tri-vantage is stupid and should not be used. WotC may agree because the latest PHB errata (https://media.wizards.com/2018/dnd/downloads/PH-Errata.pdf) for page 173 arguably make this trick illegal: Lucky lets you reroll or replace one of the disadvantage d20s, not both of them.


Nope. From the Sage Compendium (https://media.wizards.com/2019/dnd/downloads/SA-Compendium.pdf):


The Lucky feat is a great example of an exception to a general rule. The general rule I have in mind is the one that tells us how advantage and disadvantage work (PH, 173). The specific rule is the Lucky feat, and we know that a specific rule trumps a general rule if they conflict with each other (PH, 7).

[MaxWilson]

Enhance Ability lasts 1 hour. It can only work for copying level 1 spells from the school from which you are a savant.

It can be recast.

Re: Sage Advice, halfling Lucky trait is also "specific" and yet it explicitly is meant to interact with page 173. The "which rule is more specific?" meta rule just causes arguments, not clarity. It's lazy writing.

[ProsecutorGodot]

It can be recast.
Playing a bit of devil's advocate here but does it even need to be recast? Sure, it's almost certainly intended that you're supposed to role with modifiers you would keep for the duration of the copy process but it doesn't actually say that in the rules.

It's bordering on metagaming but you only need to be under the effects of Enhance Ability at the time the skillcheck is rolled, which in my opinion, since we're given no alternative makes the most sense to be made at the end of the process to determine if your time spent was successful.

[MaxWilson]

Playing a bit of devil's advocate here but does it even need to be recast? Sure, it's almost certainly intended that you're supposed to role with modifiers you would keep for the duration of the copy process but it doesn't actually say that in the rules.

It's bordering on metagaming but you only need to be under the effects of Enhance Ability at the time the skillcheck is rolled, which in my opinion, since we're given no alternative makes the most sense to be made at the end of the process to determine if your time spent was successful.

As a DM I'd say the skill check is made over the entire span of time you're engaged in the activity, and you get only bonuses which apply to the whole time. That's why I wouldn't mention Guidance here: it's just not plausible to keep it up the whole time.

[Keravath]

This is another minor reason why some wizards start as a level 1 knowledge cleric. Expertise in the arcana skill goes along way towards making it trivial to copy spells from any source.

A level 18 character, 1 knowledge cleric/17 wizard can copy Wish into his spell book with a DC 19 copy check. However, they will have a total of +17 due to expertise in the arcana skill resulting in only a 5% chance of failure to copy even the highest level spells in the game. At level 5 they will have +10 arcana (with 18 int) which makes it impossible to fail to copy level 1 spells and even level 3 spells will only fail 10% of the time.

I think one area where the Lucky feat might be underestimated for Wizards would be counterspell checks.

[MaxWilson]

I think one area where the Lucky feat might be underestimated for Wizards would be counterspell checks.

It's not bad for initiative rolls either, in a pinch.

[kazaryu]

The "Lucky" feat is criminally underrated by Wizard guides.

Treantmonk (https://docs.google.com/document/d/1IeOXWvbkmQ3nEyM2P3lS8TU4rsK6QJP0oH7HE_v67QY) doesn't even list it in his guide, and RPGBOT (https://rpgbot.net/dnd5/characters/classes/wizard/) lists it as orange ("OK").

Lucky should be rated as amazing for all Wizards though. Why? Because it helps tremendously at copying spells successfully from scrolls.

!

i think you're overestimating how useful it is to copy spell scrolls. i mean, yes its useful. but a feat that only benefits a small chunk of what a wizard does is not 'amazing'. (yes i know, lucky isn't just used for spell scrolls, but those and dispel/counterspell checks are the only benefits specific to wizard. and wizards generally make fewer skill checks than alot of other classes.

compare it to feats like warcaster or resilient (con). which have a significant impact on an entire pillar of play. its a fun interaction, to be sure, but its by no means 'underrated.

also consider: how often do treantmonks guides use spells copied from books?

[Merudo]

compare it to feats like warcaster or resilient (con). which have a significant impact on an entire pillar of play.

Lucky obviously has "significant impact" on all pillars of play, and can substitute for War Caster for concentration purposes.

[Eriol]

Clockwork amulet + proficiency in Arcana + wizard intelligence = guaranteed to succeed your whole career, once a day at a minimum. For the higher-level ones, once a day is all you can do anyways.

[kazaryu]

Lucky obviously has "significant impact" on all pillars of play,

not generally for a wizard. wizards contributions to social and exploration pillars of play are almost entirely based on spells. which either 'just work' (mostly in exploration) or they require a save from someone else. neither of which can be influenced by lucky. now, you can build a wizard with some social or exploration skills. but in most adventuring parties they're not going to be the go to for those.

that leaves combat: wizards only infrequently make skill checks in combat. which leaves attack rolls (again, not common, although not impossible. also, if you're making an attack roll, chances are its not something you'll wanna burn luck on. i.e. firebolt.) and saving throws.



and can substitute for War Caster for concentration purposes.

in a pinch, its not *as* good. but it is useable i'll grant you. obviously, how useful it is depends entirely on how often the DM tosses damage your way. regardless warcaster is far more reliable, and also has other useful benefits.

[Aaron Underhand]

I agree lucky is useful in this case, but a lot of people play divination wizards, who can copy guaranteed

[Gignere]

I agree lucky is useful in this case, but a lot of people play divination wizards, who can copy guaranteed

Some of us play lucky diviners. It is glorious amount of dice manipulation.

[ZorroGames]

It's not needed for spellbooks, only for reverse-engineering from spell scrolls.

Right but it just is weirder because of that quirk/anomaly.

[ZorroGames]

i think you're overestimating how useful it is to copy spell scrolls. i mean, yes its useful. but a feat that only benefits a small chunk of what a wizard does is not 'amazing'. (yes i know, lucky isn't just used for spell scrolls, but those and dispel/counterspell checks are the only benefits specific to wizard. and wizards generally make fewer skill checks than alot of other classes.

compare it to feats like warcaster or resilient (con). which have a significant impact on an entire pillar of play. its a fun interaction, to be sure, but its by no means 'underrated.

also consider: how often do treantmonks guides use spells copied from books?

The original guide was written in 2015 and years later in his subclass builds he pointed out he had changed his views on several aspects of spells, etc.,

Edit - he clearly states in those builds that he assumes a worst case scenario of no spellbooks as treasure or scrolls available for the build progression. A build as a framework you can hang other rewards on.

[Keravath]

Lucky obviously has "significant impact" on all pillars of play, and can substitute for War Caster for concentration purposes.

In actual play, I have found Lucky to be far less useful than some folks seem to expect. There are so many better and more reliable feats that taking Lucky is just a last ditch "let's pick up a feat because there isn't anything better" choice.

Lucky's downfall is that it could be useful for any character. Re-rolls are always handy. However, you get to re-roll three times a day and that is it. I've seen folks use up Lucky on to hit rolls, "Yay I hit", then fail a hold person save later in the day because they used all the re-rolls. That didn't go well. They might have failed the hold person even with Lucky but saving the usage for important rolls becomes the deciding factor in using Lucky.

In addition, if you need to roll a 15 to save and fail and use Lucky, you still have only a 25% chance on the second roll, the same as the first. Yes, it helps, and can make a difference but more often than not it doesn't.

Between the limited uses and saving those re-rolls for dice rolls that REALLY matter, I don't find Lucky to be that great in actual play.

On the other hand, if the DM runs a 1 fight/day game and then long rest to recover resources, the character with Lucky can afford to use all the uses in one fight and that will have a more noticeable effect.

[kazaryu]

Some of us play lucky diviners. It is glorious amount of dice manipulation.

ahh, a man of culture i see. xD



he clearly states in those builds that he assumes a worst case scenario of no spellbooks as treasure or scrolls available for the build progression. A build as a framework you can hang other rewards on.


right. in most cases their ability to copy spells from scrolls/spell books is going to act more like a ribbon ability than a mainstay feature. it comes up every so often, but wizards get so many spells just from leveling that this feature isn't likely to make or break a build. obviously, as stated, DM dependent.

[Merudo]

I've seen folks use up Lucky on to hit rolls, "Yay I hit", then fail a hold person save later in the day because they used all the re-rolls. That didn't go well. They might have failed the hold person even with Lucky but saving the usage for important rolls becomes the deciding factor in using Lucky.

In addition, if you need to roll a 15 to save and fail and use Lucky, you still have only a 25% chance on the second roll, the same as the first. Yes, it helps, and can make a difference but more often than not it doesn't.

Between the limited uses and saving those re-rolls for dice rolls that REALLY matter, I don't find Lucky to be that great in actual play.


I won't deny that Lucky is a poor choice for those who strategize poorly. It is by far the feat that relies the most on the player's intelligence.

Lucky is best used to cancel out events that are both awful and unlikely.

I personally can't imagine using Lucky on a simple hit roll, unless said hit would result in the target getting killed before it can act again. Missing just isn't that big of a deal.

I would also be reluctant to use Lucky for a roll I have a high chance of failing again. A 25% of success is quite low; it's rather not use a Luck point on it unless failing it would have dramatic consequences, in which case Lucky could save my life.

Lucky is best used against unfortunate rolls that have a low chance of occurring again on a reroll: getting hit by a Critical Hit, failing a DC10 Concentration roll, failing an ability check with a low DC, etc. In these cases you spend a Luck point for a high chance to cancel out the bad outcome.

A consequence of this is that Lucky is best for builds for which "awful and unlikely" events occur often. For example, the Bladesinger engaging in melee with AC28 will dodge near anything but Critical Hits; having Lucky to cancel out these criticals greatly improve the survivability of the class.

Similarly a spellcaster with a +7 to CON saving throws will fail DC10 concentration roll when rolling a 1 or 2; being able to cancel those failed rolls is a wonderful use of Lucky.


Lucky [...] can substitute for War Caster for concentration purposes.



in a pinch, its not *as* good. but it is useable i'll grant you. obviously, how useful it is depends entirely on how often the DM tosses damage your way.

It also depends on the CON bonus compared to the damage received.

If you have a +6 to CON saving throws (con 16, con saves proficiency, level 5-8) you'll fail about 15% of the DC10 concentration rolls. Suppose you spend all your Luck rolls on rerolling these failed concentration rolls. On average, you'll have to take damage 26 times before you fail four times, in which case Warcaster will be preferable to Lucky.

With +6 to CON saving throws, Lucky is far more desirable than Warcaster, at least to keep concentration.

[MaxWilson]

I won't deny that Lucky is a poor choice for those who strategize poorly. It is by far the feat that relies the most on the player's intelligence.

I agree that Lucky is rotten for those who strategize poorly (and I also agree with those who say it's more of a good "might as well" pick than a strong first choice). However,
Lucky is not a clear frontrunner when it comes to relying on a player's intelligence. Mobile feat, for example, comes to mind as a feat which is terrible when used poorly and fantastic when used well. Ditto Prodigy and Skulker.


Lucky is best used to cancel out events that are both awful and unlikely.

Yes, exactly! Fully agree with this. One of the things that makes it a good Warcaster-lite is when you use it to cancel crits that could potentially cause more-difficult concentration save DCs which would be likely instead of unlikely to break your concentration.


I personally can't imagine using Lucky on a simple hit roll, unless said hit would result in the target getting killed before it can act again. Missing just isn't that big of a deal.

I would also be reluctant to use Lucky for a roll I have a high chance of failing again. A 25% of success is quite low; it's rather not use a Luck point on it unless failing it would have dramatic consequences, in which case Lucky could save my life.

Again, yes. Only if the consequences for failure are dire, and they're usually not as dire as it feels like emotionally, unless your allies are outnumbered and already losing the fight, or you're fighting a disintegrate-on-kill monster like a beholder or a Nightwalker.


A consequence of this is that Lucky is best for builds for which "awful and unlikely" events occur often. For example, the Bladesinger with AC28 with shield will dodge near anything but Critical Hits; having Lucky to cancel out these criticals is extremely valuable.

Well, maybe. A critical is unlikely for this PC, but is it awful? Depends on what you're fighting. I assume by "extremely valuable" you still mean "should be used sparingly." Sometimes it's better to take 20 HP of damage than to spend a Lucky die.


Similarly a spellcaster with a +7 to CON saving throws will fail DC10 concentration checks when rolling a 1 or 2; being able to cancel those failed checks is a wonderful use of Lucky.

Yes, although again, sometimes it's better to just let the spell drop and save the Lucky die, depending on the tactical situation and how close you are to wrapping up the fight already.

[Merudo]

Lucky is not a clear frontrunner when it comes to relying on a player's intelligence. Mobile feat, for example, comes to mind as a feat which is terrible when used poorly and fantastic when used well. Ditto Prodigy and Skulker.

When I said Lucky is the feat that relies the most on a player's intelligence, I had in mind a somewhat narrow view of intelligence: "analytic" intelligence (solving optimization problems / making decisions).

I do agree that other feats better tap other aspect of intelligence (especially "creative intelligence", or coming up with novel ways to use a feat).

I can see Skulker being fantastic for a player creative enough to make good use of light obscuration. I don't really see that for Prodigy and Mobile, though.

I especially don't see it for Prodigy. That's probably because I value skill proficiency poorly: beside Perception, Athleticism and Stealth, in my experience skills ultimately have little impact on gameplay. At low levels, the +2/+3 (+1/+2 for Bard) you get from skill proficiency is just too little to matter, and at high levels, spells essentially replace skills.

As for Mobile, it's a good feat that can be employed well, but I don't really it relying on intelligence except for a some very specific builds (off the top of my head: the Ancestral Barbarian, the Moon Druid, and some Monks). I like getting it for Fighters / Paladins in heavy armor that don't have 15 STR. My main issue with the feat is that the "free" disengage requires a melee attack, and most classes who would like to disengage don't want to do melee attacks. So the fancy footwork you'd like to do with the feat is heavily restricted because the enemies can just ignore you and attack the squishies. And if it's the +10 to speed that you want, just cast Longstrider.

I think Actor is arguably the feat that relies the most on a player's creativity. With it, all kinds of crafty deceptions become possible. It definitely should be talked about more.

[MaxWilson]

When I said Lucky is the feat that relies the most on Intelligence, I had in mind a somewhat narrow view of intelligence: "analytic" intelligence (solving optimization problems / making decisions).

I do agree that other feats better tap other aspect of intelligence (especially "creative intelligence", or coming up with novel ways to use a feat).

I can see Skulker being fantastic for a player creative enough to make good use of light obscuration. I don't really see that for Prodigy and Mobile, though.

I especially don't see it for Prodigy. That's probably because I value skill proficiency poorly: beside Perception, Athleticism and Stealth, in my experience skills ultimately have little impact on gameplay. At low levels, the +2/+3 (+1/+2 for Bard) you get from skill proficiency is just too little to matter, and at high levels, spells essentially replace skills.

Just like Mobile and Skulker, it takes a fair amount of mathematical, quantitative intelligence to know when is the appropriate time to switch tactics: should you be Action Surging a bunch of attacks right now, or grapple/proning the biggest enemy (via Prodigy: Athletics + Extra Attack) to take it mostly out of play NOW even though it will mean it takes more rounds until it dies? It depends on the numbers, and having a good gut feel for numbers will help you recognize a good tactic for the situation vs. a bad tactic.

I just recently saw a battle where the Mobile Crossbow Expert Sharpshooter EK made a serious mistake: when faced with two or three Githyanki Warriors, instead of relying on Mobile or Dodge to get him out of trouble and hopping down from the ledge he was on into the nest of Giant Poisonous Snakes conjured up by other PCs, he fired off three crossbow bolts hoping to bring down one of the Githyanki, and wound up getting seriously wounded over the next couple of rounds of combat. (Took several hits including a Githyanki crit for 40+ HP of damage!) Doing the math afterwards, it turned out that wasn't even very unlikely--he just made a bad decision. (This was a practice battle I was running against myself for fun BTW, but that doesn't change the fact that it was a dumb tactic which better intelligence would have prevented.)

Lesson learned: regardless of your individual superiority to an individual opponent, seek first to achieve an advantageous overall tactical position which effectively leverages your allies. Don't gamble on finishing the whole fight yourself this round.


As for Mobile, it's a good feat that can be employed well, but I don't really it relying on intelligence except for a some very specific builds (off the top of my head: the Ancestral Barbarian, the Moon Druid, and some Monks). I like getting it for Fighters / Paladins in heavy armor that don't have 15 STR. My main issue with the feat is that the "free" disengage requires a melee attack, and most classes who would like to disengage don't want to do melee attacks. So the fancy footwork you'd like to do with the feat is heavily restricted because the enemies can just ignore you and attack the squishies.

I think Actor is arguably the skill that relies the most on a player's creativity. The feat makes it possible to pull off all kinds of crafty deceptions. I definitely think it should be talked about more.

Knowing when to throw a couple of punches for a free pseudo-Disengage, even at the cost of damage this round, is a mark of intelligence.

P.S. I hate fighting githyanki. :) They are so mobile!

[Pufferwockey]

Not saying OP is wrong, but the idea that Lucky, which is widely considered to be so good that many dms ban it from their tables, may still be unstated, is pretty funny

[MaxWilson]

Not saying OP is wrong, but the idea that Lucky, which is widely considered to be so good that many dms ban it from their tables, may still be unstated, is pretty funny

Many DMs ban it from their tables?

[ProsecutorGodot]

Many DMs ban it from their tables?

In my experience, we had a knee jerk reaction of banning it from the table as a result of very loud complaints we saw on forums from a small majority of people. We don't ban it at our table anymore and most players are more interested in other feats or a straight up ASI.

[Merudo]

Many DMs ban it from their tables?

A few years ago, the hate for Lucky was omnipresent.

[stoutstien]

A few years ago, the hate for Lucky was omnipresent.

I do remember when 5e was just out lucly was considered one of the strongest feats and now it's in the 'ok but not amazing' group and feats like inspiring leader and healer went from poor picks to solid for most groups.

It's fun to see which options stand the test of long-term play time.

[Keravath]

Not saying OP is wrong, but the idea that Lucky, which is widely considered to be so good that many dms ban it from their tables, may still be unstated, is pretty funny

Personally, I've never banned it nor would consider doing it in any game I run. I have also never played at a table where the Lucky feat was banned. I've seen quite a few posts with opinions on how great and overpowered it is but it has NEVER played out like this in any game I have been involved in. I have the feat on one rogue character (out of about 20 different characters) since it fit the character concept and it has been useful on occasion but not often.

I play it (and have seen it played) much the same as MaxWilson and Merudo have described. People save the Luck re-rolls for situations that matter and as a result entire sessions can go by without any Luck points being used. I have the impression that those who seem to consider it overpowered are in groups that have one major encounter/day and that the party knows that so they feel free to use all their resources in the 3 or 4 rounds of combat that they will have that day. In this circumstance, three re-rolls might feel quite powerful. However, I've never played in a campaign like that or run one that way. If someone burned all their resources in the first encounter, they would find themselves struggling a bit later in the day on the fourth or fifth event of the day.

This is campaign and situation dependent of course, some days may be short with one or two events while others could be quite long with [rarely] 8-10 encounters though they wouldn't usually all be combat encounters. The point is the players do not know what kind of day it will turn out to be so they have to think about how they want to allocate resources.

Either way, in the typical games I play, Lucky is far from the best choice of feat, though it is a good generic choice if you don't need an ASI or it fits a character concept.

[kazaryu]

I do remember when 5e was just out lucly was considered one of the strongest feats and now it's in the 'ok but not amazing' group and feats like inspiring leader and healer went from poor picks to solid for most groups.

It's fun to see which options stand the test of long-term play time.

honestly, i feel like a big part of this was Critical role. Liam o'brian used luck really well, and with the players generally overrreactions (not meaning this in a demeaning way, but c'mon, the cast is an excitable bunch) it made the feat seem like it was a bigger deal. couple that with general confimartion bias (people remembering when luck worked out well, more frequently than when the luck re-roll didn't matter) and it makes sense that it became sensationalized for a while.