DM question on double smite.

[MThurston]

Would you allow a warlock/paladin to use Eldritch Smite and Smite on the same hit?

I can see by Rules as Written it is allowed. I however am not sure if I would allow it.

[HappyDaze]

Would you allow a warlock/paladin to use Eldritch Smite and Smite on the same hit?

I can see by Rules as Written it is allowed. I however am not sure if I would allow it.

I would not allow it.

I also do not allow that multiclass combination.

[stoutstien]

It not really a big deal. ES is lv +5 in warlock so they don't have any high level slots to fuel smite in exchange until late tier 3.

[BloodSnake'sCha]

I will allow it. It is a huge resource dedicated for a single combo that is just a damage spike.

It is a lot of damage but with a very heavy price.

It is a great way to make the party to use resources (the goal of an encounter).

[OldTrees1]

Worst comes to worst they are Paladin 2 / Warlock 9

2nd level: Smite costs 1 1st level slot, Deals +2d8 damage, twice per long rest
3rd level: Smite costs 1 1st level slot, Deals +2d8 damage, once per short rest
4th level: Smite costs 1 1st level slot, Deals +2d8 damage, twice per short rest
5th level: Smite costs 1 2nd level slot, Deals +3d8 damage, twice per short rest
7th level: Smite costs 2 3rd level slots, Deals +8d8 damage, once per short rest
9th level: Smite costs 2 4th level slots, Deals +10d8 damage, once per short rest
11th level: Smite costs 2 5th level slots, Deals +12d8 damage, once per short rest

So the cost to use the ability skyrockets until it is consuming nearly all of their spellcasting and limiting it to one attack per short rest. In return it is basically the same as the Rogue hitting on 2 turns. So by turn 2-3 the Rogue is better. And the Rogue can do it every combat. So if there are 2 combats per short rest, the Rogue is already better on turn 1.

Yes, I would allow it, but I would caution against it because it is a poor strategy. A different invocation, or single classed warlock is a better plan than double smiting.

Different invocation
2nd level: Smite costs 1 1st level slot, Deals +2d8 damage, twice per long rest
3rd level: Smite costs 1 1st level slot, Deals +2d8 damage, once per short rest
4th level: Smite costs 1 1st level slot, Deals +2d8 damage, twice per short rest
5th level: Smite costs 1 2nd level slot, Deals +3d8 damage, twice per short rest
7th level: Smite costs 1 3rd level slot, Deals +4d8 damage, twice per short rest
9th level: Smite costs 1 4th level slot, Deals +5d8 damage, twice per short rest
11th level: Smite costs 1 5th level slot, Deals +6d8 damage, twice per short rest
Basically the equivalent benefit but with 1 more free invocation slot.

Single classed warlock
5th level: Smite costs 1 3rd level slot, Deals +4d8 damage, twice per short rest
7th level: Smite costs 1 4th level slot, Deals +5d8 damage, twice per short rest
9th level: Smite costs 1 5th level slot, Deals +6d8 damage, twice per short rest
Same benefit, earlier level.

[Lupine]

OldTrees1, I think your numbers are off. Warlocks and paladins don't get sixth level spell slots. Their slots cap at fifth. Warlocks get mysic Arcanum instead of sixth and higher levels spell slots, allowing access to those spell choices. But the Mystic Arcanum cannot be used as spell slots: it's a once per long rest class feature.

[OldTrees1]

OldTrees1, I think your numbers are off. Warlocks and paladins don't get sixth level spell slots. Their slots cap at fifth. Warlocks get mysic Arcanum instead of sixth and higher levels spell slots, allowing access to those spell choices. But the Mystic Arcanum cannot be used as spell slots: it's a once per long rest class feature.

There was a typo (meant to write 5th at Pal2/Warlock9) but the math was right. Smites are (1+spell level)d8 per slot.

[JNAProductions]

I would not allow it.

I also do not allow that multiclass combination.

Why not?

RAW, it's fine.
RAI, pretty sure that's it.
RAF, I see no issues.

It's a big resource expenditure for big damage.

[Dork_Forge]

The rules allow it and Eldritch Smite has it's own built in limiters to stop it getting out of hand. If they want to feel powerful for two hits a short rest then all the power to them.

[Mikal]

No reason not to. Heck let em have a smite spell as well. If they’re using that many resources it’s for a reason, even if that reason is bragging rights. Whatever, let the players have their fun. They may regret spending it on overkill when they realize more enemies are around the corner

[MaXenzie]

RAW allows it since there's no clarification saying only one kind of smite can be used during an attack.

A Pallock can stack Smite, Eldritch Smite, and a spellcast Smite on the same attack if they want to go supernova.

[Chronos]

A paladin/warlock in general is fine, but not when the warlock side is hexblade. Hexblade is way overpowered as a paladin dip.

[sithlordnergal]

If you're just going RAW and RAI, yes they can double smite. Heck, hand that build a Rod of Lordly Might and not only can they paralyze someone with that same strike, but they can also force a Con save for 4d6 Necrotic Damage on the same strike, or if they have a Staff of Power they can add 1d6 Force damage. And they can do it for as many hits as they have spell slots.

[JNAProductions]

A paladin/warlock in general is fine, but not when the warlock side is hexblade. Hexblade is way overpowered as a paladin dip.

Hexblade is overpowered as a dip, period. Or at least, out of whack with what other dips provide with a single level. I don't think it's so powerful as to wreck anything, but it's definitely stronger than most other dips.

[DwarfDM]

Why not?

RAW, it's fine.
RAI, pretty sure that's it.
RAF, I see no issues.

It's a big resource expenditure for big damage.

Probably because it is very hard to explain in game how someone would would be a paladin and a warlock. It depends on the type of game you are playing.

Personaly I don't allow multiclassing unless there is a good explanation for the second class. And I am not talking about any retroactive "I learned this as a child" explanation.

[JNAProductions]

Probably because it is very hard to explain in game how someone would would be a paladin and a warlock. It depends on the type of game you are playing.

Personaly I don't allow multiclassing unless there is a good explanation for the second class. And I am not talking about any retroactive "I learned this as a child" explanation.

Celestial Warlock/Devotion Paladin doesn't make sense because...
Feylock/Ancients Paladin doesn't make sense because...

I'll agree that a Devotion Paladin into Fiendlock generally doesn't make sense (unless you put in some effort-making a deal for the greater good, for instance) but the opposite (Fiend into Devotion) is actually pretty easy-just a redemption story.

[DwarfDM]

I agree with that point of view. I was just offering an explanation for your question.

I can also imagine people (me) not allowing it to prevent powergaming. Like players taking 1 level of war cleric for heavy armor and wizard from level 2 onward.

Although personaly in the case of a redemption story I would replace the warlock levels with the paladin levels. Because I can think of no good reason that a fiend would continue to give its powers to a PC that has renounced him.

[BloodSnake'sCha]

Celestial Warlock/Devotion Paladin doesn't make sense because...
Feylock/Ancients Paladin doesn't make sense because...

I'll agree that a Devotion Paladin into Fiendlock generally doesn't make sense (unless you put in some effort-making a deal for the greater good, for instance) but the opposite (Fiend into Devotion) is actually pretty easy-just a redemption story.
I will like to add that Hexblade and vengeance/conquest also make sense. You make a pact with instruments of war in order to punish/conquer more effectively.

[MrCharlie]

Hexblade is overpowered as a dip, period. Or at least, out of whack with what other dips provide with a single level. I don't think it's so powerful as to wreck anything, but it's definitely stronger than most other dips.
Life, Forge, or War cleric would like a word. Heavy armor and full-spellcaster progression for one level was hexblade before hexblade was hexblade. It's more niche because you need Wis and to want shields and heavy armor, but it's been making wizards into Iron Mage since the edition was released. Wis is never a bad thing, after all.

In general, hexblade isn't as bad as people make it out to be. While it's unquestionably strong, it's delaying whatever your base class is, the spellcasting it gives is highly limited, and the main benefit is it makes your SAD on charisma-but with point buy in 5e and the way armor calculations work that's hard to truly exploit versus your typical-"MAD" on STR/DEX, CON, and one mental (although changelings get a pretty solid game out of it). Hexblades curse is a very strong bonus action but the one target per short-rest means it's really a "screw this guy in particular" button-something which has always been a niche. A good niche, but still a niche.

This tactic just doubles down on what Paladin/Hexblade does best, which is turn one guy into paste. It's really, really good at that, and it's very flashy, but that's not the end-all be-all of DnD combat.

[Millstone85]

Celestial Warlock/Devotion Paladin doesn't make sense because...
Feylock/Ancients Paladin doesn't make sense because...
I will like to add that Hexblade and vengeance/conquest also make sense.All good examples. I will now add a couple evil ones.

I wouldn't put it past a vengeance paladin to make a deal with a fiend. One of their tenets is "by any means necessary" and they are willing to "sacrifice even their own righteousness".

Even better with a conquest paladin and a fiend, specifically a devil. "Some of these paladins go so far as to consort with the powers of the Nine Hells" and become "hell knights".

[Boci]

Although personaly in the case of a redemption story I would replace the warlock levels with the paladin levels. Because I can think of no good reason that a fiend would continue to give its powers to a PC that has renounced him.

Two good reasons I can think of are:

1. Fiends don't monitor indevidual warlocks and often don't notice when one of them has a change of heart
2. The pact is already made, the power if for services already renedered. Fiends can't cancel the pact, but they can refuse to grant more (represented by no longer taking warlocks levels)

Beyond that though, its just more fun is you're warlock/paladin. Id you're just a paladin, then you're the same as someone who was always loyal. Is nice to have a character who has mechanics that reflect an significant aspect of their backstory/development.

[Zhorn]

Definitely allow double smites.

The only problem pops up when running single encounter days (not specific to double smites, it's a general nova problem); where players can pop everything and burn the threat to the ground in one round.

If you know there isn't going to be any more threats to the party that day then going nova is the optimal strategy. As a DM you should make sure players are uncertain of what else could happen before their next short/long rest, nova abilities are used a bit more cautiously then as players don't want to wander into the next encounter on an empty fuel tank.

[sithlordnergal]

Although personaly in the case of a redemption story I would replace the warlock levels with the paladin levels. Because I can think of no good reason that a fiend would continue to give its powers to a PC that has renounced him.

I wouldn't do that...Just make it so the PC can't take anymore Warlock levels. Or have it so the Warlock is stealing power from the Fiend without the Fiend's knowledge. Personally that's always been my favorite Warlock fluff, instead of a proper Pact you're using your Patron's power without them knowing. Otherwise there is no mechanical difference between a Paladin that served a Warlock and redeemed themselves and any other Paladin in the world. Sure, there is technically the RP fluff...but I find RP fluff is pointless and is quickly forgotten/ignored unless there are mechanics to back it up.

[MThurston]

I wouldn't do that...Just make it so the PC can't take anymore Warlock levels. Or have it so the Warlock is stealing power from the Fiend without the Fiend's knowledge. Personally that's always been my favorite Warlock fluff, instead of a proper Pact you're using your Patron's power without them knowing. Otherwise there is no mechanical difference between a Paladin that served a Warlock and redeemed themselves and any other Paladin in the world. Sure, there is technically the RP fluff...but I find RP fluff is pointless and is quickly forgotten/ignored unless there are mechanics to back it up.

They have Celestrial as an option. So why couldn't a holy weapon make a hexblade?

[BloodSnake'sCha]

They have Celestrial as an option. So why couldn't a holy weapon make a hexblade?

There is no reason it can't.

[Zuras]

They have Celestrial as an option. So why couldn't a holy weapon make a hexblade?

The problem is power-wise, the Hexblade acts more like their patron is the Raven Queen or some other entity from the Shadowfell rather than a sentient weapon. Hex Warrior would still fit, but the other features not so much, especially creating undead and otherwise powering your abilities with the life/souls of your victims.

[MoiMagnus]

It's a big resource expenditure for big damage.

A lot of tables have problem with that kind of features of the game (that's one of the criticise of Counterspell too: it allows you to burn your high level spell slots at a rate higher than one per round).

When your table doesn't follow recommended pacing (even without going toward the extreme case of 15min workday), things that allow you to burn resources can create noticeable unbalances. [Additionally, if you play with immature players, it enables the kind of player that burn all its resources and then complain about not having anything remaining]

On the DM side, some DMs really like to make semi-scenarised encounters, and the more there is a difference between "average DPT of the PCs" and "maximal DPT of the PCs", the more it is difficult to actually plan for what will happen. (Illusionist DMs "solve" that by raising the difficulty of the encounter if the PCs chose to burn a lot of resources early on, but that kind of solution is very arguable)

[Zhorn]

Generally I think changing your player's class levels out is bad form. It just comes across as negating that player's agency of choice for their character if the DM is over ruling them.
This goes for either subclasses or full class level swaps. Once a player has earned their class levels in their chosen subclass, leave those locked in. Don't override them with some choice of your own. I hear of enough RPG horror stories of DMs forcing players to class-change entire builds for the sake of their precious campaign story (converting fighters to paladins is what I hear most often).

Restricting choices for future levels is one thing, but what the player is left with is still their own choice on what to invest in.

[Tes]

What's the issue with double Smites for other DMs?

Does it feel like it trivializes encounters? Or is this a fluff concern?
Frontline bruisers usually are weak to mental control attacks but have relatively high HP.
Increasing CR a bit generally raises HP levels and reduces the impact of "big Smites".
You could go so far to include something like a Barlgura or a Hillgiant with Barbarian levels with Reckless Attack just to allow your Paladin Player to feel awesome when he gets a crit to Smite on.

I could see a rather big problem if your Player's end encounters in 2-3 rounds and the Paladin steals the show every time. But that's also something to build your encounters for. Maybe the Hillgiant/Orc Warchief/Bugbear/Gnoll/whatever is now a Bearbarian or a Druid Circle has some beef with the Hexblade and continues sending Agents that slap i.e. Heat Metal on him.
Generally speaking you might want want to balance encounters around more total HP for the monsters. Or just add a horde fight where AoE can shine ever so often.

Or just find something to add that caters to the other players. Looks like the Hexadin won't need much consideration as long as you give him things to Smite.

[ScoutTrooper]

No problems here, fully allowed. I've also played a 'Padlock' in CoS. Albeit I messed up the level plan, but still got the same result. Yeah it Nova damages like no buddy's business, but then that's it. You either burn fast, or you draw it out, and use some other Warlock tricks to maneuver through encounters. I did go the extra distance and play Dark Elf, picked up Elven Accuracy for that 3d20 on attacks, criting on 19's and 20's. Nice. Not over power tho, takes a few rounds to setup. Not to mention concentration saves for some of the other cheese

[MThurston]

No problems here, fully allowed. I've also played a 'Padlock' in CoS. Albeit I messed up the level plan, but still got the same result. Yeah it Nova damages like no buddy's business, but then that's it. You either burn fast, or you draw it out, and use some other Warlock tricks to maneuver through encounters. I did go the extra distance and play Dark Elf, picked up Elven Accuracy for that 3d20 on attacks, criting on 19's and 20's. Nice. Not over power tho, takes a few rounds to setup. Not to mention concentration saves for some of the other cheese

Its not 3d20. Its 2d20 with a reroll. Very different with the odds.

[Boci]

Its not 3d20. Its 2d20 with a reroll. Very different with the odds.

Most people on the forum interpret the reroll to come before the advantage's clause of keeping the highest dice, so it is functionally the same as 3d20. Your interpretation seems just as valid, but also seems to be a needlessly weaker version of the feat.

[Keravath]

Just another data point for the OP.

As DM, I have no issues whatsoever with a character wanting to use both smite and eldritch smite in the same turn. All it does is damage to one target in a single round. It lets the player roll lots of dice and feel good. It is a major threat to at least one creature and as DM you need to consider it in terms of encounter design from time to time. However, the DM also needs to balance the encounter day so that the ability will not be available every fight.

As a player, I think eldritch smite is a trap for a warlock below level 11 since they only have two spell slots and those are usually best used for spells that will increase the warlock's effectiveness for the entire combat (eg darkness+devils sight or shadow of moil) rather than smiting.

[AHF]

Its not 3d20. Its 2d20 with a reroll. Very different with the odds.


Most people on the forum interpret the reroll to come before the advantage's clause of keeping the highest dice, so it is functionally the same as 3d20. Your interpretation seems just as valid, but also seems to be a needlessly weaker version of the feat.

Can someone spell out the interpretation of being able to reroll one of the dice when you have advantage that makes the odds different? I guess I’m not following. Seems like you will always pick the lower die roll to reroll which would lead to the same odds as three dice.

[Boci]

Can someone spell out the interpretation of being able to reroll one of the dice when you have advantage that makes the odds different? I guess I’m not following. Seems like you will always pick the lower die roll to reroll which would lead to the same odds as three dice.

You roll with two dice, pick the highest as per advantage. Then elven accuracy applies and you may reroll the dice you got from the advantage roll, but the reroll will stand. I don't like it as a ruling, but it seems to be a valid reading.

[MThurston]

Most people on the forum interpret the reroll to come before the advantage's clause of keeping the highest dice, so it is functionally the same as 3d20. Your interpretation seems just as valid, but also seems to be a needlessly weaker version of the feat.

It is not anything close to the same thing.

20 crit.

1d20 = 5% chance to crit.
2d20 = 39/400 = 9.75%
3d20 = 13.5%

19/20 crit
1d20 = 10%
2d20 = 76/400 = 19%
3d20 = 27%

Now rolling 3 dice is cheating. Because the rule says you can reroll one of those 2 dice.

[Boci]

It is not anything close to the same thing.

Yes it is. Rolling twice, with the option to reroll one dice, and keep the highest is the same as rolling 3 times and keeping the highest, plus the later saves time. Its only different if you require that elven accuracy happens after you get the highest of the dice rolled from advantage, otherwise it is very much the same. And not cheating, both readings of the feat are valid.

What sequences of numbers rolled are you imagining where there would be a difference between "roll twice, then you can reroll one and keep the highest" and "roll 3 times, keep the highest"?

[OldTrees1]

Now rolling 3 dice is cheating. Because the rule says you can reroll one of those 2 dice.


Whenever you have advantage on an attack roll using Dexterity, Intelligence, Wisdom, or Charisma, you can reroll one of the dice once.

Order of operations: Reading A

Roll 2 dice.
Reroll one of the dice once.
Choose the higher result.
If the player always chooses to reroll the lower die, this is mathematically equivalent to 3d20.

A > B, reroll B,

C > A results in C > A > B, just like 3d20.
C < A results in A > B & C, just like 3d20.

A < B, reroll A,

C > B results in C > B > A, just like 3d20.
C < B results in B > A & C, just like 3d20.



Order of operations: Reading B

Roll 2 dice.
Choose the higher result.
Reroll one of the dice once.
(Why one of the dice if there is only one left?)
This is not equivalent to 3d20. It just allows you to ditch your advantage roll in favor of a straight roll.

[BloodSnake'sCha]

The way I rule adv/dis with lucky is that you roll 2 dice. Take the higher if adv. Or lower is dis.

You can only reroll with lucky the final die of the adv/dis. You choose which dice to use.

Mike have dis, he rolled 14 and 12. He can used lucky. His lucky roll is 13. He can choose 12 or 13 because after the disadvantage roll the 14 don't exist.

Mika have adv, she rolled 15 and 18. She chooses to use lucky. She rolled 17, she can choose 17 or 18 because the 15 don't exist after the advantage roll.

[MThurston]

The way I rule adv/dis with lucky is that you roll 2 dice. Take the higher if adv. Or lower is dis.

You can only reroll with lucky the final die of the adv/dis. You choose which dice to use.

Mike have dis, he rolled 14 and 12. He can used lucky. His lucky roll is 13. He can choose 12 or 13 because after the disadvantage roll the 14 don't exist.

Mika have adv, she rolled 15 and 18. She chooses to use lucky. She rolled 17, she can choose 17 or 18 because the 15 don't exist after the advantage roll.

This is how I do it.

[Alucard89]

Actually double smitting on normal Hex/Paladin build is not that effective resource to cost.

However situation change if we make Elven Accuracy + Curse cirt fishing Vengeance Paladin/Hexblade. Then it makes sense to burn both smites on one crit.

Burning two 3d8 smites on one crit gives us bonus 12d8 damage to one hit, burning two 4d8 smites gives us 16d8 damage. It looks better now.

But it's a niche combo that works well only on crit fish build so I wouldn't worry about it. Normal smitting on hit is better on pure Paladin, Sorcadin or just with Hexblade dip than double smitting on non-crit-fish build.

[MoiMagnus]

The way I rule adv/dis with lucky is that you roll 2 dice. Take the higher if adv. Or lower is dis. You can only reroll with lucky the final die of the adv/dis. You choose which dice to use.

Is it a willing nerf? [Which I can understand, I don't really like feats that are universally good for everyone]. Because I don't see how you read the Lucky feat like that:


Lucky: "You may spend 1 luck point to roll an additional d20. You choose which of the d20s is used for the attack roll, ability check, or saving throw."

So with advantage, you roll two dice, and obtain for ex 8 and 12. The 8 is eliminated* and you have 12. You then use lucky to roll an additional die, and obtain 4. Lucky allow you to keep the old result, so you keep the 12.

*According to RAW, the true resolution of Lucky is "you roll two dices, obtain 8 and 12, then use lucky to roll an additional die and obtain 4, and then chose between 4 & 8 & 12." And it also works for disadvantages, meaning that if you use lucky on a roll with disadvantage, you get to chose between the three dice and can take the highest of the three. This was confirmed in a Sage Advice. But I disagree with this RAW.

Elven Accuracy: "Whenever you have advantage on an attack roll, you can reroll one of the dice once."

This one is more ambiguous. The feat clearly allows you to reroll the die that obtain the smallest result ("one of the dice"). But it doesn't specifies if you are forced to take the result of this reroll, or can take the result of the big die you didn't reroll. I'd probably say the later, but both reading are clearly possible. But I don't see how one can read that you are forced to reroll the highest die and then can swap with the lowest if you want.

[Boci]

This is how I do it.

Using Lucky on advantage is the same as 3d20 keep the highest. (Its technically better, since you roll 2d20, then get to choose whether or not to use it, so its less likely to be wasted).

[MrCharlie]

You roll with two dice, pick the highest as per advantage. Then elven accuracy applies and you may reroll the dice you got from the advantage roll, but the reroll will stand. I don't like it as a ruling, but it seems to be a valid reading.
The problem is that you are explicitly re-rolling "one of the dice"-I.E. one of the advantage rolls. It isn't a matter of order of operations where you "choose the highest dice" then re-roll...because you never actually choose a dice with advantage. You just take the highest one on any roll that has advantage-and nothing about elven accuracy cancels that rule.

In a similar way, a lot of people make you commit to smiting before you roll, either because that used to be the way it was done or because it seems "right" to declare a smite attack first. But it's not the way the game works.

It's an easy fix to the crit fishing build mentioned earlier to implement one or both of these fixes, but does this really need a fix? Again, Paladins are over dramatic action heroes, and paladin hexblades are edgy over dramatic antiheroes, so it absolutely fits the class to wait for a dramatic moment them blow someone the f up. As this isn't the end-all or be-all of combat, I see no problem with it-but then again, as a DM, I shy from encounters with a single enemy for precisely this reason.

[Keravath]

Strict RAW for Lucky is that you roll an extra d20 and then choose which of the available d20s you want to use for the roll. The feat does not mention advantage or disadvantage, it doesn't mention replacing or otherwise changing die rolls. All it says is roll an additional d20 and pick which one you want to use for the result.

Some folks don't like this when applied to disadvantage rolls. Personally, I don't have an issue with it since Lucky isn't a great choice of feat. At least with this case, a Lucky character is actually more likely to come through when the chips are down which I think is fairly thematic.

----

In the case of elven accuracy, it only applies in cases of advantage. It allows you to re-roll one of the die. Since you have advantage, you will always choose to re-roll the lowest so elven accuracy is functionally identical to rolling three d20 and choosing the highest one. There is no difference in probability or outcomes between rolling 3d20 or rolling 2d20 and then re-rolling the lower one. You end up with the results of 3 d20 rolls and choose the highest result.

[Chronos]

Let's say I'm rolling at disadvantage. The two dice show a 17 and a 6. If I don't do anything further, I get a 6, which probably isn't good enough, and this is an important roll that I really want to make.

But wait, I have the Lucky feat! I can take one of those dice and re-roll it. Naturally, I'm going to choose the 6. I re-roll it, and I get a 14. Now I get to choose between those dice, namely, the 6 and the 14. I choose the 14, and it replaces the 6. Now my two disadvantage rolls are a 17 and a 14, and I take the lower of those two, the 14, because I have disadvantage.

That's what the rules say to do. The fact that Sage Advice says something different is just one more example out of many of Sage Advice not knowing what the rules say.

[MrCharlie]

Let's say I'm rolling at disadvantage. The two dice show a 17 and a 6. If I don't do anything further, I get a 6, which probably isn't good enough, and this is an important roll that I really want to make.

But wait, I have the Lucky feat! I can take one of those dice and re-roll it. Naturally, I'm going to choose the 6. I re-roll it, and I get a 14. Now I get to choose between those dice, namely, the 6 and the 14. I choose the 14, and it replaces the 6. Now my two disadvantage rolls are a 17 and a 14, and I take the lower of those two, the 14, because I have disadvantage.

That's what the rules say to do. The fact that Sage Advice says something different is just one more example out of many of Sage Advice not knowing what the rules say.
" You choose which of the d20s is used for the attack roll, ability check, or saving throw."

That line of text is what you're missing and where the interpretation comes from. The reading is that this overrides advantage-rather than resolving the roll normally, you choose which of the rolls is used for the check, despite whatever might normally be said. You never technically replace your roll with the lucky roll-you just directly use the one you want.

[NaughtyTiger]

But wait, I have the Lucky feat! I can take one of those dice and re-roll it.

i disagree. lucky doesn't say you reroll a die, it says you roll an additional die.

"you can spend one luck point to roll an additional d20".
now there are 3 d20s on the table. lucky says

" You choose which of the d20s is used for the attack roll, ability check, or saving throw."

[MThurston]

Strict RAW for Lucky is that you roll an extra d20 and then choose which of the available d20s you want to use for the roll. The feat does not mention advantage or disadvantage, it doesn't mention replacing or otherwise changing die rolls. All it says is roll an additional d20 and pick which one you want to use for the result.

Some folks don't like this when applied to disadvantage rolls. Personally, I don't have an issue with it since Lucky isn't a great choice of feat. At least with this case, a Lucky character is actually more likely to come through when the chips are down which I think is fairly thematic.

----

In the case of elven accuracy, it only applies in cases of advantage. It allows you to re-roll one of the die. Since you have advantage, you will always choose to re-roll the lowest so elven accuracy is functionally identical to rolling three d20 and choosing the highest one. There is no difference in probability or outcomes between rolling 3d20 or rolling 2d20 and then re-rolling the lower one. You end up with the results of 3 d20 rolls and choose the highest result.

There is a difference.

3d20 = 13.5%
2d20= 39/400 = 9.75%
1d20 = 5%.

But a 9.75% with a 5% isn't the same as a 13.5%.

The only argument I will accept is that it takes less time.

[Millstone85]

3d20 = 13.5%
2d20= 39/400 = 9.75%
1d20 = 5%.

But a 9.75% with a 5% isn't the same as a 13.5%.Sorry, but I don't follow your reasoning.

What is "9.75% with a 5%" ? Is it 14.75% ? If so, why make this addition now when you correctly didn't just do 5+5+5% before?

[OldTrees1]

There is a difference.

3d20 = 13.5%
2d20= 39/400 = 9.75%
1d20 = 5%.

But a 9.75% with a 5% isn't the same as a 13.5%.

The only argument I will accept is that it takes less time.

Could you help me understand your argument?
There are 6 possible orderings of 3 dice regardless of whether those dice are rolled as 3d20 or 2d20 & 1d20. Numbers are either higher or lower than other numbers. I am going to call the Advantage Dice "A" & "B". I am going to call the Elven Accuracy Die "C".


Case 1: A > B > C
3d20 picks A.
Advantage & using Elven Accuracy to reroll the lower before Advantage resolves the choices
A > B so reroll B. A > C therefore Advantage chooses A.

Case 2: A > C > B
3d20 picks A.
Advantage & using Elven Accuracy to reroll the lower before Advantage resolves the choices
A > B so reroll B. A > C therefore Advantage chooses A.

Case 3: B > A > C
3d20 picks B.
Advantage & using Elven Accuracy to reroll the lower before Advantage resolves the choices
B > A so reroll A. B > C therefore Advantage chooses B.

Case 4: B > C > A
3d20 picks B.
Advantage & using Elven Accuracy to reroll the lower before Advantage resolves the choices
B > A so reroll A. B > C therefore Advantage chooses B.

Case 5: C > A > B
3d20 picks C.
Advantage & using Elven Accuracy to reroll the lower before Advantage resolves the choices
A > B so reroll B. C > A therefore Advantage chooses C.

Case 6: C > B > A
3d20 picks C.
Advantage & using Elven Accuracy to reroll the lower before Advantage resolves the choices
B > A so reroll A. C > B therefore Advantage chooses C.


That speaks to the general case. Now to your crit specific case.:
19 out of 20 times A is not a crit.
19 out of 20 times B is not a crit.
Therefore 19^2 times out of 20^2 none of A, or B are crits.
(20^2 - 19^2) = 39 times out of 20^2 = 400 at least one of A, or B are crits. This is the 39/400 number.
19 out of 20 times C is not a crit.
Therefore 19^3 times out of 20^3 none of A, B or C are crits.
Therefore (20^3 - 19^3) times out of 20^3 at least one of A, B or C is a crit. This is the 1141/8000 number.
1141 times out of 8000 = 14.2625%

[Chronos]

Where do you get that there are three d20s on the table? In my group, there are usually more like 10. But only two of them are relevant for the Lucky feat.

[Millstone85]

Where do you get that there are three d20s on the table? In my group, there are usually more like 10. But only two of them are relevant for the Lucky feat.I think it would be nice if we kept the discussion on Elven Accuracy for now.

Elven Accuracy has similarities with Lucky, specifically when Lucky is used on a roll with advantage, but even then the wording is a bit different, and there is already disagreement on how Elven Accuracy works.

[MoiMagnus]

There is a difference.

3d20 = 13.5%
2d20= 39/400 = 9.75%
1d20 = 5%.

But a 9.75% with a 5% isn't the same as a 13.5%.

The only argument I will accept is that it takes less time.

You are mathematically wrong. Because "2d20 + one reroll" DOES give the same as 3d20.

Probability of a critic with 3d20: ~14.3% (precisely 1141/8000)
Probability of a critic with 2d20: 9.75% (39/400)
Probability of a critic with 1d20: 5% (1/20)

Probability of a critic with 2d20 + 1 reroll: 9.75% + (90.25% * 5%) = ~14.3% (precisely 1141/8000)
Explanation: Roll 2d20, you made a critic with 9.75% proba. Then, in the 90.25% where you didn't get a critic, you reroll and obtain a critic in 5%.

And in general, you can mathematically prove that "2d20, reroll the lowest, then take the highest" to be the same as "3d20, take the highest". And OldTrees1's post is essentially the proof.

[MThurston]

You are mathematically wrong. Because "2d20 + one reroll" DOES give the same as 3d20.

Probability of a critic with 3d20: ~14.3% (precisely 1141/8000)
Probability of a critic with 2d20: 9.75% (39/400)
Probability of a critic with 1d20: 5% (1/20)

Probability of a critic with 2d20 + 1 reroll: 9.75% + (90.25% * 5%) = ~14.3% (precisely 1141/8000)
Explanation: Roll 2d20, you made a critic with 9.75% proba. Then, in the 90.25% where you didn't get a critic, you reroll and obtain a critic in 5%.

And in general, you can mathematically prove that "2d20, reroll the lowest, then take the highest" to be the same as "3d20, take the highest". And OldTrees1's post is essentially the proof.

I agree to.disagree.

[Boci]

I agree to.disagree.

Its maths, not an opinion. You can't really agree to disagree. Either your maths is wrong, or the maths of everyone who disagreed with you is, but this is a case of right and wrong.

[Keravath]

There is a difference.

3d20 = 13.5%
2d20= 39/400 = 9.75%
1d20 = 5%.

But a 9.75% with a 5% isn't the same as a 13.5%.

The only argument I will accept is that it takes less time.

I think the other posters have explained it but they did not include the math.

Your argument appears to be that there is a different chance of getting a critical hit by rolling 3d20 vs rolling 2d20 and then re-rolling 1. This is completely incorrect.

To obtain a critical hit or even just a hit, you either need to roll a 20 or greater than a target number. We'll just do crits but the math is the same either way.

We want to know what are the chances of at least one crit - we don't care if there is more than one since it doesn't matter and figuring out the odds of rolling 2 or 3 crits isn't needed.

On one die the odds of rolling 1 crit is 1/20 or 5%
What are the odds of rolling at least one crit when rolling 2d20? The chances of any die roll result is 1.0 since you have to roll the die. The easiest way to figure out if there is at least one crit is to find out the odds of rolling no crits and then the number you want is 1.0 - (the chances of rolling no crits)

What is the chance of not rolling a crit on one die? 19/20 ... what are the chances for two die? (19/20)^2 and for three die it is (19/20)^3

Thus the chance of at least one crit on two die is
1.0 - (19/20)^2 = 0.0975 = 9.75%

The chance of at least one crit on three die is
1.0 - (19/20)^3 = 0.142625 ~= 14.3%

So if I roll 2d20 and then roll another d20 separately - what are the odds of a crit?
9.75% for the first two ... but if I already rolled a crit then I don't need to roll another die. The odds of not rolling a crit with the first two is 90.25% ... so what are the odds of rolling a crit with three dice when I roll 2d20 and then reroll 1?

0.0975 + (0.9025 * 0.05) = 0.142625 ~= 14.3%

So the odds of rolling a crit whether you roll 3d20 or roll 2d20 and reroll 1d20 are exactly the same. In fact the odds of succeeding on ANY required to hit number by rolling 3d20 or rolling 2d20 and reroll 1d20 are also exactly the same. As a result, folks just roll 3d20 when rolling with elven advantage since the result is EXACTLY the same as rolling 2d20 and re-rolling one.

[NaughtyTiger]

Where do you get that there are three d20s on the table? In my group, there are usually more like 10. But only two of them are relevant for the Lucky feat.

valid interpretation, but not the only interpretation given the RAW.

[Keravath]

I agree to.disagree.

As the other poster said, there isn't an option to agree or disagree. It is just math. There is a right and wrong answer. One of the nice things about math is that in most cases it is not ambiguous (though there are cases like 1/0 or 0/0 :) ).

2+2 = 4 in base ten :) .. always.

1/2 * 1/2 = 1/4

The probability of a result of a process with fixed individual event probabilities (like the results of a die roll) is another one of the completely predictable mathematical problems though it can get more complicated.

[OldTrees1]

I agree to.disagree.

We can agree to disagree, but before you go, would you please, pretty please, show me your math? If I am wrong, I would appreciate the chance to learn. So please, please, please can you explain it to me?

What do you see on the "3d20" vs the "Elven Accuracy"? Please explain it to me so that I can learn.

[Nagog]

Life, Forge, or War cleric would like a word. Heavy armor and full-spellcaster progression for one level was hexblade before hexblade was hexblade. It's more niche because you need Wis and to want shields and heavy armor, but it's been making wizards into Iron Mage since the edition was released. Wis is never a bad thing, after all.

In general, hexblade isn't as bad as people make it out to be. While it's unquestionably strong, it's delaying whatever your base class is, the spellcasting it gives is highly limited, and the main benefit is it makes your SAD on charisma-but with point buy in 5e and the way armor calculations work that's hard to truly exploit versus your typical-"MAD" on STR/DEX, CON, and one mental (although changelings get a pretty solid game out of it). Hexblades curse is a very strong bonus action but the one target per short-rest means it's really a "screw this guy in particular" button-something which has always been a niche. A good niche, but still a niche.

This tactic just doubles down on what Paladin/Hexblade does best, which is turn one guy into paste. It's really, really good at that, and it's very flashy, but that's not the end-all be-all of DnD combat.

In addition to the reasons you've said it's powerful, i'd like to add on these:

SADness is nice, yes, but SADness on Charisma is gamebreaking. Charisma is the top tier of stats, followed closely by Dexterity. Charisma saves are very common, and Charisma skills are powerful enough to alter the course of the story. Now if we were discussing SADness on Int, like Artificer can get, it'd be an entirely different story. Charisma SADness also means this works with Elven Accuracy for any weapon, a trick that can be further exploited with the already broken as all hell PAM/GWM combo to all but guarantee a hit with +10 damage on your weapon that already deals 1d10 damage on a normal swing. Oh, and Hexblade's Curse means those attacks you roll 3 times for? They have a 10% chance to crit rather than a 5%. And then you can Smite.

Further, a single level investment grants one short rest spell slot. That's not something to laugh at, particularly for half casters that typically only have one or two spell slots per day. That one level dip boosts their spellcasting far more than a similar dip into any other spellcasting class would, as every single other class refreshes their slots on a long rest.

Even more boosting the spellcasting prowess is Cantrips, including the top tier Eldritch Blast. Eldritch Blast scales with overall character level, not spellcasting level, and a one level dip ensures that you'll get a form of Extra Attack at level 5, no matter what. It negates the Attack Progression delay that normally comes with multiclassing, which is one of the greatest downsides of multiclassing a martial class like Pally.

But no, a Hexblade dip isn't broken. Particularly not on Paladins, why would anybody ever think that?

[MThurston]

So if I flip 1 coin I have a 50% chance of tails.

If flip it twice I still have a 50% and a 50% chance.

It's not until you flip that coin 100 times until you see close to a 50/50 split

[JNAProductions]

So if I flip 1 coin I have a 50% chance of tails.

If flip it twice I still have a 50% and a 50% chance.

It's not until you flip that coin 100 times until you see close to a 50/50 split

Any time you flip a coin, you have about a 50/50 chance of seeing Heads or Tails.

If you need ALL Tails, then you halve the chance each time.

1 Flip has a 1/2 chance of getting all Tails.
2 Flips have a 1/4 chance.
3 Flips have a 1/8 chance.

But any given flip has that 50/50 chance.

[OldTrees1]

So if I flip 1 coin I have a 50% chance of tails.

If flip it twice I still have a 50% and a 50% chance.

It's not until you flip that coin 100 times until you see close to a 50/50 split

Yes this is right. You have a 50% and a 50% chance because those are independent events. You can even plot them in a 2D table



Coin1 Heads
Coin1 Tails


Coin2 Heads
2 Heads
1 Heads


Coin2 Tails
1 Heads
0 Heads


So the chance of at least one heads is 3 / 4. The expected value might be 1 heads, but the probability of having at least 1 heads is only 75%. I follow so far, this is exactly what you did to arrive at 39 / 400 for 2d20 criticals.

What is the next step? With 3 rolls (regardless of if it is 3d20 or Elven Accuracy) we will have a 3D table. What is the next step in the math? How should I be counting those 8000 possibilities? Would it be easier to demonstrate with coins or d4s?

Thank you for taking the time to explain.

[Millstone85]

How should I be counting those 8000 possibilities?If I may participate...

I tried imagining the table as a cube made of 8000 smaller cubes, or voxels, 20 on each edge.

Now, we have got three faces of "critical voxels".

Not counting the same voxel twice, there are 400 + 380 + 361 = 1141 critical voxels.

That's 1141/8000 = ~14.3% of the cube.

[MrCharlie]

In addition to the reasons you've said it's powerful, i'd like to add on these:

SADness is nice, yes, but SADness on Charisma is gamebreaking. Charisma is the top tier of stats, followed closely by Dexterity. Charisma saves are very common, and Charisma skills are powerful enough to alter the course of the story. Now if we were discussing SADness on Int, like Artificer can get, it'd be an entirely different story. Charisma SADness also means this works with Elven Accuracy for any weapon, a trick that can be further exploited with the already broken as all hell PAM/GWM combo to all but guarantee a hit with +10 damage on your weapon that already deals 1d10 damage on a normal swing. Oh, and Hexblade's Curse means those attacks you roll 3 times for? They have a 10% chance to crit rather than a 5%. And then you can Smite.

Further, a single level investment grants one short rest spell slot. That's not something to laugh at, particularly for half casters that typically only have one or two spell slots per day. That one level dip boosts their spellcasting far more than a similar dip into any other spellcasting class would, as every single other class refreshes their slots on a long rest.

Even more boosting the spellcasting prowess is Cantrips, including the top tier Eldritch Blast. Eldritch Blast scales with overall character level, not spellcasting level, and a one level dip ensures that you'll get a form of Extra Attack at level 5, no matter what. It negates the Attack Progression delay that normally comes with multiclassing, which is one of the greatest downsides of multiclassing a martial class like Pally.

But no, a Hexblade dip isn't broken. Particularly not on Paladins, why would anybody ever think that?
Charisma saves are extremely uncommon. Charisma skills are excellent but don't convince people to take actions they would otherwise be either incapable or unwilling to do-that said they are probably the second most useful set after wisdom, but only one party member really needs them. Group checks are somewhat plausible but unlikely. In terms of group investment charisma is probably the fifth most likely check for everyone to have to make, higher than intelligence (there only needs to be one smart guy) and worse than everything else, including the rare constitution check.

Intelligence SADness works with elven accuracy too and battlesmiths can use it with any weapon. So does dexterity SADness actually, and some of the best elven accuracy builds are simply archers.

Half casters have between 2-10+ spell slots a day. That 1/short rest slot translates to around 1-4 extra first or second level spells; more than four hours spent short resting is extraordinary, and more than 2 hours is odd. This is good but not a huge payoff, and isn't terribly reliable-having normal spell slots makes this easier to manage though, but still.

Eldritch Blast requires 2 levels in warlock to work as without agonizing it's a mediocre spell that at best synergizes with other spells. Any GWM build requires 3 levels for pact of the blade. 3 levels is more than half a tier of gameplay spent working towards a major payoff-a bit more paltable if you combine the two mentioned, but still. And Eldritch blast has supportive synergy with Paladin (ranged attacks are a weakness), but does not have primary synergy with smite.

The main strength of charisma as a SAD is that five classes use it extensively-swashbuckler rogue, warlock, bard, sorcerer, and Paladin. Alone, charisma is needed on a single party member and the worst or second worst attribute to emphasize otherwise.

So if I flip 1 coin I have a 50% chance of tails.

If flip it twice I still have a 50% and a 50% chance.

It's not until you flip that coin 100 times until you see close to a 50/50 split
What relevance does this have to how elven accuracy math works?

[MThurston]

Charisma saves are extremely uncommon. Charisma skills are excellent but don't convince people to take actions they would otherwise be either incapable or unwilling to do-that said they are probably the second most useful set after wisdom, but only one party member really needs them. Group checks are somewhat plausible but unlikely. In terms of group investment charisma is probably the fifth most likely check for everyone to have to make, higher than intelligence (there only needs to be one smart guy) and worse than everything else, including the rare constitution check.

Intelligence SADness works with elven accuracy too and battlesmiths can use it with any weapon. So does dexterity SADness actually, and some of the best elven accuracy builds are simply archers.

Half casters have between 2-10+ spell slots a day. That 1/short rest slot translates to around 1-4 extra first or second level spells; more than four hours spent short resting is extraordinary, and more than 2 hours is odd. This is good but not a huge payoff, and isn't terribly reliable-having normal spell slots makes this easier to manage though, but still.

Eldritch Blast requires 2 levels in warlock to work as without agonizing it's a mediocre spell that at best synergizes with other spells. Any GWM build requires 3 levels for pact of the blade. 3 levels is more than half a tier of gameplay spent working towards a major payoff-a bit more paltable if you combine the two mentioned, but still. And Eldritch blast has supportive synergy with Paladin (ranged attacks are a weakness), but does not have primary synergy with smite.

The main strength of charisma as a SAD is that five classes use it extensively-swashbuckler rogue, warlock, bard, sorcerer, and Paladin. Alone, charisma is needed on a single party member and the worst or second worst attribute to emphasize otherwise.

What relevance does this have to how elven accuracy math works?

That you can not look at it as 3d20. Its 2d20 with a reroll.

[Boci]

That you can not look at it as 3d20. Its 2d20 with a reroll.

Example A: I flip 3 coins.

Example B: I flip 2 coins, then can choose to reflip 1.

For purposes of getting at least 1 heads, multiple heads being irrelevant as long as I get 1, is there any difference between option A and B?

[MoiMagnus]

Trying to explain the maths in different ways, let's compute the probability of NOT obtaining a critic.
First fact is that rolling 3 dice one by one, or simulatenously, is the same.

Rolling three d20:
Probability of not obtaining a critic on the first die: 19/20 = 95%
Probability of not obtaining a critic on the second die: 19/20 = 95%
Probability of not obtaining a critic on the third die: 19/20 = 95%
=> Probability of never obtaining a critic: (19*19*19/20*20*20) = (6859/8000) = ~85.7%

Rolling two d20 and then rerolling the lowest:
Probability of not obtaining a critic on the first die: 19/20 = 95%
Probability of not obtaining a critic on the second die: 19/20 = 95%
Probability of not obtaining a critic on the rerolled die: 19/20 = 95%
=> Probability of never obtaining a critic: (19*19*19/20*20*20) = (6859/8000) = ~85.7%

EDIT: swapped 5% by 95%. Dummy me.

[Millstone85]

Probability of not obtaining a critic on the first die: 19/20 = 5%
Probability of not obtaining a critic on the second die: 19/20 = 5%
Probability of not obtaining a critic on the third die: 19/20 = 5%I think you made a typo and copy-pasted it. 19/20 is 95%, not 5%.

But yeah, I agree with you and Keravath. The crux of the problem is that, whether you decide to roll three times or to roll two times and then a third if you haven't already obtained a crit, you are facing a probability of failure of (19/20)^3.

And so your probability of success is 1 - (19/20)^3, in both scenarios.

It also works if you need a result of 13 or higher. Then it is 1 - (12/20)^3, again in both scenarios.

[MrCharlie]

That you can not look at it as 3d20. Its 2d20 with a reroll.
What you said has no relevance to that. I'm not sure what you meant by what you said-it appears to be something about how the law of large numbers-but it has no relevance to the discussion at hand.

Why do you think 3d20 is statistically different from 2d20 with a reroll?

[Satori01]

What are the odds of rolling 3 Twenties on 3D20 Super Advantage roll?

Now what are the odds of rolling 3 Twenties on a a roll of 2D20, player gets to reroll one die?

The answer is None, because as soon as one gets a critical hit, one stops rerolling.

So if 2d20 plus one reroll has a chance of 0% to roll 3 Twenties
and 3D20 has a chance Greater than Zero to roll 3 twenties,

...then, definitionally, the methods are not “equivalent”.

[Boci]

What are the odds of rolling 3 Twenties on 3D20 Super Advantage roll?

Now what are the odds of rolling 3 Twenties on a a roll of 2D20, player gets to reroll one die?

The answer is None, because as soon as one gets a critical hit, one stops rerolling.

So if 2d20 plus one reroll has a chance of 0% to roll 3 Twenties
and 3D20 has a chance Greater than Zero to roll 3 twenties,

...then, definitionally, the methods are not “equivalent”.

But that outcome is irrelevant. One 20 and three 20s are identical for the purposes of the game, so they are equivilant for the purpose of the game. The claim was that 3d20 over 2d20 with reroll changed the odds and that 3d20 was cheating, as demonstrated here:


Its not 3d20. Its 2d20 with a reroll. Very different with the odds.

and here:


Now rolling 3 dice is cheating.

[Millstone85]

Let's go back to that simple coin toss.

You can toss a coin up to 2 times. You win if you obtain heads at least once.

Method A, you toss a coin twice.

1st toss: heads

2nd toss: heads ---> You obtained heads twice (25%)
2nd toss: tails ---> You obtained heads then tails (25%)

1st toss: tails

2nd toss: heads ---> You obtained tails then heads (25%)
2nd toss: tails ---> You obtained tails twice (25%)


As you can see, there are 4 possible final outcomes, each with a probability of 25%.

And 3 of those include heads, so the probability of winning is 25 + 25 + 25 = 75%.

Method B, you toss a coin, then toss it again if you lost.

1st toss: heads ---> You obtained heads and stopped (50%)
1st toss: tails

2nd toss: heads ---> You obtained tails then heads (25%)
2nd toss: tails ---> You obtained tails twice (25%)


There are now only 3 possible final outcomes. But does each one have a probability of 33.33%? Of course not. That first heads still has a probability of 50%, and the other two outcomes have to divide the remaining 50% between themselves.

Probability of winning? 50 + 25 = 75%.

It changes nothing.

[MoiMagnus]

...then, definitionally, the methods are not “equivalent”.

Indeed, the good comparison is "roll 3d20, keep the highest" and "roll 2d20, reroll the lowest, then keep the highest".
If you keep all the die (which you never do), then there is a difference. If you keep only the highest die, there is none.

Those are mathematically equivalent on EVERY point, since at the end you have only one die remaining, and the same probability for each of the 20 faces of the die. (In particular, roughly 14% of obtaining the face 20).

[Satori01]

Language and precision matter. Functional Equivalence is different then equivalence.

Agreed, not much different, but precision matters.

The question of 3D20 being functionally or otherwise equivalent to the actual Text in the Rules, truly depends on the interpretation of this clause in Elven Accuracy:
• Whenever you have advantage on an attack roll using Dexterity, Intelligence, Wisdom, or Charisma, you can reroll one of the dice once.

Does the new role automatically replace the old number, or are all three results, selectable?

The amount of time saved by rolling 3d20 at once’s is negligible, yet already we see this method leads to a more generous outcome then the Text of the Feat as written.

[Corran]

Does the new role automatically replace the old number, or are all three results, selectable?

The amount of time saved by rolling 3d20 at once’s is negligible, yet already we see this method leads to a more generous outcome then the Text of the Feat as written.

What makes you think that the extra roll (that presumably replaces the lowest of the two previous rolls, though that the player's call) has to be picked though? There is nothing suggesting that, so in the absence of any further content, we are still operating in the same way as when rolling with advantage. Ie we pick the roll of our choosing highest roll.

Edit:

Roll with advantage -> 5, 11
Result = pick the highest roll = 11

Roll with advantage -> 5, 11
Elven accuracy -> reroll 5 -> 8
Result = pick the highest roll = 11

Roll with advantage > 5, 11
Elven accuracy -> reroll 11 -> 8
Result = pick the highest roll = 8

In other words, elven accuracy does not change the fact that you have advantage. You are still picking the die that rolled highest. Only now you have two chances with one of these two dice. Which essentially translates exactly the same as rolling 3 dice. At least that's how I understand it. For the sake of completeness (also realization), rolling 3 dice is not the same only in the case where you choose to reroll your highest roll between the two initial rolls. But why would anyone do that if given the choice.

[MrCharlie]

Language and precision matter. Functional Equivalence is different then equivalence.

Agreed, not much different, but precision matters.

The question of 3D20 being functionally or otherwise equivalent to the actual Text in the Rules, truly depends on the interpretation of this clause in Elven Accuracy:
• Whenever you have advantage on an attack roll using Dexterity, Intelligence, Wisdom, or Charisma, you can reroll one of the dice once.

Does the new role automatically replace the old number, or are all three results, selectable?

The amount of time saved by rolling 3d20 at once’s is negligible, yet already we see this method leads to a more generous outcome then the Text of the Feat as written.
Advantage never lets you select a roll. You simply pick the highest roll on any check with advantage. Therefore rerolling one of the dice can never be worse; there is no game mechanism to take the re-roll if it's lower.

As such, it is equivalent.

And functional equivalence is exactly the same as equivalence in all relevant ways. That's why we call it functional equivalence. In fact, this isn't functional equivalence; it's faster and therefore better.

[Corran]

Advantage never lets you select a roll. You simply pick the highest roll on any check with advantage. Therefore rerolling one of the dice can never be worse; there is no game mechanism to take the re-roll if it's lower.

As such, it is equivalent.

And functional equivalence is exactly the same as equivalence in all relevant ways. That's why we call it functional equivalence. In fact, this isn't functional equivalence; it's faster and therefore better.
To be fair, there is one case when rolling 3 d20's would not be appropriate. That's the case when for some reason you are using EA to reroll the highest of your two initial rolls (two initial rolls because of advantage), and that reroll ends up being lower than the lowest of your two initial rolls. For example, with advantage you roll a 10 and an 11, you use elven accuracy to reroll the 11 and get a 9, and now advantage has you take the highest roll between 10 and 9 (cause you used elven accuracy to replace your 11 with a 9). Rolling 3 d20's cannot accommodate this scenario. Now, there is no good reason using elven accuracy the way we just did (ie choose to reroll the highest number instead of the lowest), but theoretically that's the one case where rolling 3d20's is not equivalent to how advantage + EA functions. Of course, if you intend on using EA to your benefit, rolling 3 d20's is correct and the easiest way to go.

Edit:
@MThurston: Apologies if you already mentioned this, but could it be that the whole disagreement boils down to how the die that EA rerolls is chosen?

[Satori01]

Faster = Better, yet the time savings is negligible, and at least in one agreed upon situation, results in a very minor boost in the odds for the PC.

So I find it convenient that so many people value expediency, when it expediency resounds to their own benefit,even if that benefit is minor, or pyrrhic.

The Samurai Fighter with Elven Accuracy is typically NOT taking to much time rolling to hit, even using the RAW reroll verbiage in the feat.

Rolling 3d20, and declaring it a major time saver is a disingenuous position.

The calculation of damage from multiple critical hits does take time.

So, to ask a question, how many of you that roll 3d20 to “save time”...those precious seconds lost from re-rolling a d20....

.....also use pre calculated average damage for their PC?

[OldTrees1]

Language and precision matter. Functional Equivalence is different then equivalence.

Agreed, not much different, but precision matters.

The question of 3D20 being functionally or otherwise equivalent to the actual Text in the Rules, truly depends on the interpretation of this clause in Elven Accuracy:
• Whenever you have advantage on an attack roll using Dexterity, Intelligence, Wisdom, or Charisma, you can reroll one of the dice once.

Does the new role automatically replace the old number, or are all three results, selectable?

The amount of time saved by rolling 3d20 at once’s is negligible, yet already we see this method leads to a more generous outcome then the Text of the Feat as written.

Could you please demonstrate this claim?
If I always reroll the lower die with Elven Accuracy, how is it less generous than 3d20 Advantage?
In both cases there are 3 d20s rolled. In both cases only the highest result is used, because advantage mandates the highest is used. Where is the difference you see? Can you find one example in the 8000 possibilities that demonstrates this behaviour?

[Satori01]

Have you not read the thread?

How many times will you roll 3 Twenties when rolling 2d20 with one re-roll?

On the surface the differences between RAW and 3d20 Super Advantage are negligible and subtle...this is agreed to by all.

Yet in symbolic logic terms, clauses like “re-roll one die”, often have unclear, but not insignificant repurcusions.

Again the time savings on 3d20 roles is negligible.

[JNAProductions]

Have you not read the thread?

How many times will you roll 3 Twenties when rolling 2d20 with one re-roll?

On the surface the differences between RAW and 3d20 Super Advantage are negligible and subtle...this is agreed to by all.

Yet in symbolic logic terms, clauses like “re-roll one die”, often have unclear, but not insignificant repurcusions.

Again the time savings on 3d20 roles is negligible.

Why is that relevant? Three 20s has the same effect as one 20 and two 1s.

[MrCharlie]

Faster = Better, yet the time savings is negligible, and at least in one agreed upon situation, results in a very minor boost in the odds for the PC.

So I find it convenient that so many people value expediency, when it expediency resounds to their own benefit,even if that benefit is minor, or pyrrhic.

The Samurai Fighter with Elven Accuracy is typically NOT taking to much time rolling to hit, even using the RAW reroll verbiage in the feat.

Rolling 3d20, and declaring it a major time saver is a disingenuous position.

The calculation of damage from multiple critical hits does take time.

So, to ask a question, how many of you that roll 3d20 to “save time”...those precious seconds lost from re-rolling a d20....

.....also use pre calculated average damage for their PC?
No one. I doubt anyone plays "DnD" without attributes either, but I suppose you could ask hypotheticals regarding that as well. I have no idea what the relevance is of either question.

I have no idea what your position even is now; is it that wasting time is good? Or is it that there is some benefit we're trying to hide behind "it's expedient"?

The second is simply false; there is no benefit besides that it speeds up play past the actual rolling of dice. It's exactly the same as re-rolling a dice statistically. If done correctly rolling 3d20 and keeping the highest is not a benefit to the player, and the chance to hit isn't higher nor is the chance of a critical higher in comparison to rolling 2d20 and re-rolling a dice.

I can't state this clearly enough; Rolling 3d20 and keeping the highest has the same mathematical result as rolling 2d20 and re-rolling the lowest. You've provided one argument which might refute that, and it relies on what I view as a misreading of the rules. If you want to argue the interpretation that the re-rolled dice has to be the one you use, even if it's worse than the higher of your first two, then fine, that's a valid discussion. But it's not what we appear to be talking about. The only time it's ever different, at least with the interpretation I've stated, is if the dice are rigged or poorly unbalanced.

Finally, it's supremely easy to have a dice-roller like roll20 automatically process 3d20KH1, and rolling 2d20 then "re-rolling" a dice by rolling another dice and "replacing" the rolled dice is the same exact thing, except it isn't easily automated and thus takes longer. The program obviously can't actually re-roll a dice and is simply reporting the next number from the RNG, so it's supremely silly to insist there is a difference. Given that a lot of DnD is done online with dice-rollers, particularly now, this is a non-trivial time saving.

[OldTrees1]

Have you not read the thread?Yes, extensively. But what I haven't seen is your counter example. But since I have read the thread extensively, I am asking for your counter example rather that post another 3 proofs of the equality. Yours is the voice being drowned out. Yours is the voice I want to get a chance to speak. Yours is the voice that is unfortunately not giving a proof or counter example.


How many times will you roll 3 Twenties when rolling 2d20 with one re-roll? 1 in 20^3 times
When rolling 2d20 with one reroll if I get 2 20s I can reroll one of them and see a 3rd 20 5% of the time. It won't effect the outcome but I can still do it. Just like the 3rd die of the 3d20 does not matter if the first one is a 20. Since that event only happens once every 400 times, that 3rd 20 is once every 8000 times.


On the surface the differences between RAW and 3d20 Super Advantage are negligible and subtle...this is agreed to by all.If it is agreed by all, or even if only you believe it, then demonstrate it. Show me which of the 8000 cases is different.


Yet in symbolic logic terms, clauses like “re-roll one die”, often have unclear, but not insignificant repurcusions.
Can you please demonstrate how 3d20 is not equivalent to 2d20 reroll the lowest?


Again the time savings on 3d20 roles is negligible.Nobody has been arguing about that. We just want to hear which of the 8000 cases is mathematically different. Demonstrate the probability is different to back up your claim so that we can learn. Otherwise we will continue to use the math that proves 3d20 advantage is equivalent to (2d20 reroll the lower of the 2 dice) advantage.


Satori01, let's talk about Bob. Bob only owns one d20. Bob likes to roll dice and will roll if allowed even if not required. They roll that d20. Then they roll it again to represent the 2nd die of the 2d20 with advantage, even if they already got a 20. Then they roll it again to represent rerolling the lower of the first two results, again even if they already got a 20. They have rolled the d20 3 times and their result is the highest of the 3 rolls. How is this not 3d20 advantage? Which trio of results gives a different outcome one way vs the other?
Is it 1,1,1? 2d20reroll gives 1. 3d20 gives 1.
Is it 7,11,13? 2d20reroll gives 13. 3d20 gives 13.
Is it 11,13,7? 2d20reroll gives 13. 3d20 gives 13.
Is it 13,7,11? 2d20reroll gives 13. 3d20 gives 13.
Is it 1,10,20? 2d20reroll gives 20. 3d20 gives 20.
Is it 1,20,0? 2d20reroll gives 20. 3d20 gives 20.
Is it 20,10,1? 2d20reroll gives 20. 3d20 gives 20.
Is it 20,20,20,? 2d20reroll gives 20. 3d20 gives 20.

[Satori01]

oldtrees1 is the amount of time that poor old bob needs to roll his one twenty sided die 3 times so onerous?

Mine is ultimately less a mathematical argument, and more an argument regarding self interest.

If either method of rolling for Super Advantage yield identical results, and the time savings from either method is negligible then using the method as described in the text seems the easiest, and less controversial way of resolving Super Advantage rolls.

I posit again, it is not the rolls to hit with Super Advantage present, that take time...it is the rolls for damage.

How many of the posters advocating for 3d20 method, due to “Time Savings” use PC Average Damage?

My guess is: probably not many.

Again: if either rolling method yields

1) Identical Results
2) Near Identical Time Expenditures

Then Why Alter the text of the feat to roll 3d20 simultaneously?

[Corran]

If either method of rolling for Super Advantage yield identical results, and the time savings from either method is negligible then using the method as described in the text seems the easiest, and less controversial way of resolving Super Advantage rolls.

Yet the only reason for controversy is because of this idea suggested that there is in fact a difference between the 2 ways mentioned. While that's not really the case. Satori01, no one is meaning to argue how you, or me, or anyone else should perform their dice rolling ritual. No one cares about that. If you want to be typical for the sake of being typical, go right ahead. If you find overcomplicating things to be simpler, again, no argument there. But let us be clear that there is nothing controversial over the equality of the outcome of both methods, results wise.

[OldTrees1]

oldtrees1 is the amount of time that poor old bob needs to roll his one twenty sided die 3 times so onerous?

Um, this hypothetical Bob only has 1 die. It takes the same time to roll it 3 times regardless. Time, despite being your pet, was not relevant, so I ensured it was irrelevant.


Mine is ultimately less a mathematical argument, and more an argument regarding self interest.

If either method of rolling for Super Advantage yield identical results, and the time savings from either method is negligible then using the method as described in the text seems the easiest, and less controversial way of resolving Super Advantage rolls.

I posit again, it is not the rolls to hit with Super Advantage present, that take time...it is the rolls for damage.

How many of the posters advocating for 3d20 method, due to “Time Savings” use PC Average Damage?

My guess is: probably not many.

Again: if either rolling method yields

1) Identical Results
2) Near Identical Time Expenditures

Then Why Alter the text of the feat to roll 3d20 simultaneously?

I see. So you agree completely. You agree that the rolling methods are mathematically equivalent. 3d20 advantage is equivalent to (2d20 reroll lowest) advantage. Which means no text was being altered. 3d20 is 2d20 with the lowest rerolled. The reroll is just being done at the same time as the roll. While it is technically a controversy when a single person disagrees, regardless of veracity. Colloquially it takes a lot more than just 1 person for mundane probability to be controversial.

Your timing considerations are not relevant to me. If you didn't notice, most of this thread was about the mathematical equivalence, the 1 person that disagreed with that conclusion, and your posts being misunderstood as more of the same. Timing was not a concern. People just understand fundamental probability and could demonstrably prove the equivalence between rolling the Elven Accuracy d20 with or after the 2d20

PS: Since even I know about rolling damage simultaneously with the attack roll, I presume you know about it too.

[MrCharlie]

oldtrees1 is the amount of time that poor old bob needs to roll his one twenty sided die 3 times so onerous?

Mine is ultimately less a mathematical argument, and more an argument regarding self interest.

If either method of rolling for Super Advantage yield identical results, and the time savings from either method is negligible then using the method as described in the text seems the easiest, and less controversial way of resolving Super Advantage rolls.

I posit again, it is not the rolls to hit with Super Advantage present, that take time...it is the rolls for damage.

How many of the posters advocating for 3d20 method, due to “Time Savings” use PC Average Damage?

My guess is: probably not many.

Again: if either rolling method yields

1) Identical Results
2) Near Identical Time Expenditures

Then Why Alter the text of the feat to roll 3d20 simultaneously?
I don't want to repeat myself, but I feel the need for repetition here-It does have significant time-saving impact with computers and thus online. Which is, ultimately, why we're talking about it as 3d20 in the first place; the math is easier to model this way.

As for controversial-let me pose a hypothetical to you. Do you think it is controversial for someone to roll the same dice twice for advantage? There are all sorts of reasons for why they do it-they may own only one d20 and not want to reach over to borrow another, or have dice superstitions-but the point is that I've never heard of this as controversial. Yet is is exactly as illegal, according to the rules, as rolling 3d20 for elven accuracy.

If it's statistically the same, it shouldn't be controversial. And until you've mentioned it, I've only heard controversy from people who think that there is a statistically difference, and saying "I always re-roll the lower one" makes that go away.

[MoiMagnus]

Do you think it is controversial for someone to roll the same dice twice for advantage? There are all sorts of reasons for why they do it-they may own only one d20 and not want to reach over to borrow another, or have dice superstitions-but the point is that I've never heard of this as controversial. Yet is is exactly as illegal, according to the rules, as rolling 3d20 for elven accuracy.

Additional arguments against rolling the same dice twice for advantage: more errors + trust issues. The player can forget/misremember the first roll, and if you play with strangers (or known cheaters, but why would you do that?) this could cause some trust issues if the reroll is done immediately without other peoples having the time to see it.

Additional argument against rolling 3d20 for Elven accuracy: you cannot be suboptimal. I mean, that would be some very weird situation where you're making an attack roll with advantage that you want to fail (or not critic), but in this situation Elven accuracy allows you to reroll the highest, which 3d20 does not.

[OldTrees1]

Additional argument against rolling 3d20 for Elven accuracy: you cannot be suboptimal. I mean, that would be some very weird situation where you're making an attack roll with advantage that you want to fail (or not critic), but in this situation Elven accuracy allows you to reroll the highest, which 3d20 does not.

In those situations, I would suggest rolling (the 2d20 and the damage dice) in a first roll and (the Elven d20 and the critical dice) in a second roll. That gives you the time to pick up the higher of the 2d20.

So I don't find that as an argument against rolling the 3d20, it is just a situational case where you would roll 2d20 & 1d20 instead.