GWF Paladin Reroll 1s & 2s on Divine Smite?

[Snig]

Great Weapon Fighting
When you roll a 1 or 2 on a damage die for an attack you
make with a melee weapon that you are wielding with
two hands, you can reroll the die and must use the new
roll. The weapon must have the two-handed or versatile
property for you to gain this benefit.

Would this apply to Divine Smite?

[MaxWilson]

As I read the rule, no. No way.

[TopCheese]

Great Weapon Fighting
When you roll a 1 or 2 on a damage die for an attack you
make with a melee weapon that you are wielding with
two hands, you can reroll the die and must use the new
roll. The weapon must have the two-handed or versatile
property for you to gain this benefit.

Would this apply to Divine Smite?

Only the weapon damage would be rerolled, not the extra damage.

Same thing with the cleric's ability to maximize lightning damage. Weapon damage isn't maximized, only the lightning part of you attacked with a sword that did lightning damage. Like 2d6+3+1d6 lightning. Only the d6 if lightning would be maximized. Only the 2d6 would get 1 and rerolls.

[Aetol]

Sage Advice is inconsistent : No (http://www.sageadvice.eu/2014/10/06/great-weapon-smite/) Yes (http://www.sageadvice.eu/2015/09/09/great-weapon-fighting-and-smite/)

Both from Mike Mearls, even.

[bid]

Only the weapon damage would be rerolled, not the extra damage.
"damage die for an attack" has nothing about weapon-only. Moreover, GladiusLegis guide asssumes this and it hasn't been challenged.

[TopCheese]

Sage Advice is inconsistent : No (http://www.sageadvice.eu/2014/10/06/great-weapon-smite/) Yes (http://www.sageadvice.eu/2015/09/09/great-weapon-fighting-and-smite/)

Both from Mike Mearls, even.
There was a recent one about the cleric's maximize lightning, same set up really.abilities only apply to what the ability says it applies to.

[Ruslan]

"damage die for an attack" has nothing about weapon-only.Yup. Assuming you're wielding a weapon with two hands, Smite damage is indeed a "damage roll for an attack with a melee weapon that you are wielding with two hands".

[Strill]

Yes it does apply to smite. If it didn't, 2-handed weapons wouldn't be worth using over shields for Paladins.

[Longcat]

Yes it does apply to smite. If it didn't, 2-handed weapons wouldn't be worth using over shields for Paladins.

That's not true at all, as 2-handed weapons have access to both a -5/+10 feat as well as Bonus Action attacks, allowing you to smite more often and deal more consistent damage.

[Strill]

That's not true at all, as 2-handed weapons have access to both a -5/+10 feat as well as Bonus Action attacks, allowing you to smite more often and deal more consistent damage.

You mean less consistent damage. -5/+10 makes you hit less often, dealing less consistent damage, and allowing you to smite less often.

Also, you're aware that -5/+10 is worst on a Paladin right? They have the most to lose from missing, due to all the damage bonuses they get already. Improved Divine Smite, together with Hunter's Mark or Divine Favor or Crusader's Mantle.

If you want more consistent damage, you take Shield Master to knock enemies prone for advantage.

[Malifice]

As I read it: Yes.

[bid]

Also, you're aware that -5/+10 is worst on a Paladin right? They have the most to lose from missing, due to all the damage bonuses they get already. Improved Divine Smite, together with Hunter's Mark or Divine Favor or Crusader's Mantle.
Absolutely. Doing 10 damage with 10/20 to hit is the same as doing 20 damage with 5/20 to hit. And at that point you paid for a feat that offers nothing.

It's great for barbarian who can frenzy for advantage and come out 2 damage ahead on average, but you need dependable advantage to gain something.

Since a pally averages {2d6 + 3 + 2d8 = 19} base damage, the balance point becomes 19 on 15/20 vs 29 on 10/20 (aka you need 19/20 hit to do as well as barbarian).

[djreynolds]

Yes. And Rogue's should look at the savage attacker, it may be worth it as well for them.

[Strill]

Yes. And Rogue's should look at the savage attacker, it may be worth it as well for them.

No it's not. Rerolling a d6 is terrible. Rogues are much better off with just +2 DEX.

[Ruslan]

Regarding Savage Attacker, it actually doesn't work, because it only lets you reroll "the weapon’s damage dice". Compare and contrast with GWF's "damage roll for an attack"...

[GandalfTheWhite]

I let them do it. Otherwise, every Paladin would take a shield and the Defensive fighting style.

[PoeticDwarf]

Sage Advice is inconsistent : No (http://www.sageadvice.eu/2014/10/06/great-weapon-smite/) Yes (http://www.sageadvice.eu/2015/09/09/great-weapon-fighting-and-smite/)

Both from Mike Mearls, even.

Wow that is very strange xD

I would say only the weapon damage dice, GWF isn't the best fighting style maybe but it is still a good option for many builds.

[PoeticDwarf]

I let them do it. Otherwise, every Paladin would take a shield and the Defensive fighting style.

Shield and defensive is so bad. Rapier + shield + duelist would be better than greataxe + GWF style but something like polearm master is better anyway.

[Tenmujiin]

RAW definately
RAI probably not
Balance wise? It makes GWF by far the best fighting style for a paladin but without it GWF is kinda weak on a paladin.

[Strill]

Balance wise? It makes GWF by far the best fighting style for a paladin but without it GWF is kinda weak on a paladin.

No it doesn't. Shield Master is great for Paladins.

[MaxWilson]

Yup. Assuming you're wielding a weapon with two hands, Smite damage is indeed a "damage roll for an attack with a melee weapon that you are wielding with two hands".

No, it isn't. It's damage for a smite. Smiting is clearly not an attack because it doesn't fit the criteria on PHB 194: there is no attack roll, so it's not an attack. It's just a free action that happens after a successful attack.

If you had to declare smites before you made your attack roll I might view it differently, as part of the attack, but RAW it clearly is not.

[MeeposFire]

No, it isn't. It's damage for a smite. Smiting is clearly not an attack because it doesn't fit the criteria on PHB 194: there is no attack roll, so it's not an attack. It's just a free action that happens after a successful attack.

If you had to declare smites before you made your attack roll I might view it differently, as part of the attack, but RAW it clearly is not.

Are there free actions anymore? I did not think so...

Also consider what you are saying if a rogue has two attacks per round and hits with his first attack on one target but ants to save his sneak attack for his second it is clear he gets to choose whether or not to use his SA dice after he hits. Your reasoning would make SA useless since he chooses after he hits which we know as false.

Sneak attack is chosen after you hit as well should it not apply to abilities that apply to attacks? Whether or not you think GWF applies to smiting a smite is certainly part of an attack hence why it doubles on a crit. Crits only apply to abilities that are applied via an attack and the rules for attacks clearly say that additional abilities (damage or otherwise) are from a result of something that usd an attack roll.

[Talin]

Great Weapon Fighting
When you roll a 1 or 2 on a damage die for an attack you
make with a melee weapon that you are wielding with
two hands, you can reroll the die and must use the new
roll. The weapon must have the two-handed or versatile
property for you to gain this benefit.

Would this apply to Divine Smite?

I as a DM have personally ruled that you get to reroll the 1s and 2s on Smite if its a Two-Handed attack, not because I think it should be allowed but because the players were trying to convince me the low rolls on their dice were always their weapons.........I wasn't in the mood to make them have different colored dice for melee and magic damages so I just say "Whatever fine reroll the smites too I guess". I guess in the end its just a matter of compromise at the table.

[Malifice]

I allow it.

Even on a 5d8 smite, its adds what... an extra 5 points of damage or so on average per attack?

If youre worried this will break your game, ban all spell casters.

[Strill]

I allow it.

Even on a 5d8 smite, its adds what... an extra 5 points of damage or so on average per attack?

If youre worried this will break your game, ban all spell casters.
5d8 = 22.5
5d8 reroll 1s and 2s = 26.25

+3.75 damage

[Lollerabe]

Sniq - Yes it does, by raw it does so without question.

I'd like to add why you should always let GWF do so.

I tried to search the forums but couldnt find the thread with the math is was looking for.

Anyway GWF with a maul/greatsword gives you 15% - 20% more damg compared to duelist when you factor nothing in but ability mod.

However you give up a shield (a shield that can add up to +5 armor lategame) for that damg advantage,
but if GWF dosent allow you reroll smites,inspiration,hunter's mark, elemental weapons etc that 15%-20% damg increase dosent scale at all.

See the idea behind GWs is that they should deal about 20% more damg than sword and board, throughout the game by default - regardless of feats or other modifiers (magic weapons fx).

Imagine you are playing an oov paladin with a str score of 16, a flametounge wep and hunters mark on your target, and you smite with a lvl 2 slot.

A SB with duelist would deal - 4d8 + 3d6 + 5 damg
same pala with maul and GWF - 5d6 + 3d8 + 3 damg

If you then only allowed the GWF to reroll 2d6 out of that entire dice pool, his overall DPR advantage would be barely noticeable compared to the duelist. If you did allow the rerolls tho (which is the rules btw) you would most likely end up with a 15-20% damg advantage in favor of the GWF pala - as the rules intended.

The argument that the GWM feat makes GW viable is completely besides the point.
A. Players shouldn't be forced to take a feat to make GW the best damg option throughout the game.
B. GWM dosen't require GWF as a prereq. You can go GWM with the defense fight style no probes, which you absolutely should, if your DM makes such a horrible ruling as allowing GWF not to scale AT ALL.
C. There are feats for SB Palas as well, shieldmaster being strong as hell, it's not like GW wielders get GWM for free.

The entire idea that allowing GWF to reroll any dice is broken, is false and not based on any math.
The players and DMs that enforce this (stupid) rule might have had an experience where a pala oneshotted a BBEG and therefore deem it OP. If you look at how often rerolling smites actually comes into play, it's not a problem at all.

TL;DR: Yes it does allow you to reroll smites - and it should. People who say otherwise haven't looked into the math, and are gimping GWF style to a point where it dosent make sense for any pala to even glance at.

[Malifice]

TL;DR: Yes it does allow you to reroll smites - and it should. People who say otherwise haven't looked into the math, and are gimping GWF style to a point where it dosent make sense for any pala to even glance at.

Yep. An extra 2-3 damage per attack is nothing at high levels (compared to +2 AC for a shield and +1 for defence style).

From 11th level (when Paladins are getting 4d8 smites) Fighters are getting 3-7 attacks per round (and BM dice are re-rollable, and champions double those weapon dice on an 19-20) so an extra 2-3 damage for a Fighter per attack is really an extra 6-21 per round from GWM (the same boost that a Paladin can expect from re-rolling 1's and 2's on smite dice).

Even at +4d8 smites, the Paladin is getting what.. an extra 4-5 damage over what he would do normally? In exchance for a +2-6 to his AC?

Rerolling all dice sounds totally balanced; heck its even arguably suboptimal.

[GoodbyeSoberDay]

No, it isn't. It's damage for a smite. Smiting is clearly not an attack because it doesn't fit the criteria on PHB 194: there is no attack roll, so it's not an attack. It's just a free action that happens after a successful attack.

If you had to declare smites before you made your attack roll I might view it differently, as part of the attack, but RAW it clearly is not.Interesting; I hadn't seen this interpretation before.

Let's look at Smite's text:
Starting at 2nd level, when you hit a creature with a melee weapon attack, you can expend one paladin spell slot to deal radiant damage to the target, in addition to the weapon’s damage. The extra damage deals...Emphasis mine. I read the first sentence and started to believe it didn't apply. Then I read "in addition to the weapon's damage" and "extra damage," and it became clear that the radiant damage adds to the attack's damage, which makes it part of the damage of the weapon attack and not a separate effect.

[steppedonad4]

Wow that is very strange xD

He's always been inconsistent on these things which is why he usually states what he's saying is for how he would rule, not that his opinion is official. Also why there's a Sage and he's not it :smallbiggrin:

[Kryx]

Even at +4d8 smites, the Paladin is getting what.. an extra 4-5 damage over what he would do normally? In exchance for a +2-6 to his AC?
If you're playing with the optional rule for +X magic items, which many GMs do not.

GWF is not nearly as suboptimal as people proclaim. GWM does about 30% more damage than a sword and board (who has dueling and is proning an enemy). Same with Polearm.

It'll be slightly different for a Paladin, but not much.

If you think GWM is a bad fighting style then boost it, but having it apply to spells and everything else is far outside the scope of any fighting styles imo.

[Lollerabe]

Guess we can call feats optional as well.

Kryx care to elaborate on how "having it apply to spells and everything else" makes it "far outside the scope of any fighting styles" ?

Use my example:
You are playing an oov paladin with a str score of 16, a flametounge wep and hunters mark on your target, and you smite with a lvl 2 slot

A SB with duelist would deal - 4d8 + 3d6 + 5 damg
same pala with maul and GWF - 5d6 + 3d8 + 3 damg

No feats involved, how does the rerolling break anything? If you didnt allow the rerolls, why on god's earth would I go GWF over SB + defense/duelist (feats aside)?

I don't intend to be rude here, but statements like yours, seem to be what causes the confusion in the first place.

Look at Strills example:

5d8 = 22.5
5d8 reroll 1s and 2s = 26.25

+3.75 damage caused by GWF - Is that broken? Is that not a fair trade for -2 ac (or -3 if sb and defense style)

Again this is about the fighting styles, not the feats related to them, and what they offer as 5e dosen't use alot (any?) prereqs for weaponstyled feats.

[Kryx]

Kryx care to elaborate on how "having it apply to spells and everything else" makes it "far outside the scope of any fighting styles" ?
Archery - flat bonus to attack with ranged weapons
Defense - flat bonus to AC
Dueling - flat bonus to damage with one handed weapons
Great Weapon Fighting - reroll weapon dice (basically flat bonus to damage) when using 2 handed weapons
Protection - You can impose disadvantage on attacks against adjacent allies
TWF - You can apply ability mod on offhand weapon

I think the wording of GWF is quite clear how it is intended to be used - to boost hit/damage just like the other 4 fighting styles (besides protection which is the outlier).



You are playing an oov paladin with a str score of 16, a flametounge wep and hunters mark on your target, and you smite with a lvl 2 slot

A SB with duelist would deal - 4d8 + 3d6 + 5 damg
same pala with maul and GWF - 5d6 + 3d8 + 3 damg
Not using feats heavily favors S&B as GWM is an amazing feat.

For Fighter with all the feats it's a 30% difference between S&B and GWM. I would expect the paladin to be around a 15-20% difference once the full math is done. I'll do the full DPR math in a bit here.



statements like yours, seem to be what causes the confusion in the first place.
Statements like what? Saying it is an outlier? It is.
If you want to argue on math then argue on math. I've happily done so many times on this forum and I provide the most comprehensive DPR numbers of anyone here.

Yet in this case I'm not arguing on numbers because I haven't done the Paladin numbers (I'll do them now).
In this case I'm arguing on RAI that applying it to spells is entirely out of place


Again this is about the fighting styles, not the feats related to them, and what they offer as 5e dosen't use alot (any?) prereqs for weaponstyled feats.
And yet you through all kinda of additives onto the math to make it not about fighting styles. I'll do the full math, but my arguement is based on RAI, not math. Again, if the math is bad then fix the math.

[djreynolds]

GWF and GWM are great. I'm not using -5/+10 with every enemy in a battle. I have only so many smites. I use the Power Attack portion and a great sword for easy hits, and I then use my long sword and shield for cleave portion and smite for the BBG.

And if the BBG is far away and I use a bow and save my spells for later. I got 20 arrows, bound to hit him once. Feats give options as do styles and different ways to damage opponents or defend yourself. Smites run out quick in the early levels, and quicker when you start stacking and using higher spell slots. But if you run out of smites, keep one spell for bless and it helps out with the GWM feat.

[HoarsHalberd]

If you rule that Smite is affected by the crit of the weapon it is associated with then then to be consistent GWF has to apply to smite. "When you roll a 1 or 2 on a damage die for an attack you make with a melee weapon that you are wielding with two hands" It doesn't say the weapon's damage dice, it says any die associated with the attack.

[Lollerabe]

Allright I apologize, I might have been a bit out of line.

And you are right, im throwing on a bunch of additives, flametounge however being the only one you can't rely on.

Yes removing feats favors SB - but that's my point, GWs becomes viable through GWM not through GWF, I don't believe that's RAI at all.

If a DM opt to leave rerolls for weapon damage dice only, then that DM forces his players to take the GWM feat, to make the GW wielder viable. Something that would upset me alot as a player, GWs should be the kingpin of damage regardless of my choice in further customization (feats).

And by that I mean they should be the kingpin by a fair margin.

I think it's fair to say that a Paladin with 20 ac (plate + shield) and a damg interval of 2-10 + mod (duelist)
is alot more attractive to play, than a paladin with 18 ac and a damage interval of 2-12 + mod (rerolling weapon damg only) if you know that the damage advantage of GWF won't scale unless you feat tax it.



And I really disagree on this:

"I think the wording of GWF is quite clear how it is intended to be used - to boost hit/damage just like the other 4 fighting styles"

I think the wording is quiet clear on the exact opposite.

PHB p, 84(GWF): Damage die for an attack you make with a melee weapon.

PHB p, 49 and 169 (Brutal critical,Savage attacker): Weapons damage die/dice.

At least its RAW and I believe RAI as well, otherwise as said GWF falls short.

[GoodbyeSoberDay]

Archery - flat bonus to attack with ranged weapons
Defense - flat bonus to AC
Dueling - flat bonus to damage with one handed weapons
Great Weapon Fighting - reroll weapon dice (basically flat bonus to damage) when using 2 handed weapons
Protection - You can impose disadvantage on attacks against adjacent allies
TWF - You can apply ability mod on offhand weapon

I think the wording of GWF is quite clear how it is intended to be used - to boost hit/damage just like the other 4 fighting styles (besides protection which is the outlier).Going by RAW and letting it apply to bonus damage dice doesn't go beyond the scope of what you're saying. It just adds more damage.

Moreover, if we are to compare GWF to its closest counterpart, Duelist, and restrict it as you suggest, it sucks. In the best case, greatsword, it adds 4/3 damage (plus pocket change from crits) per hit compared to Duelist's flat 2.

Now, I agree that PCs who happen to use great weapons tend to do well, especially early game, where feats like Polearm Master really shine. But GWF style is in no way the reason for their success. It's actually quite tempting to pick defensive fighting style; the only thing that tips the scales back to GWF is bonus damage dice, and even then just barely.

[Kryx]

Going by RAW and letting it apply to bonus damage dice doesn't go beyond the scope of what you're saying. It just adds more damage.
The scope laid out there adds more damage to the weapon. Dueling doesn't apply on smites either.


if we are to compare GWF to its closest counterpart, Duelist, and restrict it as you suggest, it sucks.
I'm not restricting it anymore than you are buffing it. Please don't assume what you think is what the designers intended and then say I'm "nerfing it".
Please see the math below that disputes that GWF sucks.




Math:
DPR of Classes (Paladin) (https://docs.google.com/spreadsheets/d/1d-9xDdath8kX_v7Rpts9JFIJwIG3X0-dDUtfax14NT0/edit#gid=199488908) and graph below.

Paladin GWM does about 25-30% more damage than a S&B build does.
Even taking away -5/+10 would be between 15-20% more DPR than S&B (See houserules version).
To fully isolate it the GWF style is worth ~5-10% more DPR itself, but the concept should be viewed holistically.

Allowing GWF to work on smite alone would make Paladin GWM do about 30-35% more DPR than S&B. It's not mathematically an issue. However I do not think it is RAI.

Both 25-30% and 30-35% are within the same range that a Fighter has between GWM and S&B.

Though these numbers would change by a fair amount if you add other additive dice.

http://i.imgur.com/tYibRW1.png

[Strill]

The scope laid out there adds more damage to the weapon. Dueling doesn't apply on smites either.Why should it? Dueling gets a shield to compensate.


*graph*
Looks perfect! That's what I would expect from 2-handed weapons vs 1h+shield, especially since shield master is also giving your allies advantage, and boosting your DEX saves.


Though these numbers would change by a fair amount if you add other additive dice.Adding more damage dice would actually narrow the gap, not widen it. For example, GWF increases the damage from d8s by 16.7%. If you keep adding d8s, the gap between 2-hand and 1-hand would eventually close in on 16.7%, which is less than what it's at now.

[GoodbyeSoberDay]

The scope laid out there adds more damage to the weapon. Dueling doesn't apply on smites either.Gonna need textual support for that one, because the way I see it, each style provides a bonus whenever you use equipment of a particular type (except Protection, which as you noted doesn't fit the mold either way).
I'm not restricting it anymore than you are buffing it. Please don't assume what you think is what the designers intended and then say I'm "nerfing it".Speaking of assumptions, I never once mentioned designer intent. I quoted RAW and talked about RAW. The text makes it perfectly clear that GWF works with bonus damage dice. From that baseline, making GWF only apply to weapons is indeed a restriction.
Please see the math below that disputes that GWF sucks.Nice graph, but it's the wrong comparison here. Like I already stated, I know that THFers can do well. The problem is that Polearm Master and/or GWM are doing all the work. GWF style itself has to compete with other fighting styles, defensive in particular, because that's the relevant build choice. And if it only applies to weapons (happy with my phrasing?), the obvious choice is +1 AC.

[Lollerabe]

Kryx thanks for providing the math!

I gotta agree with Strill here - looks how it should IMO.

As you said, the GWF is not adding all that much, GWM is.

But as Strill mentioned shieldmaster adds alot as well.
Advantage for every other melee char in the party is hard to place a value on.

Plus shield ac value to dex saves is huge for paladins, especially in conjunction with the 3rd part of the feat (that Strill didnt mention) no damg taken from AOE effects if you succeed on the saving throw, as a reaction.


All in all, I think it's safe to say that allowing GWF to reroll every additive dice dosen't break anything, and is indeed both RAW and RAI.

Good discussion btw:)

[Kryx]

Adding more damage dice would actually narrow the gap, not widen it. For example, GWF increases the damage from d8s by 16.7%. If you keep adding d8s, the gap between 2-hand and 1-hand would eventually close in on 16.7%, which is less than what it's at now.
If you go by the ruling that GWF doesn't affect smite, yes.

Otherwise the DPR of GWM grows - probably staying the same percentage relative to shield as it is now.



Gonna need textual support for that one, because the way I see it, each style provides a bonus whenever you use equipment of a particular type
What? I've already outlined everything above. EVERYTHING but Protection modifies the weapon hit/damage only. None of the others modify anything but weapon hit/damage. That's the scope I've been talking about the whole time.



I quoted RAW and talked about RAW. The text makes it perfectly clear that GWF works with bonus damage dice.
If your only concern is RAW then this conversation does not matter. RAW is there, and it is ambiguous. RAW was praised in 3.X. Happily 5e has moved heavily away from RAW and toward intent.


The text makes it perfectly clear that GWF works with bonus damage dice. From that baseline, making GWF only apply to weapons is indeed a restriction.
I, and others, think the text does not specify that. Even more people think that it applying on smite/spells/etc isn't the intent whether it's RAW or not.

If it was abundantly clear they would've specified. They did not, therefore it is ambiguous and will vary table to table.


The problem is that Polearm Master and/or GWM are doing all the work.
Your assumption seems to be that GWM after feats should be ~50% DPR increase over shield. I find 25-30% a perfectly acceptable range.


GWF style itself has to compete with other fighting styles, defensive in particular, because that's the relevant build choice.
No it doesn't. Every style needs to be looked at holistically. Otherwise Archery is by far the best style in terms of damage. +2 to hit is far better than +2 dmg, or +1.33 damage.


if it only applies to weapons (happy with my phrasing?), the obvious choice is +1 AC.
One could choose that option, yes. No matter if GWF applies on other dice or not +1 AC is likely a better choice. The small DPR difference between it applying on other dice or not is likely too small to matter, really. Though if you stack many items with dice then it matters for sure.

5e is designed around bounded accuracy, and in that system the cost to reward ratio gets more costly the higher in damage you go. Under that paradigm it makes perfect sense for GWF to scale worse than dueling or another middleground option as it is the top damage choice.

[Mara]

I would rule no on smites or flametong or any of that. I find the dpr increase too good to pass up otherwise. I also don't want rogues trying to make two handed finesse weapons just to dip fighter to boost sneak attack.

[Lollerabe]

Mara out of curiosity can I ask you: Did you read the numbers? How does the damage become to good to pass up? It's likely the other way around, if you rule against RAW GWF becomes next to useless.
Look at the graph and the math, GWM is the 'cause of the high dps, not GWF.

And im pretty sure there is no finesse heavy/heavy polearm weapon because of that - it would make dex martials the best option every.single.time.

[GoodbyeSoberDay]

What? I've already outlined everything above. EVERYTHING but Protection modifies the weapon hit/damage only. None of the others modify anything but weapon hit/damage. That's the scope I've been talking about the whole time.Let's go over those styles (except Prot), shall we?
Archery: A bonus to your attack rolls, when using a ranged weapon.
Defense: A bonus to your AC, while you wear armor.
Duelist: A bonus to your damage rolls, when using a one-handed weapon and no other.

Now, I ask again: Is there something in the text, or hell, even in a dev tweet, that would imply they're talking about weapon damage? As opposed to "while using [X equipment]" simply being a requirement for the benefit to apply?
If your only concern is RAW then this conversation does not matter. RAW is there, and it is ambiguous. RAW was praised in 3.X. Happily 5e has moved heavily away from RAW and toward intent.1. The RAW is far from ambiguous. If you'd like to actually debate me on the RAW instead of simply disparaging the words in the rule book as a basis of a rules discussion then please see my reply to MaxWilson.
2. Lame edition wars aside, a change of books didn't stop the fact that RAW + Errata + Word of God are the main things any two tables can expect to have in common. Speaking of Word of God...
3. If you have a tweet from a dev, I'll take one. It looks like Mearls contradicted himself, as is WotC tradition in these matters. But in a case like this, intent is always in the eye of the beholder. That's why people adopted RAW in 3.5 discussions.
I, and others, think the text does not specify that.Stating an opinion without argument isn't very meaningful.
Even more people think that it applying on smite/spells/etc isn't the intent whether it's RAW or not.Based on what, exactly? You yourself have admitted it doesn't even matter that much, balance-wise. You even state later that it's worse than +1 AC when you allow it to apply to bonus damage dice (outside of magic item backed special cases), meaning there's no balance issue at all. So, what's the intent? To make the GWF fighting style (not THFing in general) a crap option, or to make it usually worse than other options and sometimes a bit better?
If it was abundantly clear they would've specified. They did not, therefore it is ambiguous and will vary table to table.They did specify. In the book. Where I quoted upthread.
Your assumption seems to be that GWM after feats should be ~50% DPR increase over shield. I find 25-30% a perfectly acceptable range.Nope. My assumption is that GWF, ideally, is competitive with other relevant fighting styles.
No it doesn't. Every style needs to be looked at holistically. Otherwise Archery is by far the best style in terms of damage. +2 to hit is far better than +2 dmg, or +1.33 damage.Hence why I emphasized +1 AC.
One could choose that option, yes. No matter if GWF applies on other dice or not +1 AC is likely a better choice. The small DPR difference between it applying on other dice or not is likely too small to matter, really. Though if you stack many items with dice then it matters for sure.

5e is designed around bounded accuracy, and in that system the cost to reward ratio gets more costly the higher in damage you go. Under that paradigm it makes perfect sense for GWF to scale worse than dueling or another middleground option as it is the top damage choice.I need some clarification here. I think you're saying that damage options should decrease in efficacy as one takes more and more of them. And hence, GWF should suck, because it's the expensive cherry on top of the damage ice cream sundae for those people who took everything possible to increase damage. In short, optimization should favor balance of defense/offense over specialization in one or the other.

If that's what you're saying... I guess we have to agree to disagree. I think (relevant) options should be relatively equal in value, when you can help it, and that means GWF providing as much value as +1 AC, for two-handed melee weapon wielders.
I would rule no on smites or flametong or any of that. I find the dpr increase too good to pass up otherwise. I also don't want rogues trying to make two handed finesse weapons just to dip fighter to boost sneak attack.No worries on that front. GWF isn't that big no matter what, and your rogue should never be dipping a class just to get it, from an optimization perspective.

[Kryx]

Mara out of curiosity can I ask you: Did you read the numbers? How does the damage become to good to pass up? It's likely the other way around, if you rule against RAW GWF becomes next to useless.
Look at the graph and the math, GWM is the 'cause of the high dps, not GWF.
Next to useless is an exaggeration. Again, it must be looked at holistically, not just on the fighting style otherwise Archery is the best by far.




...
Ignoring your passive aggressive rudeness, let's look at the actual words:

When you roll a 1 or 2 on a damage die for an attack you make with a melee weapon that you are wielding with two hands, you can reroll the die and must use the new roll. The weapon must have the two-handed or versatile property for you to gain this benefit.
I have bolded the important words. When wielding a melee weapon in two hands you can reroll "the die". "the die" in this case would presumably be talking about the weapon as that is what is being modified. It doesn't say "all die". I can see how it can be read both ways.

Really there isn't much more to debate here - no need to go line by line.
You, and some others, believe RAW says one thing. I, and some others, believe it says another. I'm not here to debate RAW as that argument cannot be won as it's ambiguously defined.

Please stop being passive aggressive about it - your opinion of what RAW says is not the default and I don't have to disprove it. It's impossible to prove or disprove.

[Lollerabe]

Indeed no need to be rude or passive aggressive, doubt he was tho'.

Next to useless might be an exaggeration, but Kryx can I ask a few questions then?

Would you yourself ever go for GWF, if no additive dice rerolls were allowed, as opposed to defense style? (bearing in mind you can go GWM with either option)

From a mechanical perspective, based on your own math - do you believe that allowing the rerolls of additive dice breaks anything? Does it blow other options out of the water then? (as in simply making SB an inferior option almost always)

IF we look completely away from RAW or RAI and put egos aside, would I as player at your table, be allowed to reroll additive dices - if it meant a great deal for me, in regards to character creation and fluff?

It might not need to be said, but im genuinely keeping this discussion going because i'm interested in it - not 'cause I want to "win" the debate.

[MaxWilson]

Are there free actions anymore? I did not think so...

I'm not using the word in a technical sense. There are things you can do, like the Ranger's "Horde Breaker" attack or making a saving throw to grab something when hit by Reverse Gravity, which do not cost a bonus action.


Also consider what you are saying if a rogue has two attacks per round and hits with his first attack on one target but ants to save his sneak attack for his second it is clear he gets to choose whether or not to use his SA dice after he hits. Your reasoning would make SA useless since he chooses after he hits which we know as false.

I'm not sure what you're arguing here, but if you think that the rogue gets to roll both attacks before deciding which one is a sneak attack then I disagree. He has to decide when he hits.

[Kryx]

Would you yourself ever go for GWF, if no additive dice rerolls were allowed, as opposed to defense style? (bearing in mind you can go GWM with either option)

From a mechanical perspective, based on your own math - do you believe that allowing the rerolls of additive dice breaks anything? Does it blow other options out of the water then? (as in simply making SB an inferior option almost always)
Well my choices would entirely be based on the math, so I'm probably not the best candidate to ask.

I likely wouldn't ever choose GWF over Defensive because I think +1 AC is more valuable than 5-10% more DPR. However some people may prefer the damage.

If GWF is broken then it needs to fixed for ALL classes. A Fighter is in the same situation of GWF not being the best option. Allowing GWF to effect other dice doesn't fix it for Fighter and Paladin. As a baseline allowing it on smite would improve a Paladin's DPR by about 5% based on my numbers above.

An option to fix it for everyone would be to make it always reroll, not just once. That would make it worth it.
Plus you'd have to make a Greataxe reroll 1-5 to make it equal to 9 average damage - the same as a Greatsword ignoring 1-2 on 2d6.


would I as player at your table, be allowed to reroll additive dices - if it meant a great deal for me, in regards to character creation and fluff?
Nope - I would find an alternate solution that fixed it for both the Fighter and the Paladin. The option above would do so.


It might not need to be said, but im genuinely keeping this discussion going because i'm interested in it - not 'cause I want to "win" the debate.
100% agreed. I much prefer productive discussion.

[bid]

Paladin GWM does about 25-30% more damage than a S&B build does.
Even taking away -5/+10 would be between 15-20% more DPR than S&B (See houserules version).
To fully isolate it the GWF style is worth ~5-10% more DPR itself, but the concept should be viewed holistically.

Allowing GWF to work on smite alone would make Paladin GWM do about 30-35% more DPR than S&B. It's not mathematically an issue. However I do not think it is RAI.
[/IMG]
See, this is why the best graph for this is "percent of S&B", where S&B would stay at 100% throughout the graph. I think it would yield a much better picture with a GWF-all and GWF-weapon.


It would be fairer to compare damage from GWF paladin to GWF fighter. Their 3rd and 4th attack means even S&B fighter scales up by 4 points, whereas improved divine smite would only adds 3/4. If anything, it's S&B pally which is the outlier.

Moreover, extra damage from lack of shield scales up. Another argument for smite to scale up.

[GoodbyeSoberDay]

passive aggressive rudenessBaseless ad hominem
I have bolded the important words. When wielding a melee weapon in two hands you can reroll "the die". "the die" in this case would presumably be talking about the weapon as that is what is being modified. It doesn't say "all die". I can see how it can be read both ways.My first thought is, "Greatsword."

More importantly, the text you bolded first describes "a damage die," and then, when referring to that particular damage die, says "the die." Hence, it is referring to damage dice in general.
Really there isn't much more to debate here - no need to go line by line.
You, and some others, believe RAW says one thing. I, and some others, believe it says another. I'm not here to debate RAW as that argument cannot be won as it's ambiguously defined.I find that the RAW in this edition is quite often ambiguous, but it is not here.
Please stop being passive aggressive about it - your opinion of what RAW says is not the default and I don't have to disprove it. It's impossible to prove or disprove.You don't have to do anything, including taking offense at a simple set of objections to your arguments.

[Kryx]

See, this is why the best graph for this is "percent of S&B", where S&B would stay at 100% throughout the graph. I think it would yield a much better picture with a GWF-all and GWF-weapon.
Your wish is my command:
http://i.imgur.com/phoK81h.png
http://i.imgur.com/X18Uu7G.png

Allowing Smite to be affected by GWF doesn't really fix much at all in comparison to S&B.

[Kryx]

Here is what it would look like if GWM was 9 average (always reroll 1 and 2 on a greatsword, or 1-5 on a greataxe):

http://i.imgur.com/1AcMi7v.png
http://i.imgur.com/2pNeKcI.png

[Lollerabe]

Okay so your answer confused me a bit.

You would never opt for a 5-10% damg increase over +1 ac on a melee martial. I think it's safe to say that within bounded accuracy, few of us would.

Even though you don't say it outright, im taking a leap, and assuming you don't find rerolling additive dice to be broken at all.

Maybe im a bit slow today, but your suggestion to GWF would be:

Greatsword/maul: Whenever you roll less than 3 on a WEAPON damage die, you can treat it as a 3.

Greataxe: Whenever you roll less than 5 on a WEAPON damage die you can treat it as a 5.

Is this correct? Thus only applying the effect to weapon damage die/dice ?

I know you stated you would allow rerolling always/more than once(as long as it is a 1-2 on d6 and 1-5 on d12), but applying this technique, might prove extremely tidious at a table, so above "translation" gets the gist of it I hope.

So in short we are having a RAW vs RAI debate more than a balance/RAF debate at this point.

So to actually help out OP, if homebrew GWF isn't a option, can we atleast agree that rerolls on additives dosen't break anything DPR wise, and that the assumption that it does is a fallacy?

(im asking this since you can see from Maras post, that people assume that it actually makes GWF overpowered. Which might lead to DM's ruling against GWF rerolls based on nothing but empty hearsay)

[bid]

Your wish is my command:
Allowing Smite to be affected by GWF doesn't really fix much at all in comparison to S&B.
Looks like allowing rerolls on smite helps keep a constant ratio.

The drop in ratio is because of the -5/+10 of GWM, no?



Here is what it would look like if GWM was 9 average (always reroll 1 and 2 on a greatsword, or 1-5 on a greataxe)
Do you mean perma-reroll, aka 2d4+4?


In both cases, the dip at level 5 is weird. Is it because of shield master proning the target?

And the rebound at level 6 is even weirder. Looking at the source data, why does GWM starts there?

[Kryx]

Greatsword/maul: Whenever you roll less than 3 on a WEAPON damage die, you can treat it as a 3.

Greataxe: Whenever you roll less than 5 on a WEAPON damage die you can treat it as a 5.

Is this correct? Thus only applying the effect to weapon damage die/dice ?

I know you stated you would allow rerolling always/more than once(as long as it is a 1-2 on d6 and 1-5 on d12), but applying this technique, might prove extremely tidious at a table, so above "translation" gets the gist of it I hope.
Forget what I said above, it's too complicated. I would probably make the fighting style add 2 damage and fix greataxe by making it "brutal 1" which always rerolls 1s. It would then average as 7 and with the 2 from the fighting style would be 9 as well.


So in short we are having a RAW vs RAI debate more than a balance/RAF debate at this point.
A balance discussion is healthy.


So to actually help out OP, if homebrew GWF isn't a option, can we atleast agree that rerolls on additives dosen't break anything DPR wise, and that the assumption that it does is a fallacy?
For the base smite there is no mathematical problem. Adding in more dice like hunter's mark or flametongue would heavily alter that.




Looks like allowing rerolls on smite helps keep a constant ratio.
It entirely depends based on the magical items given out, as mentioned above.


The drop in ratio is because of the -5/+10 of GWM, no?
In both cases, the dip at level 5 is weird. Is it because of shield master proning the target?
And the rebound at level 6 is even weirder. Looking at the source data, why does GWM starts there?
I actually don't have -5/+10 in those numbers. As you have guessed it is because S&B gets a second attack at 5th level so their proning strategy peaks at that point.
Fighter gets a feat at 6th level is why it starts there for paladin, but that's wrong.
You should effectively ignore level 5&6 - you get the idea and it would take me too long to fix it.

[Xetheral]

I read GWF as applying to the attack's damage dice. Smite explicitly adds to the attack's damage dice (rather than being a separate source of damage). Therefore, GWF applies to the smite dice.



Math:
DPR of Classes (Paladin) (https://docs.google.com/spreadsheets/d/1d-9xDdath8kX_v7Rpts9JFIJwIG3X0-dDUtfax14NT0/edit#gid=199488908) and graph below.

Kryx, I took a look at your spreadsheet and believe there are two errors relating to how you are handling Paladin damage calculations.

First, your field "Smite Dmg/Use" takes the total smite damage for the day and divides it by the expected number of rounds per day, applying that total as a constant amount of damage each round. This overlooks the fact that smite damage will often be applied on a crit, and thus frequently will be doubled.

Second, you're applying both GWM and Smite independently of each other, ignoring their interaction. In particular, against an opponent dangerous enough that the Paladin wishes to use smite, they will not want to use GWM in the same round because it will decrease their average damage for that round. (Link in my signature has the specific breakpoints beyond which the potential damage (such as smite) is high enough that the Paladin won't want to use GWM for a particular hit chance.) Put another way, your handling of Smite and GWM is calculating the average DPR assuming the Paladin is trying to maximize his total damage per day, and thus doesn't care what or when he smites. This is different from calculating the average DPR assuming the Paladin is trying to maximize DPR on a round-by-round basis.

I didn't take a long look, so if I missed the place where you take these factors into account, please let me know.

[Kryx]

First, your field "Smite Dmg/Use" takes the total smite damage for the day and divides it by the expected number of rounds per day, applying that total as a constant amount of damage each round. This overlooks the fact that smite damage will often be applied on a crit, and thus frequently will be doubled.
A crit would happen 1.25 times in a typical day. I've put it in, but smite at 20 would change from 10.1 to 10.6 DPR. It's so negligible. I don't use crit fishing smites in my home games.
It would happen more than that as there are between 2 and 3 attacks per round. So about 3 times per day which is still incredibly small.
EDIT: Calculated smite crits. Went from 10.1 DPR to 11.2 DPR at 20. Incredibly small.


against an opponent dangerous enough that the Paladin wishes to use smite, they will not want to use GWM in the same round because it will decrease their average damage for that round.
DPR is Damage Per Round. If it only works on 1 round then it is burst damage which is an entirely different statistic.
DPR over a day assumes an average AC at each level as defined in the "Scaling" sheet. Choosing when to use -5/+10 and when not to based on the enemy AC is not part of the scope of this DPR sheet as it would MASSIVELY complicate the issue.
That said if we take GWM Fighter at 20 for example.
Enemy AC of 19, -5/+10. DPR is 72
Enemy AC of 19, normal. DPR is 64
Enemy AC of 21, -5/+10. DPR is 56
Enemy AC of 21, normal. DPR is 56
Enemy AC of 23, -5/+10. DPR is 39
Enemy AC of 23, normal. DPR is 47

So the breaking point where GWM is no longer worth it is for enemies with 22 AC or higher.
That's a 6 Ancient Dragons (Black, Blue, Red, Bronze, Gold, Silver)[22], the Empyrean[22], and the Tarrasque [22].
That's 8 monsters.

The numbers probably differ slightly on the Paladin. In his case the breakpoint is AC 21 or higher which includes the 8 monsters from above plus Solar and Ancient Green/Copper dragons.
That's 11 monsters total.

This extreme edge case isn't worth calculating. Plus I don't use -5/+10 so those numbers are only there for people to see the RAW.

[Lollerabe]

So can I ask another favor?

Can you do the math for damage difference between the GWF pala, and the SB/duelist pala I gave as an example earlier?

The one with flametounge, lvl 2 smite, hunter's mark, 16 str etc.

Since your saying that allowing rerolls on additives beyond smites, would inflate the % number by a large margin - i'd like to see how much the difference is, with no rerolls on additives and rerolls on ALL of them.

That would be awesome, since im not known as a mathmagician :)

[bid]

First, your field "Smite Dmg/Use" takes the total smite damage for the day and divides it by the expected number of rounds per day, applying that total as a constant amount of damage each round. This overlooks the fact that smite damage will often be applied on a crit, and thus frequently will be doubled.
Actually with 35% hit for GWM, that's 1/7 crit or +9/7 damage per attack. That's extra 2.5 DPR (2.7 with rerolls).

At 60% for S&B, 9/12 is 1.5 DPR.

[bid]

Since your saying that allowing rerolls on additives beyond smites, would inflate the % number by a large margin - i'd like to see how much the difference is, with no rerolls on additives and rerolls on ALL of them.
Napkin approximation, reroll add:
- 1/2 to 1d4 or a 25% boost
- 2/3 to 1d6 or a 20% boost
- 3/4 to 1d8 or a 17% boost
- 4/5 to 1d10 or a 15% boost
- 5/6 to 1d12 or a 13% boost

If hunter's mark represents 25% of the damage, the boost would be {25% of 20% = 5%}

[Kryx]

The one with flametounge, lvl 2 smite, hunter's mark, 16 str etc.
I made a graph that scales the whole way (starts using hunter's mark and Flame Tongue at level 2)

http://i.imgur.com/RfBtHSL.png

[Strill]

If you go by the ruling that GWF doesn't affect smite, yes.

Otherwise the DPR of GWM grows - probably staying the same percentage relative to shield as it is now.That's a mathematical impossibility. If you keep adding damage dice, the proportional damage difference between each weapon will eventually approach the contribution from GWF. Since GWF's proportional damage difference (20% max) is LESS than the current difference (~30%), that means the proportional difference can only shrink.

[Kryx]

That's a mathematical impossibility.
Percentages are really a poor way to model it. See the DPR graph above for how it turned out.

[Pex]

If your only concern is RAW then this conversation does not matter. RAW is there, and it is ambiguous. RAW was praised in 3.X. Happily 5e has moved heavily away from RAW and toward intent.


I, and others, think the text does not specify that. Even more people think that it applying on smite/spells/etc isn't the intent whether it's RAW or not.

If it was abundantly clear they would've specified. They did not, therefore it is ambiguous and will vary table to table.



Which is why I do not find it "happily" 5E moved away from RAW and toward intent. I should not have to relearn the game just because I play with a different DM.

[Strill]

Percentages are really a poor way to model it. See the DPR graph above for how it turned out.

Why are they a bad way to model it? Percentages are all I care about. As long as 2-handed weapons do around 20% to 30% more than 1h + shield, I'm fine with it.

For the graph, Does the line labeled "Paladin GWM" include great weapon fighting as well? If not, then the comparison isn't particularly relevant since you can't tell how flametongue weapons and hunter's mark specifically affect the results.


Which is why I do not find it "happily" 5E moved away from RAW and toward intent. I should not have to relearn the game just because I play with a different DM.

You've got it backwards. A game adjudicated by intent is what is going to give you different results with different DMs. A game adjudicated by RAW will give you the same results.

[McNinja]

No, it isn't. It's damage for a smite. Smiting is clearly not an attack because it doesn't fit the criteria on PHB 194: there is no attack roll, so it's not an attack. It's just a free action that happens after a successful attack.

If you had to declare smites before you made your attack roll I might view it differently, as part of the attack, but RAW it clearly is not.
So you're telling me that Smite is not damage? It's part of the attack, I fail to see how rolling dice for damage isn't actually rolling dice for damage.

[Pex]

You've got it backwards. A game adjudicated by intent is what is going to give you different results with different DMs. A game adjudicated by RAW will give you the same results.

No, that's precisely what I mean. I want the rules to give the same results table to table. I accept house rules. Those are incidentals. I'm talking about the rules of the game themselves.

[coredump]

GWF says you reroll 1 or 2 on "a damage die of an attack"

Criticals say you roll "the attacks damage dice" twice

They both use the same terminology, they refer to the damage dice of an attack. So anything that would not benefit from GWF would also not benefit from a Critical hit either. If Smite is not part of the Damage dice, than neither is Sneak Attack; thus neither one should benefit from Criticals. Of course, the PHB says that SA gets doubled....so we know that isn't true.

___________________

We can easily compare GWF usage of the term "damage die" compared to Half Orcs Savage Attacks which states "Weapon's damage die". The authors understand the difference and applied it when appropriate.

________________
GWF gives a smaller bonus than Duelist or even TWF. TWF can give up to a +5 to a single attack, Duelist gives +2 to every attack. GWF gives a polearm a mere 0.8 bonus, and a greatsword 1.33 bonus.
This becomes even worse when you look at percent increase. Duelist gives 44%, GWF gives 14.5% and 19%

The part people continually skip over, is the change to *max* damage. Duelist makes a longsword go from 8 max, to 10 max. GWF makes a polearm go from 10 max to..... 10 max.

So GWF is clearly an inferior choice..... the reason is because it is designed and written to apply to all "Damage dice" in the attack. So while it is normally inferior, it adds some benefits along with things like Hunter's Mark, or Smite, or Sneak Attack, etc....

[Strill]

No, that's precisely what I mean. I want the rules to give the same results table to table. I accept house rules. Those are incidentals. I'm talking about the rules of the game themselves.

Exactly. So if you go by "intent", which is subjective, you'll get different results from table to table. If you go by RAW, you get the same results.

[CNagy]

The RAW is entirely unambiguous. Attack's damage dice means every single die you roll to determine damage for a single attack. Not liking that doesn't make it uncertain. Mearls has answered both in the affirmative and in the negative, but Mearls rulings are always how he would rule at a table, not how he'd rule according to RAW. Unless Jeremy Crawford comes out and says that GWF applying to more than weapon damage dice is unintentional, then we have no RAI to go by and only the RAW, which clearly applies the reroll mechanic to every die rolled to determine damage for an attack.

[Lollerabe]

Strill arent you confused here? Pex wrote that he DOSEN'T want the game to be run by intent - since that would leave to different tables ruling differently. he DOES want RAW to rule - as that would give a ruleset set in stone, I think you both agree here.

Anyway to sum up what i've learned (and thanks for that guys):

1. Rerolling additives is indeed RAW - arguing that it isn't, requires some imaginative interpretation of the rules.

2. Rerolling additives dosen't break anything and isn't OP.
It ensures that GWF is somewhat able to compete with other fightingstyles, and in reality it's not a mechanic that you will see utilized that often regardless.

3. Make the ruling at your table based on RAW, and if not at the very least, look at numbers and actual math before you outright go against RAW.
4. I suck at reading graphs.
5. I haven't changed my opinion whatsoever, but at least I can say that it is because I honestly believe i'm right, not just because i'm stubborn or lazy.

Anyway thanks for a healthy, well mannered debate so far guys, not something I'm used to on internet forums (I play dota... ^^ )

Cheers!

[GoodbyeSoberDay]

2. Rerolling additives dosen't break anything and isn't OP.
It ensures that GWF is somewhat able to compete with other fightingstyles, and in reality it's not a mechanic that you will see utilized that often regardless.This is the funniest part for me. All this talk of game balance, and +1 AC is still better for the PM/GWM paladin, except for the niche case where you've loaded up on bonus damage dice, in which case GWF is slightly better.

For me, it's tempting to house rule GWF to have a flat +2 bonus to damage simply for speed of play.

[Malifice]

This is the funniest part for me. All this talk of game balance, and +1 AC is still better for the PM/GWM paladin, except for the niche case where you've loaded up on bonus damage dice, in which case GWF is slightly better.

For me, it's tempting to house rule GWF to have a flat +2 bonus to damage simply for speed of play.

Go with 'double Str bonus'. It's just as quick, thematically appropriate and has precedent.

[Rusty Killinger]

Sage Advice is inconsistent : No (http://www.sageadvice.eu/2014/10/06/great-weapon-smite/) Yes (http://www.sageadvice.eu/2015/09/09/great-weapon-fighting-and-smite/)

Both from Mike Mearls, even.

Ughh.. I don't miss much about 4e, but they could write a consistent and clear rule set back then.

[GoodbyeSoberDay]

Go with 'double Str bonus'. It's just as quick, thematically appropriate and has precedent.That's too good for a feature like fighting style, though.

[Lollerabe]

Indeed GoodbyeSoberDay indeed..

If I have to take a guess as why anyone at any table ever deemed the rerolling OP, I think it because of paladins and crit mechanics - NOT the GWF reroll.

If we take my example from earlier.

Flametounge, hunters mark, lvl 2 smite, GWF with greatsword/maul:

5d6 + 3d8 dice gets thrown, that's alot of dice now say said paladin was critfishing.

That attacked turned into - 10d6 + 6d8 dice, at this point the entire table went "Sweet baby jesus! What the holy hell!?" The fact that the player then rerolled lets say 4-5 of those dice properly didn't change the damage by a huge margin.
But the sheer chaos and waow factor of 16 dice rolled, led the table to believe that the OHKO on the BBEG was caused by GWF rerolls.
When in reality any SB paladin could throw those 16 dice(well 14 in that case) and mathematically speaking, would have a higher chance of doing so, if he had the shieldmaster feat and had succes on his bonus action prone attack (im talking about the increase in crit chance from advantage here to make myself clear).

So yeah I think its common in D&D that things get a OP/Broken status based on selective memory at the table, you remember the paladin onehitting the young dragon, not all the times smite was a suboptimal option and rerolling didn't do squat.

[Kryx]

1. Rerolling additives is indeed RAW - arguing that it isn't, requires some imaginative interpretation of the rules.
That seems to be the consensus now, yup.

2. Rerolling additives dosen't break anything and isn't OP.
It ensures that GWF is somewhat able to compete with other fightingstyles
5. I haven't changed my opinion whatsoever, but at least I can say that it is because I honestly believe i'm right, not just because i'm stubborn or lazy.
This is a stretch that is unproven. GWF applying on additives definitely helps it. We still haven't reached the conclusion of "is GWF worth it over choosing defensive style with the GWM feat".

at the very least look at numbers and actual math before you outright go against RAW.
100% agreed. Informed opinions based on math are much more valuable than play experience.



Go with 'double Str bonus'. It's just as quick, thematically appropriate and has precedent.
Adding Strength bonus again would range from +3 to +5 damage - that's way too much. My math below adds 2 and that would match up with +1 AC.



Math:
The Fighter gets a 7-8% boost in DPR by choosing GWF over Defensive (+1 AC).
The Paladin gets a 9-10% boost in DPR by choosing GWF over Defensive (+1 AC).
Paladin assumes that GWF applies on smite as discussed.
http://i.imgur.com/9JkmCXy.png

The value of +1 AC is variable depending on your current AC and the enemy's attack. If you previously had a 40% chance to be hit and now have a 35% chance to be hit for instance then out of 100 times you'd be hit 88 times now. (really good improvement). Closer to the middle 50% vs 45% is 90 times out of 100.

So we're comparing 7-8% or 9-10% more DPR vs 10-12% more defense. In this case defense always wins mathematically.

If GWM were +2 damage then GWF for Fighter becomes 10-12% more DPR. However Paladin then does 7-10% (slowly dropping as levels increase) instead of 9-10%.
http://i.imgur.com/FEvOjkv.png
The percentages there are in comparison to the defensive style.
My goal would be to get the Paladin to match Fighter's percentages of 10-12%. Maybe +2 and allow additive dice to reroll 1s and 2s once.
http://i.imgur.com/ULtNYaG.png
In this version Paladin is 20-24% until 6 and then they stay at 12% for the rest of the levels. Fighter is 17-20% until 4 and then they are 12% scaling down to 10%.

I think that is the most balanced option:
"When you are wielding a melee weapon in two hands, you gain a +2 bonus to damage rolls with that weapon. When you roll a 1 or 2 on any additive damage die for an attack you make with this weapon, you can reroll the die and must use the new roll, even if the new roll is a 1 or a 2."
I've verified that this does not offset the current GWM vs Polearm balance.
That plus allowing a Greataxe to always reroll 1s would balance the whole spectrum.

[Lollerabe]

I gotta love you Kryx, you've been going back and forth with me on this subject, yet you end up suggesting a GWF version that's the same as my interpretation of the current, only better :D

Jokes aside - that seems legit.

Problem remain the same as ever tho', while people agree that homebrew and houseruling is fine (often even required) many people seem to squirm a bit, when it comes to actually doing it.

I'm definitely not a fan of the double strength bonus, reminds me of a dark and cumbersome time they call *spooky voice* 3,5e....

[McNinja]

I like the idea, but for a Fighting Style I feel like it should be simpler. Maybe just a flat +3 to damage rolls and any 1s you roll are treated as 2s?

that gives the greataxe a range of 5-15
and gives a gs or maul a range of 9-15

... in a perfect world both would be even. The main benefit of using a gs/maul over the ga is the higher minimum damage.

[Kryx]

I gotta love you Kryx, you've been going back and forth with me on this subject, yet you end up suggesting a GWF version that's the same as my interpretation of the current, only better :D
Nice working through this with you as well - you've been respectful in disagreement and curious to see the actual numbers.
I always try to make suggestions/houserules based on math. I was less articulate than I wanted to be earlier - GWF the fighting style isn't great, but how it currently works must be fully understood before any changes were to be made. And buffing it must be done only with the full numbers (which I've now given).

The trick is to make something that works for the Fighter (who is dependent on his weapon damage) and for the Paladin (who is dependent on his smite damage). This captures that. My opinion on GWF applying to smite had to change as a result of the numbers - which matter more than how I think it should work.


People do seem adverse to houserules, and yet make them all the time. Go read a thread about Short Rest frequency and you'll see people are massively unbalancing the game by ignoring short rests, and yet they are so incredibly resistant in other areas - it's quite silly.



I should really publish a full list of balance changing houserules and (math) reasons for them. I have them collected, but not with charts and stuff.

[Kryx]

Maybe just a flat +3 to damage rolls and any 1s you roll are treated as 2s?
That's significantly more than what I suggested. GWF would then outpace Defensive style by a fair margin.

As I said above the trick is to support both the Fighter (based on weapon damage) and the Paladin (based more on smite damage). Plus balancing Greataxe vs Greatsword - but that should be done for all classes, not just Paladin and Fighter.

[Strill]

People do seem adverse to houserules, and yet make them all the time. Go read a thread about Short Rest frequency and you'll see people are massively unbalancing the game by ignoring short rests, and yet they are so incredibly resistant in other areas - it's quite silly.
A lot of it depends on simplicity. People are averse to complex solutions. Throwing out short rests may be terrible for balance, but it's simpler. Likewise while your GWF change may be more balanced, it's also more complex. You have to remember which damage dice you get to reroll, and if you don't get bonus damage dice on a regular basis, it's easy to forget the perk and not apply it when it becomes relevant.

[Lollerabe]

That's a great idea, take your time and make a "Tweak/remakes, the math and why" thread.

I think that there are a bunch of players out there like me, who have a problem with a lot of the rules (twf im looking at you) but don't wanna post remake suggestions since they don't know the ramifications of said remakes.


Just keep the numbers legit, and make sure to spell out that any suggestion has to be backed by math, not by gut feeling or player experience.

[Kryx]

Likewise while your GWF change may be more balanced, it's also more complex. You have to remember which damage dice you get to reroll, and if you don't get bonus damage dice on a regular basis, it's easy to forget the perk and not apply it when it becomes relevant.
Well in my case it would only apply to things like smite, hunter's mark, etc, but I do see how it could easily be forgotten.
GWF by default has complex wording, but in practice it's simple. I'm open to suggestions if you have a way of making it simpler. I'll see if I can make it more simple.
The problem is rerolling the main weapon will inherently have different results for Greatsword vs Greataxe.


That's a great idea, take your time and make a "Tweak/remakes, the math and why" thread.

I think that there are a bunch of players out there like me, who have a problem with a lot of the rules (twf im looking at you) but don't wanna post remake suggestions since they don't know the ramifications of said remakes.

Just keep the numbers legit, and make sure to spell out that any suggestion has to be backed by math, not by gut feeling or player experience.
I meant more "Here are my fixes with reasons based on math". I wouldn't mind other contributors, but they would need to provide good math. Many people try to make napkin math on the forum and that cannot possibly capture the whole picture.

[Lollerabe]

That's what I meant - your suggestions with all the numbers etc. And (if even) a statement in bold text that explains the rules for posting remake suggestions in that thread fx:

Your numbers need to add up, if it's a ability/spell/mechanic accessible to multiple classes, it has to be balanced for them all.

Or whatever you think'll fit.

Anyway, I think it's a great idea.

[Kryx]

Simpler version of above:
"When you attack with a melee weapon that you are wielding in two hands, you gain a +2 bonus to damage rolls with the main end of the weapon and a +1 bonus to damage with any other damage dice associated with the attack."
That means main weapon gets +2 damage. Polearm Master bonus attack, smite, hunter's mark, etc get +1.

Still not as simple as people would probably like, but the balance here is very tricky.

Paladin GWF would range from 13-14%, but I think that's acceptable.

[coredump]

I meant more "Here are my fixes with reasons based on math". I wouldn't mind other contributors, but they would need to provide good math. Many people try to make napkin math on the forum and that cannot possibly capture the whole picture.
Thats kind of funny... because you seem to equate 'good math' with 'lots of colored graphs'. On a number of occasions I have tried to get the basis of your 'conclusions' from you, and it has been like pulling teeth. The one time I did get the info, I found numerous errors in your assumptions used. So while I agree on the need for 'good math', I do not trust your colorful graphs to provide it.

[coredump]

Simpler version of above:
"When you attack with a melee weapon that you are wielding in two hands, you gain a +2 bonus to damage rolls with the main end of the weapon and a +1 bonus to damage with any other damage dice associated with the attack."
That means main weapon gets +2 damage. Polearm Master bonus attack, smite, hunter's mark, etc get +1.

Still not as simple as people would probably like, but the balance here is very tricky.

Paladin GWF would range from 13-14%, but I think that's acceptable.

And thats your *simple* version?

I still don't think you have made your case for why a change is needed, and what you have said contains a plethora of assumptions that you have never bothered to support. Yet you still use them as 'reasons' for needing a much more complicated system.

[Kryx]

Thats kind of funny... because you seem to equate 'good math' with 'lots of colored graphs'.
The actual numbers are readily available on my spreadsheet. But no one will look at them. I've tried in the past to direct people to it, but they will not look. Therefore I post graphs. Please correct any math issues you see.


On a number of occasions I have tried to get the basis of your 'conclusions' from you, and it has been like pulling teeth. The one time I did get the info, I found numerous errors in your assumptions used. So while I agree on the need for 'good math', I do not trust your colorful graphs to provide it.
I do not remember the specifics of your past complaints. Though several have tried to make DPR do something that it cannot do. I have made significant improvements to Fighter, Barbarian, and added Paladin in the last week so maybe your previous issue is gone.
Seriously, please do correct any issues or assumptions.


And thats your *simple* version?
It's nearly the exact same complexity as the RAW version.
RAW: Reroll all 1s and 2s on damage dies associated with this weapon. Do not reroll more than once.
My version: Add 2 damage damage to the main weapon damage and 1 to any additional damage dice.

Again, I'd happily alter it to something simpler if the math supported that. The problem is it needs to work for Fighter and Paladin.


I still don't think you have made your case for why a change is needed
I made the case above. Please, sincerely please, point out any issues for it. GWF is an inferior choice for the Fighter and Paladin in comparison to the Defensive style.
Defensive style makes you less likely to get hit by about 10-12% depending on the base value.
GWF itself increases DPR by 7-8% for the Fighter and 9-10% for the Paladin.

Maybe that number is small enough to ignore, but you'll see people complaining about it all throughout this thread.


what you have said contains a plethora of assumptions that you have never bothered to support. Yet you still use them as 'reasons' for needing a much more complicated system.
Cut the aggression, dude. You're respected here and on enworld. No need to act this way. Please stop the attacks and address any specific issues that you have.

[Xetheral]

A crit would happen 1.25 times in a typical day. I've put it in, but smite at 20 would change from 10.1 to 10.6 DPR. It's so negligible. I don't use crit fishing smites in my home games.
It would happen more than that as there are between 2 and 3 attacks per round. So about 3 times per day which is still incredibly small.
EDIT: Calculated smite crits. Went from 10.1 DPR to 11.2 DPR at 20. Incredibly small.

Don't forget that when critting one can use the highest available spell slot to maximize the gain. That will skew the average higher.


DPR is Damage Per Round. If it only works on 1 round then it is burst damage which is an entirely different statistic.

Damage per round can be optimized for different goals. Your measure of DPR is (Total Damage per Day)/(Number of Rounds per Day). You're therefore assuming that DPR-maximizing characters try to maximize their total damage dealt over the course of the day. It would also be perfectly reasonable to measure DPR as the average of the per-round damage, instead assuming that characters are trying to maximize the damage they do on a particular round. Despite a basic mathematical equivalency, these two measures of DPR will only provide the same result if the players make identical choices. Since each involves trying maximize a slightly different number, the incentives are different and therefore the two measures of DPR will give different results.

For example, trying to maximize total damage per day results in a player with divine smite not caring which enemies the ability is used on. Therefore, accepting the -5/+10 on every attack may still result in a sufficient number of hits to expend all the available spell slots on smiting. By contrast, trying to maximize damage in a particular round will likely result in the -5/+10 being applied only against particular easy-to-hit enemies (or only when one has advantage). Therefore, the two different approaches will give significantly different raw values for the expected DPR gain from using either Divine Smite or GWM, and particularly different raw values when trying to evaluate their interplay.


DPR over a day assumes an average AC at each level as defined in the "Scaling" sheet. Choosing when to use -5/+10 and when not to based on the enemy AC is not part of the scope of this DPR sheet as it would MASSIVELY complicate the issue.

I agree that strategic use of the -5/+10 is likely too complicated to effectively model. But while that may make the decision to not include that factor a very reasonable modeling decision, it still leaves your model ineffective at capturing the highly-relevant interplay between Divine Smite and GWM in the context of a discussion on the mathematical import of Divine Smite. Ultimately, it makes your model less useful to this discussion than it is to its primary purpose of comparing the classes' DPR in the abstract.


That said if we take GWM Fighter at 20 for example.
Enemy AC of 19, -5/+10. DPR is 72
Enemy AC of 19, normal. DPR is 64
Enemy AC of 21, -5/+10. DPR is 56
Enemy AC of 21, normal. DPR is 56
Enemy AC of 23, -5/+10. DPR is 39
Enemy AC of 23, normal. DPR is 47

So the breaking point where GWM is no longer worth it is for enemies with 22 AC or higher.
That's a 6 Ancient Dragons (Black, Blue, Red, Bronze, Gold, Silver)[22], the Empyrean[22], and the Tarrasque [22].
That's 8 monsters.

The numbers probably differ slightly on the Paladin. In his case the breakpoint is AC 21 or higher which includes the 8 monsters from above plus Solar and Ancient Green/Copper dragons.
That's 11 monsters total.


I think your work making DPR charts to compare the classes is fantastic. It was a ton of work and produced a very thought-provoking analysis.

On the other hand, I think it's important to remember that your model is an abstraction based on a plethora of underlying assumptions. When comparing the output of your model for one class to the output of your model for another, that's fine: both outputs are based on those same assumptions, so they're directly comparable despite the abstraction. But the moment you isolate a single class and try to make use of your model's raw numerical output in an attempt to analyze something else (such as GWM breakpoints in the above quote) you're misapplying your model, effectively forgetting that (like all models) your numerical results hold only in the context of your underlying assumptions.

Put another way, your model doesn't output a class's numerical DPR. It instead outputs a dimensionless numerical weighting correlated to DPR. That weighting can be effectively compared only to other dimensionless numerical weightings produced by the same procedure, but cannot be usefully compared to raw numerical DPR or weightings produced via other methods.

So, in the context of calculating a GWM breakpoint as above, what you've calculated is that the breakpoint is AC 22 for a character that chooses whether or not to use GWM with the goal of maximizing the weighting produced by your model for his build. But that doesn't mean that the breakpoint for GWM for Fighter 20 is AC 22.

[Kryx]

Don't forget that when critting one can use the highest available spell slot to maximize the gain. That will skew the average higher.
An average should account for highs and lows. Sometimes it will be a high spell slot, sometimes low. There is no way to predict when that'll happen. Maybe the crit won't happen until you use your high level slots. Maybe you save your high level slots just for crits and never use them.

If you can think of how to model this then I'll implement it, but I cannot think of how to do so. I also think this would change the DPR by a very little amount.


It would also be perfectly reasonable to measure DPR as the average of the per-round damage, instead assuming that characters are trying to maximize the damage they do on a particular round.
It would, but that is an entirely different calculation. 5e's balance criteria is based around a whole day. That's the avenue that I'm interested in pursuing.


I agree that strategic use of the -5/+10 is likely too complicated to effectively model....it still leaves your model ineffective at capturing the highly-relevant interplay between Divine Smite and GWM in the context of a discussion on the mathematical import of Divine Smite
I currently use average AC as the target for attacks. That average AC accounts for some ACs that are higher and some that are lower. 22+ may suck, but there are far more cases of 16 or below that have MUCH higher damage. The point of using an average is that it evens out.
I do not agree that it inadequately models divine smite with -5/+10 in the context of a day. A round by round dissection that threw different creatures of different ACs at it would likely produce a different, though I think very similar result.


I think your work making DPR charts to compare the classes is fantastic. It was a ton of work and produced a very thought-provoking analysis.
Thanks!


The moment you isolate a single class and try to make use of your model's raw numerical output in an attempt to analyze something else (such as GWM breakpoints in the above quote) you're misapplying your model, effectively forgetting that (like all models) your numerical results hold only in the context of your underlying assumptions.
I don't see the issue here. How does using average AC or average damage not apply to a single class or specific feature? The whole concept of the model is to be able to mathematically say how valuable classes or features are.


your model doesn't output a class's numerical DPR. It instead outputs a dimensionless numerical weighting correlated to DPR.
It outputs expected DPR over a set day. The assumptions I make are meant to be averages over that day. That does not prevent it from being DPR as all D&D DPR is built on assumptions (See DPR threads)


That weighting can be effectively compared only to other dimensionless numerical weightings produced by the same procedure, but cannot be usefully compared to raw numerical DPR or weightings produced via other methods
It can only be used to compare with other classes/features from my same model or anyone who makes the same assumptions I do, correct. It cannot be used to compare to models that do not use the full math..


So, in the context of calculating a GWM breakpoint as above, what you've calculated is that the breakpoint is AC 22 for a character that chooses whether or not to use GWM with the goal of maximizing the weighting produced by your model for his build. But that doesn't mean that the breakpoint for GWM for Fighter 20 is AC 22.
I really must be missing the small details here... A fighter who uses GWM and follows the build I laid out has a breakpoint of 22 AC... The build I laid out is typical for a fighter. Why is it wrong to use that breakpoint?..

[KorvinStarmast]

5d8 = 22.5
5d8 reroll 1s and 2s = 26.25

+3.75 damage Not game breaking, and thanks for the damage breakdown.

[Xetheral]

Don't forget that when critting one can use the highest available spell slot to maximize the gain. That will skew the average higher.An average should account for highs and lows. Sometimes it will be a high spell slot, sometimes low. There is no way to predict when that'll happen. Maybe the crit won't happen until you use your high level slots. Maybe you save your high level slots just for crits and never use them.

If you can think of how to model this then I'll implement it, but I cannot think of how to do so. I also think this would change the DPR by a very little amount.

Since I'm arguing that your model is less useful in this context than in its class-vs-class context, I don't know that it would be worth the effort to account for this factor too. If you wanted to, you could introduce a new scaling factor to up the contribution from smite, but I suspect we'd never agree on what an appropriate factor was. :)



It would also be perfectly reasonable to measure DPR as the average of the per-round damage, instead assuming that characters are trying to maximize the damage they do on a particular round.
It would, but that is an entirely different calculation. 5e's balance criteria is based around a whole day. That's the avenue that I'm interested in pursuing.

I would think that most players would be trying to optimize their round-by-round damage, and not worry about optimizing their daily totals. We might just have different experiences though.



I agree that strategic use of the -5/+10 is likely too complicated to effectively model. But while that may make the decision to not include that factor a very reasonable modeling decision, it still leaves your model ineffective at capturing the highly-relevant interplay between Divine Smite and GWM in the context of a discussion on the mathematical import of Divine Smite. I currently use average AC as the target for attacks. That average AC accounts for some ACs that are higher and some that are lower. 22+ may suck, but there are far more cases of 16 or below that have MUCH higher damage. The point of using an average is that it evens out.
I do not agree that it inadequately models divine smite with -5/+10 in the context of a day. A round by round dissection that threw different creatures of different ACs at it would likely produce a different, though I think very similar result.

I could be mistaken, but it makes sense to me that the result would be entirely different. Against any opponent worthy of using Divine Smite on, it will almost never be worthwhile to simultaneously take the -5/+10 from GWM, because the proportional decrease in smite damage represented by the hit penalty will exceed the proportional increase in damage from the damage bonus. Accordingly, in play, Divine Smite and GWM will rarely be used in the same round by a player trying to maximize their damage. Your model assumes that both are used simultaneously which both overstates the impact of GWM (because it won't be used against smite-worthy opponents) and understates the impact of Smite (because overall hit chance is reduced). It would seem to me that overstating one side of the comparison while understating the other is likely to lead to a very different conclusion on the relative impact of each feature.



But the moment you isolate a single class and try to make use of your model's raw numerical output in an attempt to analyze something else (such as GWM breakpoints in the above quote) you're misapplying your model, effectively forgetting that (like all models) your numerical results hold only in the context of your underlying assumptions.
I don't see the issue here. How does using average AC or average damage not apply to a single class or specific feature? The whole concept of the model is to be able to mathematically say how valuable classes or features are.

I fear that won't quite work. As an analogy, consider credit scores. They're an abstraction that correlates strongly with risk of default, just as your model outputs a value that correlates strongly with DPR over the course of an adventuring day. One can evaluate what changes to one's credit history will do to one's credit score, but one can't use credit scores to try to measure what changes to one's credit history will do to one's risk of default. Similarly, while one can evaluate what changes to a class feature will do to the output of your model, you can't then use the output of your model to try to measure the impact of those changes on DPR.

For a more specific example, let's delve into the weeds of your model for a moment. When calculating the impact of Divine Smite you're straight out dividing the available smite damage by your assumption that there are 25 rounds of combat in a given adventuring day. If that value were changed to be 20, the impact of divine smite on your charts in comparison to the impact of GWM would instantly be increased by 25%. When comparing Fighter DPR from your model to Paladin DPR from your model, this isn't that big a deal (although it still matters). But when the discussion revolves around the relative DPR impact of GWM and Divine Smite for a Paladin, trying to draw conclusions from the output of your model is severely constrained by the fact that the value of that output depends so strongly on the values chosen for the assumptions.

Does that make more sense?



your model doesn't output a class's numerical DPR. It instead outputs a dimensionless numerical weighting correlated to DPRIt outputs expected DPR over a set day. The assumptions I make are meant to be averages over that day. That does not prevent it from being DPR as all D&D DPR is built on assumptions (See DPR threads)

In a very technical sense, you're correct that your model outputs DPR, and therefore isn't dimensionless. (*Probably. It's possible you've mistakenly divided out a unit you didn't intend to.) However, the assumptions and simplifications underlying your model create an unknown (non-linear!) scaling factor that renders the magnitude of the output of your model un-interpretable. (If you really wanted you could incorporate error analysis to your model to account for the scaling factor, but you really, really don't want to do that much work, and I can virtually guarantee that the uncertainty you'd calculate would overwhelm the output.) Your "unit" of DPR, so to speak, is not equivalent to the raw unit of DPR that is applied to basic calculations such as average damage*average hit chance*number of attacks. Because of this unknown scaling factor, I would argue that it's easiest to think of the output of your model as dimensionless, rather than labeling the output as straight DPR when doing so would be misleading.



So, in the context of calculating a GWM breakpoint as above, what you've calculated is that the breakpoint is AC 22 for a character that chooses whether or not to use GWM with the goal of maximizing the weighting produced by your model for his build. But that doesn't mean that the breakpoint for GWM for Fighter 20 is AC 22.
I really must be missing the small details here... A fighter who uses GWM and follows the build I laid out has a breakpoint of 22 AC... The build I laid out is typical for a fighter. Why is it wrong to use that breakpoint?

Let me illustrate. From your spreadsheet, a Fighter 20 deals 8.33 damage per hit and has a +11 to hit. The breakpoint, X, occurs at the AC where the proportional decrease in chance to hit exactly compensates for the proportional increase in damage, that is, where (Old Hit Chance)/(New Hit Chance) = (New Damage/Old Damage). (New Damage/Old Damage) in this case is 19.33/8.33 = 2.3205.

Old Hit Chance = (21-(X-11))*.05
New Hit Chance = (21-(X-11+5))*.05

The ratio simplifies to (X-32)/(X-27), which equals 19.33/8.33 when X = 23.2126. Your model therefore understated the actual breakpoint by 2 (your model outputs a breakpoint of 21, which I miswrote as 22).

It's certainly close, as a good model will be, but the point remains that using your model to calculate breakpoints gives inaccurate results in the general case.

[Ruslan]

You probably should have had 18.33/8.33. Where did 19.33 come from?

[Asaris]

It might be worth noting that -5/+10 is better for the Oath of Vengeance paladin, because they can get advantage at any time.

[JoeJ]

It might be worth noting that -5/+10 is better for the Oath of Vengeance paladin, because they can get advantage at any time.

Although only against one enemy per rest, which makes it hard to model because we don't know how many of the paladin's attacks are made against that one enemy.

[Kryx]

I would think that most players would be trying to optimize their round-by-round damage, and not worry about optimizing their daily totals.
Indeed players would attempt to do so, but that's not how the game is balanced. The game is balanced around 6-8 encounters and 2 short rests per day. Obviously some of those encounters are more important than others and burst classes will do better at them. I think a system that tried to model that would be great, but my model is intended to look at the overall balance of classes, archetypes, and features given 5e's system.


I could be mistaken, but it makes sense to me that the result would be entirely different. Against any opponent worthy of using Divine Smite on, it will almost never be worthwhile to simultaneously take the -5/+10 from GWM, because the proportional decrease in smite damage represented by the hit penalty will exceed the proportional increase in damage from the damage bonus.
That may be what you think, but according to the math that is not true. Duplicate the sheet and see for yourself. Take away the -5/+10 for a level 20 Paladin GWM for example and his DPR drops from 35.7 to 34.1.


Your model assumes that both are used simultaneously
My model assumes that because the math says that the DPR is higher on average when they are both used.


Does that make more sense?
I just don't agree with your conclusion that since DPR isn't 100% accurate to real situations that it can't be used to judge balance. I disagree that DPR is a weighted comparison and not just DPR. DPR is based on averages in every D&D DPR model ever used - I follow the same system used for years in DPR threads of all editions.


For a more specific example, let's delve into the weeds of your model for a moment. When calculating the impact of Divine Smite you're straight out dividing the available smite damage by your assumption that there are 25 rounds of combat in a given adventuring day.
Yes! Specifics that we can actually discuss!
Divine Smite does indeed assume 25 rounds per day - I assume 5 encounters of 5 rounds. That is well within the 6-8 encounters expected by 5e. 25 rounds is likely an average. Some days are 40 rounds, some days are 10 rounds.


But when the discussion revolves around the relative DPR impact of GWM and Divine Smite for a Paladin, trying to draw conclusions from the output of your model is severely constrained by the fact that the value of that output depends so strongly on the values chosen for the assumptions.
Again I disagree with this. The point of a model is to compare based on averages. If that assumption of 25 average rounds per adventuring day is incorrect then that should be adjusted. Though I think you'd be hard pressed to find a better average number.


I think the difference in your perspective is the average - you don't seem to value the average or any numbers derived from it.


the assumptions and simplifications underlying your model create an unknown (non-linear!) scaling factor that renders the magnitude of the output of your model un-interpretable.
Are you saying my assumptions are wrong? If so please help me fix them

If you're saying that any assumptions ruins the model then I do not agree. Everyone uses DPR based on assumptions to compare features. "straight DPR" which you praise below makes significantly more assumptions than I do.




Your "unit" of DPR, so to speak, is not equivalent to the raw unit of DPR that is applied to basic calculations such as average damage*average hit chance*number of attacks.
Yes, we discussed this above. Of course the most simplistic DPR possible is not comparable to more complex math. Fighter without taking into account Action Surge or Superiority dice looks awful. But once those are taken into account it actually excels.


I would argue that it's easiest to think of the output of your model as dimensionless, rather than labeling the output as straight DPR when doing so would be misleading.
And I would argue that "straight DPR" is entirely garbage for comparisons. See what I just wrote about the fighter. I started with "straight DPR", but it leaves out so much and heavily skews the numbers to favor classes without resources - hence the 25 rounds per day.






Let me illustrate. From your spreadsheet, a Fighter 20 deals 8.33 damage per hit and has a +11 to hit. The breakpoint, X, occurs at the AC where the proportional decrease in chance to hit exactly compensates for the proportional increase in damage, that is, where (Old Hit Chance)/(New Hit Chance) = (New Damage/Old Damage). (New Damage/Old Damage) in this case is 19.33/8.33 = 2.3205.

Old Hit Chance = (21-(X-11))*.05
New Hit Chance = (21-(X-11+5))*.05

The ratio simplifies to (X-32)/(X-27), which equals 19.33/8.33 when X = 23.2126. Your model therefore understated the actual breakpoint by 2 (your model outputs a breakpoint of 21, which I miswrote as 22).

It's certainly close, as a good model will be, but the point remains that using your model to calculate breakpoints gives inaccurate results in the general case.
You've ignored so much... what about:

Crit is a defined number - there are no assumptions made
Action Surge happens a set amount of times - easy to calculate
Superiority Dice happens a set amount of times - easy to calculate
GWM bonus attack happens when you crit (easy to calculate) or kill (fairly accurate guesstimate based on DMG table of average HP)
Opportunity Attack is the only thing here heavily DM dependent - of which I've assumed a 15% chance


Your 23.2126 being a breakpoint is far less accurate than mine is as it ignores all of those factors. The only thing you've taken into consideration is the weapon dice.. I can't comprehend how you can argue that this method is more accurate... You've literally ignored a majority of the class and features.
Superiority Dice massively changes the DPR as the ability to trip an enemy gives you advantage which significantly changes the math of -5/+10. So does action surging and attacking 4 times after you've tripped said enemy.

I'll throw some numbers through my model to get a more accurate estimate:
AC 19 Enemy. DPR with -5/+10 is 72. Without is 64. -5/+10 wins.
AC 20 Enemy. DPR with -5/+10 is 64. Without is 60. -5/+10 wins.
AC 21 Enemy. DPR with -5/+10 is 55.6. Without is 55.6. Tie

[JoeJ]

Kryx: Is there a way to incorporate effective damage rather than rolled damage in your model? By effective damage I mean damage that actually goes to reduce an enemy's hit points and is not "wasted" on overkill. Apart from the general observation that damage delivered in smaller packages is less likely to be wasted, which implies that the number of attacks per day is a factor, I'm not sure how that could be taken into account.

[Kryx]

Kryx: Is there a way to incorporate effective damage rather than rolled damage in your model? By effective damage I mean damage that actually goes to reduce an enemy's hit points and is not "wasted" on overkill. Apart from the general observation that damage delivered in smaller packages is less likely to be wasted, which implies that the number of attacks per day is a factor, I'm not sure how that could be taken into account.
I wish. That is actually a really good point that I have forgotten about for a while.

Encounters have enemies that range in CR +/-3. You may have some that have high HP, some with low. We could abstract that and assume average CR. Solvable.
Enemies have a set HP based on CR. # of attacks to get to this value with any leftover being wasted. Solvable, but also very complex
Enemy's current HP - unsolvable. For GWM I assume they are at 80% of their total HP on average. I wonder if I could use that number somehow...

[Xetheral]

You probably should have had 18.33/8.33. Where did 19.33 come from?

It is a mistake. When corrected, the breakpoint becomes 22.835, so Kryx's estimate is only off by 1, rather than by 2.


That may be what you think, but according to the math that is not true. Duplicate the sheet and see for yourself. Take away the -5/+10 for a level 20 Paladin GWM for example and his DPR drops from 35.7 to 34.1.

My model assumes that because the math says that the DPR is higher on average when they are both used.


When the accuracy/relevance of your spreadsheet's conclusions is called into doubt, we can't exactly use your spreadsheet to test it's own conclusions. :)


I just don't agree with your conclusion that since DPR isn't 100% accurate to real situations that it can't be used to judge balance. I disagree that DPR is a weighted comparison and not just DPR. DPR is based on averages in every D&D DPR model ever used - I follow the same system used for years in DPR threads of all editions.

Sure, you can use it as a rough gauge for balance evaluations so long as you keep the underlying assumptions in mind. But you can't use it for measurements. You can't look at a proposed balance chance and use your model to claim that the change increases the DPR of a specific class by a specific number.


Yes! Specifics that we can actually discuss!
Divine Smite does indeed assume 25 rounds per day - I assume 5 encounters of 5 rounds. That is well within the 6-8 encounters expected by 5e. 25 rounds is likely an average. Some days are 40 rounds, some days are 10 rounds.

Again I disagree with this. The point of a model is to compare based on averages. If that assumption of 25 average rounds per adventuring day is incorrect then that should be adjusted. Though I think you'd be hard pressed to find a better average number.

...

Are you saying my assumptions are wrong? If so please help me fix them

Your assumption may be the best assumption, but choosing your assumptions well doesn't mean that your model is more generally applicable than those assumptions permit it to be.

And I disagree with you on the point of a model. A model is an abstraction of a complicated system used to permit comparisons that would not otherwise be possible. Taking averages of variables rather than treating them as a distribution is a very common modeling technique, but it comes at a cost.


I think the difference in your perspective is the average - you don't seem to value the average or any numbers derived from it.

I value averages a ton. Your valuable model wouldn't be possible without them. But when modeling it's important to remember than when you take an average you're losing a ton of generality, and that once you do that it's improper to then use your model to try to make detailed measurements of the general case.


If you're saying that any assumptions ruins the model then I do not agree. Everyone uses DPR based on assumptions to compare features. "straight DPR" which you praise below makes significantly more assumptions than I do.

Straight DPR isn't necessarily more valuable, but it is far more generally accessible. Therefore, when using a model that outputs something other than straight DPR, as yours does, it is important to qualify your results so that people realize that 0.7 DPR in your model does not equal 0.7 DPR in the general sense of the term.


Yes, we discussed this above. Of course the most simplistic DPR possible is not comparable to more complex math. Fighter without taking into account Action Surge or Superiority dice looks awful. But once those are taken into account it actually excels.

And I would argue that "straight DPR" is entirely garbage for comparisons. See what I just wrote about the fighter. I started with "straight DPR", but it leaves out so much and heavily skews the numbers to favor classes without resources - hence the 25 rounds per day.

I'm not suggesting that straight DPR is useful for comparing one class to another; far from it. But it is useful when answering general questions such as "which character will do more damage this round"? (Note that I'm using "straight DPR" to refer to any simplistic DPR calculation limited to a specific question, not a specific type of DPR calculation).


You've ignored so much... what about:

Crit is a defined number - there are no assumptions made
Action Surge happens a set amount of times - easy to calculate
Superiority Dice happens a set amount of times - easy to calculate
GWM bonus attack happens when you crit (easy to calculate) or kill (fairly accurate guesstimate based on DMG table of average HP)
Opportunity Attack is the only thing here heavily DM dependent - of which I've assumed a 15% chance


Your 23.2126 being a breakpoint is far less accurate than mine is as it ignores all of those factors. The only thing you've taken into consideration is the weapon dice.. I can't comprehend how you can argue that this method is more accurate... You've literally ignored a majority of the class and features.
Superiority Dice massively changes the DPR as the ability to trip an enemy gives you advantage which significantly changes the math of -5/+10. So does action surging and attacking 4 times after you've tripped said enemy.

I'll throw some numbers through my model to get a more accurate estimate:
AC 19 Enemy. DPR with -5/+10 is 72. Without is 64. -5/+10 wins.
AC 20 Enemy. DPR with -5/+10 is 64. Without is 60. -5/+10 wins.
AC 21 Enemy. DPR with -5/+10 is 55.6. Without is 55.6. Tie

None of the factors on your list are relevant to answering the question "Should I use GWM on this attack?" Answering that question is the point of knowing the breakpoint, because the decision on whether or not to use the feat is made on each specific attack.

This is a prime example of where your model is very useful and where it is less useful. If a character wants to know what GWM will likely do for his DPR over his career, your model is fantastic. It can't provide meaningful numerical results (because of the unknown scaling factor separating your "unit" of DPR from the normal usage), but the comparison between two builds, one with GWM and one without, will be far easier using your charts. But, on a specific round, when trying to determine whether or not to use GWM against a particular opponent, your model is likely to give inaccurate (or coincidentally accurate) results.


Ultimately, your model is extremely useful. :) I love models and yours looks like a good one. But will all models it's important to remember the limitations of the techniques employed.

[georgie_leech]

I value averages a ton. Your valuable model wouldn't be possible without them. But when modeling it's important to remember than when you take an average you're losing a ton of generality, and that once you do that it's improper to then use your model to try to make detailed measurements of the general case.


As a quick example of why this is so important, imagine a class that has a feature that let's it once per round with a 0.001% chance deal 100000000000000000 damage, but otherwise deals only 1 damage. It's average damage per round is 1000000000000.999, but in the vast majority of specific cases it's next to useless. Such a class should never exist, but it's just an illustration of why average damage isn't necessarily an indicator of useful damage.

[Kryx]

Sure, you can use it as a rough gauge for balance evaluations so long as you keep the underlying assumptions in mind. But you can't use it for measurements. You can't look at a proposed balance chance and use your model to claim that the change increases the DPR of a specific class by a specific number.
...
Your assumption may be the best assumption

Again, we just disagree. You're happy to accept "straight DPR" even though it is missing so many factors. This logic makes no sense to me.


When the accuracy/relevance of your spreadsheet's conclusions is called into doubt, we can't exactly use your spreadsheet to test it's own conclusions. :)
For your opinion of GWM not being worth it "because smite" you have provided no data. I have provided the best data that I know.


when modeling it's important to remember than when you take an average you're losing a ton of generality, and that once you do that it's improper to then use your model to try to make detailed measurements of the general case.
...
It can't provide meaningful numerical results (because of the unknown scaling factor separating your "unit" of DPR from the normal usage)
We just simply disagree on the application of models. The whole point of a model is to compare, whereas you think a model prevents it from being able to compare in the first place. Or maybe you only think a round-by-round model allows comparison.
DPR is a tool to measure differences and should be used to do so. If you have a system to generate more accurate DPR then please do share it - I'd happily switch and contribute. Otherwise I'll use the most accurate DPR system that I know of to compare and contrast builds.



Straight DPR isn't necessarily more valuable, but it is far more generally accessible. Therefore, when using a model that outputs something other than straight DPR, as yours does, it is important to qualify your results so that people realize that 0.7 DPR in your model does not equal 0.7 DPR in the general sense of the term.
DPR doesn't have a "general sense". For this forum's sake the main "straight DPR" used ignores half the rules. Because of this it is strictly inferior. Using the 20 GWM Fighter once again "straight DPR" ignores half of his class features which in my model changes his DPR from 43.8 to 71.9. In this case "straight DPR" is incredibly inaccurate.
The same is true for many situations - Barbarian, all Fighters, Paladin, etc.



it is useful when answering general questions such as "which character will do more damage this round"? (Note that I'm using "straight DPR" to refer to any simplistic DPR calculation limited to a specific question, not a specific type of DPR calculation).
"Straight DPR" isn't good at calculating "which character will do more damage this round?" either. In every case on these forums all the factors I mentioned above plus more are ignored. A list of some that are often ignored: Crit, Opportunity Attack, likelihood of hitting, significance of advantage, class features[superiority dice, action surge, monk stun, paladin smite], feats, etc



None of the factors on your list are relevant to answering the question "Should I use GWM on this attack?"
What? Let me quote your words above:

The breakpoint, X, occurs at the AC where the proportional decrease in chance to hit exactly compensates for the proportional increase in damage, that is, where (Old Hit Chance)/(New Hit Chance) = (New Damage/Old Damage)
Your words state how important damage is. You then go on to ignore a significant amount (~40%) of the damage while making this calculation. You can't make an accurate calculation while ignoring that much.


But, on a specific round, when trying to determine whether or not to use GWM against a particular opponent, your model is likely to give inaccurate (or coincidentally accurate) results.
You keep coming back to round-by-round. My numbers are based on 5e's system which uses an adventuring day to balance classes and features. My numbers are entirely unconcerned with round-by-round.



Ultimately, your model is extremely useful. :) I love models and yours looks like a good one. But will all models it's important to remember the limitations of the techniques employed.
Thanks again for the kind words. We just disagree on the limitations and the possible applications of DPR.

Though if we go by your logic then we can never compare anything as we can never get the full picture. Simple math and my DPR math both show that the GWF fighting style is inferior for example, but you're saying we can't use those numbers to determine that. So what numbers can we use to determine that?

[Kryx]

As a quick example of why this is so important, imagine a class that has a feature that let's it once per round with a 0.001% chance deal 100000000000000000 damage, but otherwise deals only 1 damage. It's average damage per round is 1000000000000.999, but in the vast majority of specific cases it's next to useless. Such a class should never exist, but it's just an illustration of why average damage isn't necessarily an indicator of useful damage.
Averages do have a weakness in extreme outliers. Luckily those don't exist in 5e.

Though lets run with your claim that averages invalidate any data based on them. What then shall we use to calculate differences in builds? What system would be better?

[georgie_leech]

Averages do have a weakness in extreme outliers. Luckily those don't exist in 5e.

Though lets run with your claim that averages invalidate any data based on them. What then shall we use to calculate differences in builds? What system would be better?

I didn't claim that averages were useless, just that averages can skew the data in a way that you're apparently disregarding.

[Strill]

It would, but that is an entirely different calculation. 5e's balance criteria is based around a whole day. That's the avenue that I'm interested in pursuing.That's not quite true. 5e's balance criteria is also predicated around the idea that one high-health creature is equal to many medium-health creatures, and that single-target burst damage is equal to area-of-effect damage, even if the AoE does far more damage overall.

Just compare the damage of single-target spells vs AoE spells, or the statistics of minions created from the Animate Objects spell. Animating tiny objects with Animate Objects results in far more overall damage and health than animating a Huge object, but the game considers both equivalent.

[Xetheral]

For your opinion of GWM not being worth it "because smite" you have provided no data. I have provided the best data that I know.

The data is in the link in my signature. For example, if a Paladin does 8.33 base damage and plans to spend a 1st level spell slot on a hit, her base damage is 17.33. Unless she can hit the opponent on a roll of 7 or better, she shouldn't use GWM if she wants to maximize her damage that round. Few opponents at any level can be hit on a 7.


We just simply disagree on the application of models. The whole point of a model is to compare, whereas you think a model prevents it from being able to compare in the first place. Or maybe you only think a round-by-round model allows comparison.
DPR is a tool to measure differences and should be used to do so. If you have a system to generate more accurate DPR then please do share it - I'd happily switch and contribute. Otherwise I'll use the most accurate DPR system that I know of to compare and contrast builds.

You model can helps whenever the results it is being compared to were also filtered through your model. You can thus help answer the question "Which class does more damage, overall?" What it can't do is help answer the question "How much extra damage per round will I do if I take GWM?"


DPR doesn't have a "general sense". For this forum's sake the main "straight DPR" used ignores half the rules. Because of this it is strictly inferior. Using the 20 GWM Fighter once again "straight DPR" ignores half of his class features which in my model changes his DPR from 43.8 to 71.9. In this case "straight DPR" is incredibly inaccurate.
The same is true for many situations - Barbarian, all Fighters, Paladin, etc.

"Straight DPR" isn't good at calculating "which character will do more damage this round?" either. In every case on these forums all the factors I mentioned above plus more are ignored. A list of some that are often ignored: Crit, Opportunity Attack, likelihood of hitting, significance of advantage, class features[superiority dice, action surge, monk stun, paladin smite], feats, etc

Straight DPR is fantastically useful for making limited comparisons. For example, if you want to know how much more damage you'll do with a longsword rather than a dagger, all else equal, a proportional comparison using straight DPR works great: 1.89 times as much.


What? Let me quote your words above:

Your words state how important damage is. You then go on to ignore a significant amount (~40%) of the damage while making this calculation. You can't make an accurate calculation while ignoring that much.

You keep coming back to round-by-round. My numbers are based on 5e's system which uses an adventuring day to balance classes and features. My numbers are entirely unconcerned with round-by-round.

The choice of whether or not to use GWM is made round-by-round (or, more precisely, attack-by-attack). Whether or not you choose to use GWM on a specific attack against a specific opponent is unaffected by critical hits (chance to crit is the same each way, and the +10 damage isn't multiplied anyway), action surge (the ability to take more actions does not affect the damage you do on this attack), GWM bonus action attack (except insofar as the ratio between your potential damage and the target's remaining HP), or opportunity attacks (the ability to take future attacks does not affect the damage you do on this attack). Superiority dice do affect the decision, as the option to use them affects the hit chance (for precise strike) and damage totals (for the others), so I was hasty in saying that none of them matter, but the point remains that one can more accurately calculate the actual breakpoint using a traditional calculation than one can using your model.


Though if we go by your logic then we can never compare anything as we can never get the full picture. Simple math and my DPR math both show that the GWF fighting style is inferior for example, but you're saying we can't use those numbers to determine that. So what numbers can we use to determine that?

We can make comparisons all we want to long as we take into account the assumptions of the model. Which numbers are of use to answer a particular question at a particular level of generality will vary greatly. There is no one model that is the correct choice for all potential comparisons. (If there was a such model, it would be D&D itself, exactly, which would defeat the point of having a model in the first place.)

[Kryx]

averages can skew the data in a way that you're apparently disregarding.
Averages can skew the data if there are extreme outliers. In 5e there are no extreme outliers. I could run a simulation that did the same thing over and over and it'd end up being the same as the averages I provide - that's the point of averages.

I'm not trying to "win" an argument here, but I seriously want to know: If averages skew the data then what shall we use to calculate differences in builds? What system would be better?

If there is no better system then I'll use this model to compare as I have confidence that it is very accurate to real play. Any differences from actual play would likely be different for both sides of the comparison in most cases as they both use the same model (some exceptions obviously, but those can be adjusted).



That's not quite true. 5e's balance criteria is also predicated around the idea that one high-health creature is equal to many medium-health creatures, and that single-target burst damage is equal to area-of-effect damage, even if the AoE does far more damage overall.

Just compare the damage of single-target spells vs AoE spells, or the statistics of minions created from the Animate Objects spell. Animating tiny objects with Animate Objects results in far more overall damage and health than animating a Huge object, but the game considers both equivalent.
I'm not seeing how this is related to what you quoted. I was discussing balance based around an average adventuring day. These topics seem unrelated balance questions of specific things like single target vs AoE or multiple medium hp vs higher HP.

If you're saying single target vs AoE cannot be modeled then I think it can. I did so for my Wizard vs Sorcerer comparison (https://docs.google.com/spreadsheets/d/1H1SxsntIaK1IIHwJmR_6aUnUtZE6yWlP4vPzHQLE7Ak/edit) (old numbers). In it I assume that AoEs hit a certain number of people when they are cast.



The data is in the link in my signature. For example, if a Paladin does 8.33 base damage and plans to spend a 1st level spell slot on a hit, her base damage is 17.33. Unless she can hit the opponent on a roll of 7 or better, she shouldn't use GWM if she wants to maximize her damage that round. Few opponents at any level can be hit on a 7.
You're switching to the Paladin which has different concerns than the Fighter. You're also focusing on one single attack instead of all his attacks in a round. There is a breakpoint, but looking at individual attacks to see whether it is valuable to use -5/+10 on that attack is different than looking at the whole round overall. Maybe a Paladin should not use -5/+10 when he tries to smite, but that would just be a fix to the model that the Paladin uses, not invalidating it.


You model can helps whenever the results it is being compared to were also filtered through your model. You can thus help answer the question "Which class does more damage, overall?" What it can't do is help answer the question "How much extra damage per round will I do if I take GWM?"
It can answer that question for a round based on an enemy's average AC. My model does not calculate each individual attack.
Again this boils down to averages - you do not seem to value averages. In some cases the AC will be higher. In some it will be lower. In the end it all averages out and we can determine if GWM should be used or not.


Straight DPR is fantastically useful for making limited comparisons. For example, if you want to know how much more damage you'll do with a longsword rather than a dagger, all else equal, a proportional comparison using straight DPR works great: 1.89 times as much.
1.89 times as much weapon damage. That ignores all the other factors which in practice far outweigh the value of the weapon damage. In the end a dagger vs longsword would be no greater than 20% of a paladin's DPR. Therefore 1.89x as much is misinformation, at best.


The choice of whether or not to use GWM is made round-by-round (or, more precisely, attack-by-attack). Whether or not you choose to use GWM on a specific attack against a specific opponent is unaffected by critical hits (chance to crit is the same each way, and the +10 damage isn't multiplied anyway), action surge (the ability to take more actions does not affect the damage you do on this attack), GWM bonus action attack (except insofar as the ratio between your potential damage and the target's remaining HP), or opportunity attacks (the ability to take future attacks does not affect the damage you do on this attack). Superiority dice do affect the decision, as the option to use them affects the hit chance (for precise strike) and damage totals (for the others), so I was hasty in saying that none of them matter, but the point remains that one can more accurately calculate the actual breakpoint using a traditional calculation than one can using your model.
Trip attack for Superiority Dice changes the math significantly. "straight" DPR isn't going to model that without adding complexity - which my model does.

Critting, Opportunity Attack, Action Surge, and Bonus attacks all effect the overall round's DPR if you use -5/+10 or not. What you're looking to calculate is each individual attack which is different. As I said above maybe the attacks where he smites should be made without -5/+10, but that is an adjustment to the model to make it more accurate.
If you include the number of times you can trip, the chance to trip an enemy which then gives you a higher chance to hit then you're in the same territory as I am.
Without including that the numbers are not valuable as they ignore a significantly portion of the factors.
So with resource based features you're forced to ignore them or quantify how often they're used. Ignoring them is highly inaccurate. Estimating their usage based on an average day is far more accurate than ignoring them. An adventuring day is what 5e uses to balance resources, so DPR should use it too.

[Strill]

I'm not seeing how this is related to what you quoted. I was discussing balance based around an average adventuring day. These topics seem unrelated balance questions of specific things like single target vs AoE or multiple medium hp vs higher HP.

If you're saying single target vs AoE cannot be modeled then I think it can. I did so for my Wizard vs Sorcerer comparison (https://docs.google.com/spreadsheets/d/1H1SxsntIaK1IIHwJmR_6aUnUtZE6yWlP4vPzHQLE7Ak/edit) (old numbers). In it I assume that AoEs hit a certain number of people when they are cast.

If single-target attacks can do less overall damage than AoE, but still be valued as equal, then your optimization of overall damage during the adventuring day does not accurately model the value of burst damage. Spending several class features in the same turn in order to maximize damage done in a single turn, at the expense of overall damage, is a viable strategy which you're not accounting for.

Just as a single-target spell is worth the same as an AoE, even though it deals less overall damage, burst-damage on a martial character is worth sacrificing overall damage for. This is why it can be optimal to use precision attack over trip attack, or even to use precision attack multiple times in a single turn.

In a more concrete context, you're saying that it's better to be sparing with superiority dice against the boss, so that you can use the superiority dice to clean up his weaker minions after all the fight's major threats are gone. That's not a reasonable conclusion. Damage later in the fight after most of the threats are gone is worth less than damage at the beginning of the fight.

[Ruslan]

For example, if a Paladin does 8.33 base damage and plans to spend a 1st level spell slot on a hit, her base damage is 17.33. Unless she can hit the opponent on a roll of 7 or better, she shouldn't use GWM if she wants to maximize her damage that round.
There's a subtle logical flaw here. If the Paladin misses with the attack, what does she lose? She loses the chance to deal her base 8.33 damage. She does not lose the chance to deal the +2d8 Smite damage. Well, on this attack, she does, but not in general. Assuming this particular combat round is not the last one in the Paladin's lifetime, the +2d8 sits safely in the bank, so to speak, ready to be tacked on to the next attack that hits.

As an example, let's say our Paladin has 4 spell slots. And she fights 20 rounds of combat per day. She can fight normally, hitting 13 times in 20 attacks, or she can use GWM, hitting 8 times in 20 attacks. In both cases, she gets to Smite exactly 4 times. The Smite part of her damage output is unaffected.

[georgie_leech]

There's a subtle logical flaw here. If the Paladin misses with the attack, what does she lose? She loses the chance to deal her base 8.33 damage. She does not lose the chance to deal the +2d8 Smite damage. Well, on this attack, she does, but not in general. Assuming this particular combat round is not the last one in the Paladin's lifetime, the +2d8 sits safely in the bank, so to speak, ready to be tacked on to the next attack that hits.

As an example, let's say our Paladin has 4 spell slots. And she fights 20 rounds of combat per day. She can fight normally, hitting 13 times in 20 attacks, or she can use GWM, hitting 8 times in 20 attacks. In both cases, she gets to Smite exactly 4 times. The Smite part of her damage output is unaffected.

He's not talking about maximising overall damage per round though, but maximum damage that round. It's relevant for situations where you can either drop a target immediately and so eliminate their contribution to the fight, or for attacking bigger threats that need to be killed sooner rather than later.

[Ruslan]

He's not talking about maximising overall damage per round though, but maximum damage that round. It's relevant for situations where you can either drop a target immediately and so eliminate their contribution to the fight, or for attacking bigger threats that need to be killed sooner rather than later.I understand, but in that case the logic makes even less sense, because for dropping targets, there's another variable that's missing - the target's hit point value. For example, if the target has 25 hit points, there is zero chance dropping it if you attack conservatively, dealing 17.33 damage. Using GMW for +10 damage, on the other hand, gives you at least some chance to drop it.

Overall, I find the damage this round analysis worthless without adding many additional factors.

[Kryx]

This is why it can be optimal to use precision attack over trip attack, or even to use precision attack multiple times in a single turn.
Precision in comparison to Trip is never better due to -5/+10 on the second attack. That value is worth more than potentially having a slightly higher chance to hit on this attack.


Spending several class features in the same turn in order to maximize damage done in a single turn, at the expense of overall damage, is a viable strategy which you're not accounting for.
..
In a more concrete context, you're saying that it's better to be sparing with superiority dice against the boss, so that you can use the superiority dice to clean up his weaker minions after all the fight's major threats are gone. That's not a reasonable conclusion. Damage later in the fight after most of the threats are gone is worth less than damage at the beginning of the fight.
100% agreed that my model does not weight boss fights to be more important. In practice they would be. But that's not something average DPR is going to cover unless you start weighting earlier rounds of combat more. If you have any suggestions for how to weight that I'd be happy to hear them. Though many would complain about any kind of estimate like I've done with 5x5 encounters.





Overall, I find the damage this round analysis worthless without adding many additional factors.
Agreed. Even then damage this round analysis is only worthwhile for actual play and learning how to optimize your playstyle. It isn't practically useful for balancing features against each other.

[bid]

100% agreed that my model does not weight boss fights to be more important. In practice they would be. But that's not something average DPR is going to cover unless you start weighting earlier rounds of combat more. If you have any suggestions for how to weight that I'd be happy to hear them. Though many would complain about any kind of estimate like I've done with 5x5 encounters.
Any model is not a be all end all, it is only a baseline to evaluate variations. Players should think "this is a boss fight, I can't drop him if I'm lucky therefore I should aim for constant damage" and "we have the DPR to drop him in 3 rounds if I use big smites". Blaming the tool is being a tool.

[Malifice]

Not game breaking, and thanks for the damage breakdown.

Yep. 3.75 extra damage on average on 5d8?

Who cares? Give it to them, and move on.

[poobah_1]

I used to allow it. Now I do not.

My issue is that looking at the text and applying a little thought it becomes clear.

GWF allows damage rolls of 1 or 2 to be re-rolled "on a damage die for an attack you make with a melee weapon you are wielding with two hands."

Then look at Smite:
"when you hit a creature with a melee weapon Attack, you can expend one spell slot to deal radiant damage to the target, in addition to the weapon's damage.


The provision within Smite separates Weapon Damage and Smite Damage. By stating -"in addition to". The separation is specifically stated that weapon is one thing and the radiant damage is another.


Now lets think about GWF and apply some thought, some intent....The idea behind GWF is that the PC has become particularly proficient with attacking with large weapons, so that when the PC's strikes an NPC with his Greatsword, Maul, Polearm, or Glaive(etc.) the damage from that attack from that weapon is elevated. The type of damage from any physical weapon can only be slashing, piercing, or bludgeoning.

The text is specific as well, " damage die for an attack you make with a melee weapon you are wielding with two hands." ----The only types of damage that can be incurred by a melee weapon are piercing, slashing, or bludgeoning. Unless the weapon has some enchantment where additional damage is provisioned.

It does not state - "damage die for an attack you make with a melee weapon you are wielding with both hands, plus all subsequent damage die from other sources."

The increase in damage is related to the weapon it is Great Weapon Fighting. Not Great Weapon Opportunity Attacks. It is a buff related to the weapon. Generally, Weapons do slashing, piercing, and bludgeoning damage.


Within the text of GWF the intent is clearly to allow for a damage buff related to the weapon used.
Within the text of Smite, the intent is to separate the damage from the weapon and the radiant damage from the Smite.

Put the two together and in my opinion (after 30+ years of Dm'ing) all weapon specific damage (piercing, Slashing, and bludgeoning...maybe some fire, cold, or lightning if the weapon is enchanted) would re-roll 1 and 2. All additional damage buff's from spells, or smites, or effects or aura are normal rolls.

Some may disagree from a matter of balance, or comparative statistics, but for me, the rule is specific. Specific beats General

PooBah

[Arial Black]

Part 1: RAW:-


Once per turn when you roll damage for a melee weapon attack, you can reroll the weapon’s damage dice and use either total

Clearly, this would not include damage from stuff like Smite, Sneak Attack, Flame Tongue, etc. because that damage may be from the attack but not from the 'weapon's damage dice'.


When you score a critical hit, you get to roll extra dice for the attack’s damage against the target. Roll all of the attack’s damage dice twice and add them together. Then add any relevant modifiers as normal. To speed up play, you can roll all the damage dice at once. For example, if you score a critical hit with a dagger, roll 2d4 for the damage, rather than 1d4, and then add your relevant ability modifier. If the attack involves other damage dice, such as from the rogue’s Sneak Attack feature, you roll those dice twice as well

Clearly, this would include damage from stuff like Smite, Sneak Attack and Flame Tongue, because that damage may not be from the weapon's damage dice but it is from the attack's damage dice.

So, RAW, there is a clear difference between "weapon's damage dice" and "attack's damage dice".

All we need to do is check the wording of the Great Weapon Fighting style to see if it says "weapon's damage dice or "attack's damage dice, and we will have our RAW answer.

So, what does it say?


When you roll a 1 or 2 on a damage die for an attack you make with a melee weapon that you are wielding with two hands, you can reroll the die and must use the new roll, even if the new roll is a 1 or a 2. The weapon must have the two-handed or versatile property for you to gain this benefit

RAW answer: yes, all damage dice delivered by that melee weapon attack (that roll 1 or 2) are re-rolled, not just the weapon's damage dice.