Why all the hate on dual wielding?

[Mortis_Elrod]

doubling the number of attacks is a bit too much.

Only if they hit, and it's still max d6+mod . You basically take away the defensive option, better weapons. Keep the bonus action to activate, and you have 4 attacks max for someone who lands all there attacks and has xtra attack. And without the red style it's just another weak short sword stab.

Idk but someone should run those numbers and see if that works. For a 20 level fighter it would be a lot of attacks sure but they also spend an entire turn attacking trying to hit.

Maybe if you miss once you forgo the other secondary weapon hits

Ramble ramble

[Easy_Lee]

Would like to reiterate that spells like magic weapon and elemental weapon are less effective for TWF characters. Between that, crits rewarding larger damage dice, potentially needing to use two attunement slots for weapons, drawing / stowing, and competition for bonus action, a lot ought to be changed for TWF to be competitive.

I'm tempted to suggest the simplest fix: actually fighting with two weapons at once, instead of using them separately.

Base TWF:
- You may draw / stow weapons instead of one on your turn if those weapons are light.
- When you take the attack action and are wielding two light melee weapons, you may make a single attack with their combined weapon dice. If you make multiple attacks with the attack action, you may use this feature once for each attack. Your attribute bonus to the attack and damage rolls, and other static bonuses such as from magic weapons, only apply for one of the weapons, your choice.

Fighting style: You may two weapon fight with weapons that don't have the light property.

Dual Wielder:
- When you make a reaction attack and are wielding two weapons that qualify for TWF, you may make a combined attack as per TWF
- While wielding two weapons that qualify for TWF, if you take the attack action on your turn, you may spend your bonus action to make a combined weapon attack with these weapons as per TWF.

TWF would still do less damage than a great weapon barring special situations, and would still suck at critical hits. But at least this would make it more consistent with the rest of the game, bring its damage up above sword and board at least, allow it to benefit approximately equally from magic weapon and elemental weapon, and give it some interesting options (such as poison on both weapons).

[Knaight]

You claimed TWF is in its current state due to "damage overflow":

Read that quote again. I'm not claiming that TWF is in its current state at all there, I'm pointing out a failing in the DPR model as pertains to TWF in particular (it also pertains to PAM compared to GWF).

You provided no mathematical backing and still provide none. My complicated DPR/KPR and even my forum math shows that your claim has no merit as PAM does more damage with the same number of attacks. There is nothing there that can be lost in the "DPR for comparison purposes don't necessarily hold well for an actual game" claim.

You've made a claim that is not supported by even the simple math of the game.

I provided an example where DPR breaks down, and given that the entirety of my point is that DPR is an imperfect metric that can break down in the context of number of attacks that's entirely adequate.

So Knaight's point isn't wrong on the level of "things other than statistical average DPR are important to a real-gaming analysis." That's correct and he's good to have seen beyond the white-room analysis to recognize it. However, the setup of the game, as it works, based on the opponents and how they are constructed, tends to work against the advantage of the adaptability he is espousing as the benefit that redeems dual-wielding.

I'm not espousing it as a benefit that redeems dual-wielding. I don't consider dual-wielding redeemed, and would classify it as a comparatively weak option. The entirety of my point is that the DPR model is simplified; there are specific cases where that simplification will produce an inaccurate analysis; and that because of those cases there's a major distinction to be made between the claims of the theory being imperfect due to ignoring the effect of options and the theory being imperfect and therefore the same thing but with smaller numbers is just as good.

[Kryx]

I provided an example where DPR breaks down, and given that the entirety of my point is that DPR is an imperfect metric that can break down in the context of number of attacks that's entirely adequate

And yet again you fail to address PAM having the same amount of attacks as twf which totally invalidates your whole claim. We've been going in circles for several pages now with you choosing to ignore this simple fact.
You posted several pages ago that you weren't fond of the fighting style. Stick with that opinion instead of trying to make factual claims that have no evidence.



@easy: what you propose only addresses the bonus action issue. It does not address the damage issue which is still at an unacceptable 70%.
Twf doing less on crits, opportunity attacks, etc is totally a valid differentiation that doesn't need fixing imo. Bonus action needs fixing, damage needs fixing, and possibly spells giving at least magic to overcome damage resistance. Without a fix to both bonus action and damage it isn't a fix.
The issue, in my opinion, is that PAM should not offer a bonus action attack - that is the niche of twf and by doing so TWF has no niche. Removing the bonus action from PAM and -5/+10 from GWM makes those feats equivalent to an ability score improvement, instead of the current twice the value of an ability score improvement. Removing those two, often cited as OP, benefits and making other small adjustments that I outlined earlier in the thread does wonders to fix TWF by my calculations.

[Knaight]

And yet again you fail to address PAM having the same amount of attacks as twf which totally invalidates your whole claim. We've been going in circles for several pages now with you choosing to ignore this simple fact.
You posted several pages ago that you weren't fond of the fighting style. Stick with that opinion instead of trying to make factual claims that have no evidence.

PAM having the same number of attacks as TWF is totally immaterial. PAM, TWF, GWM, sword and board, and everything else are sets of inputs put into the model. I'm talking about the limits of the model itself. My claim is like saying that the ideal gas law breaks down in high pressure conditions, plus a demonstration of where it does so. In return I keep getting "both of these particular gases are stored at high pressure, which totally invalidates your whole claim", when it does nothing of the sort. We've been going in circles because you keep assigning me claims that don't particularly resemble my actual claims.

[Kryx]

Your claim is that the model did not capture "damage overflow" and implied that it was therefore not appropriate to measure the effectiveness of twf in terms of damage. I have since shown that PAM has the exact same attack structure and does more damage so the whole argument that you brought forward is invalid.

We're going in circles because you're dancing around your own claim, not taking ownership of it.
Your argument has no mathematical backing and yet you keep saying the same "DPR isn't perfect", "damage overflow" statements without any information of substance. It's like a politician making a claim and then being unable to back it up so they dance around the issue 100 times.
DPR isn't perfect, but a difference of 70% is massive - significantly bigger than the other supported martial options. Additionally as I've now said many times PAM follows the same structure and does not have the damage issue so something is clearly wrong with TWF.

[Lolzyking]

Your claim is that the model did not capture "damage overflow" and implied that it was therefore not appropriate to measure the effectiveness of twf in terms of damage. I have since shown that PAM has the exact same attack structure and does more damage so the whole argument that you brought forward is invalid.

We're going in circles because you're dancing around your own claim, not taking ownership of it.
Your argument has no mathematical backing and yet you keep saying the same "DPR isn't perfect", "damage overflow" statements without any information of substance. It's like a politician making a claim and then being unable to back it up so they dance around the issue 100 times.
DPR isn't perfect, but a difference of 70% is massive - significantly bigger than the other supported martial options. Additionally as I've now said many times PAM follows the same structure and does not have the damage issue so something is clearly wrong with TWF.

honestly wish wizards would hire you as a in house balance tester, no one else here is even on your balance level.

[Cybren]

Your claim is that the model did not capture "damage overflow" and implied that it was therefore not appropriate to measure the effectiveness of twf in terms of damage. I have since shown that PAM has the exact same attack structure and does more damage so the whole argument that you brought forward is invalid.

We're going in circles because you're dancing around your own claim, not taking ownership of it.
Your argument has no mathematical backing and yet you keep saying the same "DPR isn't perfect", "damage overflow" statements without any information of substance. It's like a politician making a claim and then being unable to back it up so they dance around the issue 100 times.
DPR isn't perfect, but a difference of 70% is massive - significantly bigger than the other supported martial options. Additionally as I've now said many times PAM follows the same structure and does not have the damage issue so something is clearly wrong with TWF.

You seem really invested in trying to argue against something knaight never said

[Knaight]

Your claim is that the model did not capture "damage overflow" and implied that it was therefore not appropriate to measure the effectiveness of twf in terms of damage. I have since shown that PAM has the exact same attack structure and does more damage so the whole argument that you brought forward is invalid.

The implication isn't that it isn't inappropriate to measure twf in terms of damage, it's that the model is imperfect and that there is a benefit to split attacks that doesn't show up in it. That means the model consistently underestimated TWF, PAM, Extra Attack, and other similar actions.

We're going in circles because you're dancing around your own claim, not taking ownership of it.
Your argument has no mathematical backing and yet you keep saying the same "DPR isn't perfect", "damage overflow" statements without any information of substance. It's like a politician making a claim and then being unable to back it up so they dance around the issue 100 times.
DPR isn't perfect, but a difference of 70% is massive - significantly bigger than the other supported martial options. Additionally as I've now said many times PAM follows the same structure and does not have the damage issue so something is clearly wrong with TWF.

That's funny. I'd liken your series of arguments to arguing against a creationist who keeps throwing out "why are there still monkeys".

[Kryx]

the model is imperfect and that there is a benefit to split attacks that doesn't show up in it. That means the model consistently underestimated TWF, PAM, Extra Attack, and other similar actions.

What... PAM is not underestimated in the model at all. PAM does about 90-95% of GWM. PAM is totally fine in that metric. TWF, however, does 70%.
The model doesn't have a problem calculating the twf structure - it does it perfectly fine for PAM. The issue is TWF.

[Cybren]

What... PAM is not underestimated in the model at all. PAM does about 90-95% of GWM. PAM is totally fine in that metric. TWF, however, does 70%.
The model doesn't have a problem calculating the twf structure - it does it perfectly fine for PAM. The issue is TWF.

Now you appear to be willfully missing the point

[Mortis_Elrod]

Now you appear to be willfully missing the point

I think he's saying that since PAM and TWF share the same structure of attacks and use, they should be the same at damage overflow control. And if that's the case then why does PAM still fall in line dmg wise, while TWF does not. The answer being TWF is just worse and there is no reason why it should be this way.

I think.

[Cybren]

I think he's saying that since PAM and TWF share the same structure of attacks and use, they should be the same at damage overflow control. And if that's the case then why does PAM still fall in line dmg wise, while TWF does not. The answer being TWF is just worse and there is no reason why it should be this way.

I think.

The actual answer is that "PAM keeping up damage wise isn't relevant to the observation that overflow damage isn't measured in DPR calculations"

[Mortis_Elrod]

The actual answer is that "PAM keeping up damage wise isn't relevant to the observation that overflow damage isn't measured in DPR calculations"

Ok so does Pam do better or worse than TWF when it comes to damage overflow?

[Cybren]

Ok so does Pam do better or worse than TWF when it comes to damage overflow?

That's not relevant. It's not about TWF or PAM or anything specifically, it's about acknowledging the limitations of your model.

[Mortis_Elrod]

That's not relevant. It's not about TWF or PAM or anything specifically, it's about acknowledging the limitations of your model.

Ok sure. Screw the model. Model is limited.

Answer the question anyway. See if we can make traction here.

[Cybren]

Ok sure. Screw the model. Model is limited.

Answer the question anyway. See if we can make traction here.

How is it relevant? I am making no claims regarding the powerlevel of any particular option.

[Mortis_Elrod]

How is it relevant? I am making no claims regarding the powerlevel of any particular option.

It's relevancy is to see if TWF lines up to similar things. You don't need to make a claim, we already established the dpr model doesn't matter, what matters is if TWF lines up In any way to everything else.

Because it seems to me like you just satisfied with pointing out flaws but not bothering with coming up with solutions, or anyway to begin with coming up with solutions. If you don't care than why say anything at all? To be right? Work with me here .

If TWF needs fixing, what if not dpr can you tell this by? If it's balanced than by what means is it in line with other similar options. If the model does not account for something. Then account for it now. This is an open question not specifically directed at you , but I ask for you to answer anyway.

[jas61292]

Ok so does Pam do better or worse than TWF when it comes to damage overflow?

Technically, it would still do worse, since its (typically) two attacks of 1d10, and one of 1d4. The net damage is practically the same as a TWF with Dual Wielder who does three attacks of 1d8. However, a more even damage spread between all three attacks makes overflow a bit less likely. But really, that's not that big a difference. Its nowhere near the the same thing as the overflow of GWF.

[Mortis_Elrod]

Technically, it would still do worse, since its (typically) two attacks of 1d10, and one of 1d4. The net damage is practically the same as a TWF with Dual Wielder who does three attacks of 1d8. However, a more even damage spread between all three attacks makes overflow a bit less likely. But really, that's not that big a difference. Its nowhere near the the same thing as the overflow of GWF.

So relatively similar. Ok . So what if you make this the reason for TWF ? First answer is more attacks right ? But suddenly that becomes too strong right? What if more attacks if you keep landing hits? Infinite dmg is posible then, so reduce to proficiency bonus in number of total attacks ?

Looking for cool solutions here. Less antagonizing more problem solving.

[Cybren]

It's relevancy is to see if TWF lines up to similar things. You don't need to make a claim, we already established the dpr model doesn't matter, what matters is if TWF lines up In any way to everything else.

I am not particularly invested in the TWF quality debate. I accept the present reality of two weapon fighting being an option best for rogues or for characters in featless games and don't really have much problem with it. If anything, my objections with TWF are aesthetic (as I have already mentioned, the lack of One handed + light being an optimal combination outside of external factors, as an example, TWF being historically a mostly defensive choice being another)

Because it seems to me like you just satisfied with pointing out flaws but not bothering with coming up with solutions, or anyway to begin with coming up with solutions. If you don't care than why say anything at all? To be right? Work with me here .

I didn't make the original claim, Knaight did. But Kryx consistently misinterpreted it because he seemed to have erroneously believed Knaight was advocating for TWF as having numbers consistent with other options. I am pointing out that Knaights claim was that purely using the model of DPR can fail to capture some of the utility in having more but weaker attacks. Which is valid. Then trying to claim that he was saying that "having more but weaker attacks is better" is immaterial. That wasn't the claim.

If TWF needs fixing, what if not dpr can you tell this by? If it's balanced than by what means is it in line with other similar options. If the model does not account for something. Then account for it now. This is an open question not specifically directed at you , but I ask for you to answer anyway.

I do not accept the premise that TWF necessarily needs fixing, and I do not give two gray oozes about comparing it to polearm master.

[Mortis_Elrod]

Snip

Ok I can accept that.

[SharkForce]

Would like to reiterate that spells like magic weapon and elemental weapon are less effective for TWF characters. Between that, crits rewarding larger damage dice, potentially needing to use two attunement slots for weapons, drawing / stowing, and competition for bonus action, a lot ought to be changed for TWF to be competitive.

I'm tempted to suggest the simplest fix: actually fighting with two weapons at once, instead of using them separately.

Base TWF:
- You may draw / stow weapons instead of one on your turn if those weapons are light.
- When you take the attack action and are wielding two light melee weapons, you may make a single attack with their combined weapon dice. If you make multiple attacks with the attack action, you may use this feature once for each attack. Your attribute bonus to the attack and damage rolls, and other static bonuses such as from magic weapons, only apply for one of the weapons, your choice.

Fighting style: You may two weapon fight with weapons that don't have the light property.

Dual Wielder:
- When you make a reaction attack and are wielding two weapons that qualify for TWF, you may make a combined attack as per TWF
- While wielding two weapons that qualify for TWF, if you take the attack action on your turn, you may spend your bonus action to make a combined weapon attack with these weapons as per TWF.

TWF would still do less damage than a great weapon barring special situations, and would still suck at critical hits. But at least this would make it more consistent with the rest of the game, bring its damage up above sword and board at least, allow it to benefit approximately equally from magic weapon and elemental weapon, and give it some interesting options (such as poison on both weapons).

suggested something like this a while back. general consensus: it mechanically helps, but doesn't give the "more attacks but each is weaker" feeling that people who love TWF style wanted.

I am not particularly invested in the TWF quality debate. I accept the present reality of two weapon fighting being an option best for rogues or for characters in featless games and don't really have much problem with it. If anything, my objections with TWF are aesthetic (as I have already mentioned, the lack of One handed + light being an optimal combination outside of external factors, as an example, TWF being historically a mostly defensive choice being another)

you've misunderstood something...

TWF isn't better for rogues or characters in featless games. TWF is better for featless rogues, and pretty much nobody else except for levels 1-4 which barely last any time at all.

two handed weapons are still better for damage, and iirc sword and board is extremely close to TWF DPR once you get two attacks, but has much better AC.

consider, for example, paired shortswords against a single rapier, with no fighting style: TWF is doing 3d6 + (2*attribute), while longsword/rapier/battleaxe/etc with no fighting style is doing 2d8 + (2*attribute). so you're looking at 3d6 vs 2d8 when you cancel out... average of 10.5 vs 9... TWF is very slightly better, but it's very close. with fighting style, it's 3d6 + (3*attribute) vs 2d8 + (2*attribute) + 4 (from dueling style), or 3d6 + attribute vs 2d8 + 4, which means 10.5 + attribute vs 13. again, TWF is very slightly better assuming an attribute modifier of at least +3, but it's only a 2.5 damage difference even with a +5 attribute modifier.

now if we start factoring in potential other uses of bonus actions on occasion (like a fighter wanting to use second wind, a barbarian using rage, and so forth), you're going to lose some of that already very small lead (say, you only get 5/6 of the value of that bonus action attack instead of all of it to reflect that maybe 1 in 6 rounds you do something else with your bonus action)

and when you factor in hit chance and the slightly higher damage gained from crits for the duelist, it gets even closer. the TWF build will be *slightly* ahead in average damage as compared to the sword & board build... but it is *very* close, even in a featless game, and in a featless game sword and board has a +2 AC advantage for their extremely small damage deficit. it's not easy to model *exactly* how much +2 AC is worth, but it's fairly simple to work from the assumption that it's fairly close to the damage difference between a greatsword build and a longsword + shield build, which means TWF isn't getting anywhere near enough.

[EvilAnagram]

TWF isn't better for rogues or characters in featless games. TWF is better for featless rogues, and pretty much nobody else except for levels 1-4 which barely last any time at all.

Yeah, this is nonsense. More attacks for a rogue is always better, featless or no. I cannot think of a single feat that removes the usefulness of a second attack.

Furthermore, TWF is perfectly useful until the Fighter gets that third attack. At that point, it slides well behind, but until that point it's competitive.

[Willie the Duck]

Yeah, this is nonsense. More attacks for a rogue is always better, featless or no. I cannot think of a single feat that removes the usefulness of a second attack.

Furthermore, TWF is perfectly useful until the Fighter gets that third attack. At that point, it slides well behind, but until that point it's competitive.

I'm pretty sure that the argument is that Crossbow Expert is the feat to take to get that second attack. It's not wrong. It just relies on the idea that the ranged rogue is always the way to go. That itself may or may not be wrong from a pure optimization perspective. However, there will always be players that want to play melee rogues. Also, regardless of any discussion on these boards or elsewhere about the skill system and whether or not hiding is extremely easy to do in combat (and thus easily setting up the ranged rogue for SAs), there are going to be plenty of DMs who will nix that idea, so the dual wielding rogue is definitely a solid build to use for the wider gaming public.

[Citan]

Please provide the math comparing TWF vs PAM showing that the loss of damage for TWF is due to "damage overflow".

Except that any comparison should include with and without feats, and without feats all arguments in favor of PAM/GWM lose.

Beyond that, taking two Barbarians, Rogues, Fighters or similar, which both start with the same starting array, just distributed differently...
While the first takes PAM/GWM which restricts him to STR, the second can take Sentinel.
While the first now takes Sentinel, the second can bump STR.
While the first can max STR, the second can grab Resilient: Wisdom, Mage Slayer or even Magic Initiate or Defensive Duelist.

Etc etc...
You always compare the styles including feats, but that's a lazy thinking, because, as you and others told yourselves, the Dual Wielder feat is not bringing much in terms of pure offense. So you are perfectly fonctional with just the related fighting style, and just a bit subpar without it.
Whereas all other builds need at least one feat, and get channeled into a very particular and limited style. So either channeling you into specific races, or pushing away other importants ASI/feats.

That's why TWF is good in itself. Dual Wielder just stresses the intrisical versatility by giving more options of build and equipment. You can build on it, or you can build without it.

Once you leave the pure theorycraft to get into actual games, you realize TWF is very worth it. The fact that other beat him in damage or defense at highest levels is irrelevant, because it's their schtick in the first place. TWF is versatility, from level 1 to level 20. :)

[MadBear]

Except that any comparison should include with and without feats, and without feats all arguments in favor of PAM/GWM lose.

Beyond that, taking two Barbarians, Rogues, Fighters or similar, which both start with the same starting array, just distributed differently...
While the first takes PAM/GWM which restricts him to STR, the second can take Sentinel.
While the first now takes Sentinel, the second can bump STR.
While the first can max STR, the second can grab Resilient: Wisdom, Mage Slayer or even Magic Initiate or Defensive Duelist.

Etc etc...
You always compare the styles including feats, but that's a lazy thinking, because, as you and others told yourselves, the Dual Wielder feat is not bringing much in terms of pure offense. So you are perfectly fonctional with just the related fighting style, and just a bit subpar without it.
Whereas all other builds need at least one feat, and get channeled into a very particular and limited style. So either channeling you into specific races, or pushing away other importants ASI/feats.

That's why TWF is good in itself. Dual Wielder just stresses the intrisical versatility by giving more options of build and equipment. You can build on it, or you can build without it.

Once you leave the pure theorycraft to get into actual games, you realize TWF is very worth it. The fact that other beat him in damage or defense at highest levels is irrelevant, because it's their schtick in the first place. TWF is versatility, from level 1 to level 20. :)

Ok, from what I can see you've not provided any evidence, just a bunch of claims.

Since, you've repeatedly done this, let me ask. What evidence and in what form would you need to see to convince you that twf is subpar? Or are you convinced that it's not possible to change your mind?

[Willie the Duck]

Etc etc...
You always compare the styles including feats, but that's a lazy thinking, because, as you and others told yourselves, the Dual Wielder feat is not bringing much in terms of pure offense. So you are perfectly fonctional with just the related fighting style, and just a bit subpar without it.
Whereas all other builds need at least one feat, and get channeled into a very particular and limited style. So either channeling you into specific races, or pushing away other importants ASI/feats.

It's not lazy to include feats. Many, probably most, 5e gamers use feats. Excluding the possibility of them being in the mix would be shortsighted. Beyond that, regardless of whether it is appropriate to the analysis or not, it isn't lazy. That really doesn't even make sense. I don't know whether I have a problem with you refusing/declining to back up your claims with anything except vague assurances that the versatility of twf makes it worth it, the way that Kryx does, but it's pretty clear that he's not being lazy, and you're not making your own claims look well backed up (not the other way around).

Once you leave the pure theorycraft to get into actual games, you realize TWF is very worth it. The fact that other beat him in damage or defense at highest levels is irrelevant, because it's their schtick in the first place. TWF is versatility, from level 1 to level 20. :)

And this is a platitude, not an argument or evidence. You may as well have said, "once you get your head out of the clouds, you'll realize how right I am." That's the exact same level of actually providing evidence to support your point.

[malachi]

Here's some math, assuming that all attacks have 60% chance to hit (35% with GWM, 84% with advantage from shield master), STR/DEX is +3 at lvl 1, +4 at lvl 4, and +5 at lvl 8. I don't factor in increased crit damage, though.

All damage listed is damage / turn.

Fighter::
With no fighting style or feat::
Lvl 1 (1 attack, +3 STR): Both greatsword and 2x shortswords gets you 6 dmg / turn.
Lvl 5 (2 attacks, +4 STR): Greatsword wins at 13.2 dmg / turn. Polearm (11.4) and 2x shortswords (11.1) are close, with S&B (10.2) in last. TWF fails here, since the other options either do a bit more damage, have comparable damage and reach, or more damage and higher AC.
Lvl 8 (2 attacks, +5 STR): same analysis as lvl 5. Greatsword (14.4) > Polearm (12.6) > TWF (12.3) > S&B (11.4).
Lvl 11 (3 attacks, +5 STR): Greatsword (21.6) > Polearm (18.9) > TWF (17.4) > S&B (17.1)
Lvl 20 (4 attacks, +5 STR): Greatsword (28.8) > Polearm (25.2) > S&B (22.8) > TWF (22.5)

Summary: TWF suffers here, but if you're taking Defensive Fighting Style, you probably want S&B to emphasize that.

With fighting style (reroll 1's and 2's for greatsword / polearm, dueling for S&B, and add stat to dmg for TWF):
Lvl 1: TWF (7.8) > Greatsword (6.8) > Dueling (5.7) > Polearm
Lvl 5: Greatsword (14.8) > TWF (13.5) > S&B (12.6) > Polearm (12.36)
Lvl 8: Greatsword (16) > TWF (15.3) > S&B (13.8) > Polearm (13.56)
Lvl 11: Greatsword (24) > Dueling (20.7) > TWF (20.4) > Polearm (20.34)
Lvl 20: Greatsword (32) > S&B (27.6) > Polearm (27.12) > TWF (25.5)

Summary: TWF is never as good as the greatsword, but is 2nd best until the 3rd attack comes up.

With Feat (-5/+10 and 5% crit chance for Greatsword, PAM*, Shield Master for advantage**, and Dual Wielder for d8 TWF) and a fighting style:
*I didn't include the easier time getting opportunity attacks, since I wasn't sure how to model it.
**Assuming that you always get to knock prone for advantage
Lvl 1: TWF (9) > PAM (8.88) > S&B (7.98) > PAM
Lvl 5: S&B (17.64) > Greatsword (16.415) > PAM (16.26) > TWF (15.3)
Lvl 8: S&B (19.32) > PAM (18.06) > Greatsword (17.15) > TWF (17.1)
Lvl 11: S&B (28.98) > Greatsword (25.725) > PAM (24.84) > TWF (22.8)
Lvl 20: S&B (38.64) > Greatsword (34.3) > PAM (31.62) > TWF (28.5)

Summary: TWF is ahead of the pack at lvl 1, but falls behind by lvl 5. It keeps a ahead of S&B assuming you never get to knock the target prone (3.3 @ lvl 1, 2.7 @ 5, 3.3 @ 8, 2.1 @ 11, 0.9 @ 20).

If you have a feat but not the fighting style, TWF falls into last place by Lvl 5.


Barbarians:
When raging without feats or fighting styles, TWF barely deals the most damage at lvls 1 and 20, is 0.9 short of the greatsword at lvls 5 and 8, and is 0.3 damage short of the greatsword at lvl 11.

When raging with feats and reckless attack (not listing S&B damage, because the feat literally doesn't help in terms of damage with reckless attack):
Lvl 1: PAM (15.12) > TWF (13.44) > GWM (13.34)
Lvl 5: GWM (27.89) > PAM (26.46) > TWF (23.1)
Lvl 8: GWM (29.11) > PAM (28.98) > TWF (24.78)
Lvl 11: PAM (31.5) > GWM (30.32) > TWF (27.3)
Lvl 20: PAM (34.02) > GWM (31.53) > TWF (29.82)

Summary: TWF is never the right choice for a barbarian.

Note:I'm too lazy to list this out, but with a feat and without reckless attacking, PAM > TWF > GWM pretty consistently throughout the career. S&B getting advantage through Shield Master beats non-reckless attacking TWF at lvl 5, and stays ahead by about 1 dmg / turn.

Note:If you dip into Fighter for a fighting style, TWF is only 1.34 damage / turn behind PAM.


Paladins:
With just a feat (assume Defensive fighting style):
Lvl 1: PAM (8.4) > GWM (7.35) > TWF (7.2) > S&B (6.3)
Lvl 5: GWM (15.44) > PAM (15.3) > S&B (14.28) > TWF (12.9)
Lvl 8: PAM (17.1) > GWM (16.17) > S&B (15.96) > TWF (14.1)
Lvl 11 & 20: PAM (25.2) > S&B (23.52) > TWF (22.2) > GWM (19.48)

Note: With no feats, style-less TWF is only 1 pt of damage less than the best fighting style (GWF) at levels 11+.
Note: Without a feat, fighting style, or expending any long-rest resources, Paladins are better at TWF than other weapon types at 11+.
Note: With a feat and fighting style (assuming you can get the TWF style), PAM beats TWF at 5+ (by about 1 pt of damage).


EDIT:
All this assumes VHuman for comparison, since that allows you to get your 1 necessary feat at level 1. The numbers will obviously be different at lvl 5 if you have to drop an ASI for your feat. And obviously, PAM + GWM or PAM + Sentinel will get you more damage - but I don't want to type up those changes.

[Easy_Lee]

I'm pretty sure that the argument is that Crossbow Expert is the feat to take to get that second attack. It's not wrong. It just relies on the idea that the ranged rogue is always the way to go. That itself may or may not be wrong from a pure optimization perspective. However, there will always be players that want to play melee rogues. Also, regardless of any discussion on these boards or elsewhere about the skill system and whether or not hiding is extremely easy to do in combat (and thus easily setting up the ranged rogue for SAs), there are going to be plenty of DMs who will nix that idea, so the dual wielding rogue is definitely a solid build to use for the wider gaming public.

That's one class out of six melee classes that might TWF. Even then, crossbow expert is better even if you want a melee rogue because it lets you fire hand crossbows in melee. Even if the DM is a stickler about having a free hand for loading, a melee rogue can still work around it.

Not holding rapier at start of round: make two crossbow attacks, load crossbow, draw rapier.
Holding rapier at start of round, crossbow loaded: attack with rapier, bonus action fire crossbow
Holding rapier at start of round, crossbow not loaded: attack with rapier, sheath rapier, load and fire hand crossbow, reload crossbow. (Or sheath rapier and fire twice if your DM is a major stickler about the holding a loaded hand crossbow part).