I recently posted What class has the most damage output per round at level 6?, asking about character optimization. Without additional context, the question was:

what are the most optimizable class/multiclass options in the early levels and later levels? Is the answer different for high and low level characters?

I've since edited the question in response to comments. The general consensus seemed to be that it was too broad, rather than too opinion based. Here are my revisions:

what are the most optimizable class/multiclass options for both damage output and survivability at levels 5-6?

what are the most optimizable class/multiclass options for damage output at levels 5-6?

what are the most optimizable class/multiclass options for damage output per round in a single encounter against a small group of high HP enemies at levels 5-6, assuming the enemies have no damage resistances and the character has just finished a long rest (i.e., all abilities are available)?

and finally:

What is the most optimizable class from the Players Handbook for damage output per round in a single encounter against a small group of high HP enemies at levels 5-6, assuming the enemies have no damage resistances and the character has just finished a long rest (i.e., all abilities are available), and has no buff effects from allies?

It was closed with this message:

This question is opinion-based. It is not currently accepting answers.

Can someone explain how this is opinion based? I'm not saying it was a great question, but to me, while it is overly broad (or at least was before I edited it), is is inherently objective. Survivability and damage output are both well defined and quantifiable. So how is this opinion based? I'd like to stress that I'm not trying to say it shouldn't have been closed, but that I disagree with the reason stated for closing it.


The question was re-opened, and then re-closed, and then re-opened again. We'll see what happens in the future. Here is the final(?) version:

What is the most optimizable class from the Players Handbook for damage output per round in a single encounter, assuming the character:

  • is fighting four Brown Bears
  • has just finished a long rest
  • has no magic items
  • uses the standard ability array or point buy,
  • has no buff effects from allies.

and to narrow it down further:

  • use only RAW content from the Players Handbook

For the sake of clarity, let's define damage per round as the amount of damage a character can do averaged out between 5 consecutive rounds (assume the character doesn't die). Damage from Area of Effect spells hit all four bears. For simplicity, leave out summoning or animating spells.

  • 1
    \$\begingroup\$ Related/of use from the FAQ: Why was my question closed as too broad, unclear, or opinion-based? and probably also useful reading (as it related to character build and optimization): Are character build or optimization questions on topic? \$\endgroup\$
    – Someone_Evil Mod
    May 23, 2021 at 19:36
  • 1
    \$\begingroup\$ Why are you making another meta? \$\endgroup\$ May 24, 2021 at 16:40
  • 1
    \$\begingroup\$ Because I think it's good to have a more general discussion about questions that get closed and reopened and closed again (and possibly reopened again) \$\endgroup\$
    – Jasmine
    May 24, 2021 at 17:52
  • \$\begingroup\$ @JakeFuller I look forward to it. \$\endgroup\$ May 24, 2021 at 18:36
  • \$\begingroup\$ Do you feel like the question is one that is helpful to you now? Or has it been made 'stackable' to the detriment of a useful answer to you? Are you really limited to only PHB at your table? Why don't you want to include summoning/animation as a possible solution? \$\endgroup\$
    – NotArch
    May 25, 2021 at 14:05
  • \$\begingroup\$ I feel like summoning/animating would make things overly complicated, but maybe I’m wrong. The question is certainly far more specific than I originally intended, but remains relevant to me, although it’s now more like a “because I was curious” question. The campaign I was in that inspired this was limited to PHB, so that’s why I went that way with it. Also, I didn’t know bounties made it so a question couldn’t be closed, so I apologize for that \$\endgroup\$
    – Jasmine
    May 25, 2021 at 14:26
  • 1
    \$\begingroup\$ The goal of the figuring out constraints is in giving us the guidelines for what you'd expect to see when using this. So the only PHB thing makes sense. The lack of summoning because complexity doesn't if that's going to be allowed at the table. Think of the other possible constraints as well for what could help drive an answer. If there are 'none', then the question is probably too broad at that point. But you've done a great job in working to narrow, I was just making sure the constraints match what you actually want answered. \$\endgroup\$
    – NotArch
    May 25, 2021 at 14:34

4 Answers 4


This type of theorycrafting often gets an uneven reception here

My best advice is to not ask about multiclassing from all possible options, but that's not a guarantee that the question won't still be closed.

My impression based on experience (read: I don't have hard evidence) is that there are relatively few users here who like to answer optimization questions, at least compared with the number of users who are likely to attempt answers to other questions. And answering them tends to involve a substantial amount of work, as it is necessary to compare all available options within the given constraints.

It follows from that that those constraints are very important. As mentioned in Trish's answer here, a single extra element can add a huge number of additional possibilities to compare. And people vary in which constraints they think are crucial to answering, prompting lots of requests for clarification which may be orthogonal to each other.

Usually this works itself out by answerers self-selecting: people who are interested in tackling an optimization question and feel there is enough information presented to do so will post answers. The problem comes from close votes. People can vote to close for any reason (no matter what the close vote explanation they list is). So you can easily get five people each voting to close due to five totally different reasons which are blurred together by the reasons presented in the VTC popup.

To date, there isn't a great way to proactively avoid this. You can specify as much detail as you like, but as long as there are at least five people that feel a question is lacking they can close it. And people seem to get a bit... territorial over close votes in optimization questions. Which can make it hard to shake a previous closure. We mostly keep voting wars contained, but the sort of swing you've noticed on this question happens from time to time.

The earliest versions of this question really were underspecified and not feasibly answerable in the SE format (which isn't a huge deal; it happens!). Iteration is normal and desirable here. But now peoples' eyes are on the question and you will have to overcome the objections of, at least, all but four of those people.

Stick with it, and it's very likely that eventually the question will be open, at least for long enough to attract an answer or two. It's just rarely a smooth road.

I present here a few similar(ish) questions without further comment:

  • \$\begingroup\$ Thanks, I appreciate this. I'll keep working on it. They reason I posted the question here rather than somewhere else is precisely because RPG.SE has such rigorous standards for questions and answers. In other words, when I post it here I know that I'll get a better answer than "generally the most notable for durability are bear totem barbarian and moon druid, damage vs multiple foes favors evocation wizard, single target damage favors paladin or fighter." (actually taken from the same question on DNDbeyond) \$\endgroup\$
    – Jasmine
    May 24, 2021 at 22:01

For me, it needs more focus and more details.

We do handle optimization questions. See . When you look at the optimization questions dealing with optimizing character performance that have scored well, you will observe two formulae that work well for the stack:

  • A mostly complete character build, where the stack is asked to optimize a few remaining build choices for a particularly outcome
  • An open ended question, where the stack is asked to create character build that optimizes one well defined outcome.

It is this second point that you need to try to emulate. Take for example this question: What is the highest possible AC?

Here, we are given few constraints, but the sought outcome is well defined - it needs no further details. The trouble with your question is that the outcome is not yet well defined. In comments, several users point out a plethora of variables you should consider constraining:

  • Right now this question is very broad. Which class/multiclass is the best damage dealer/damage taker changes nearly every level, and would change depending on burst damage or longevity. I would recommend you narrow this down: pick a specific level, pick a specific combat style, pick a length of time they will need to deal/receive damage over, etc.
  • Some of this depends on the group - you don't need to pour so many resources into your health and AC if you're surrounded by barbarians who are absorbing all the hits - you could be focusing a little more on damage. Are we to assume your character is on their own? (Just a warning, if you play like this it isn't the nicest play style for the group) –
  • How many enemies can we assume an AoE includes? What are the enemy AC and saving throw bonuses? When you say we should use "damage output per round" how does this account for abilities that cannot be used every round? Or did you mean that we are trying to optimize damage output of one round? If so, do we have a round or several to set up effects like wall of fire, spiritual weapon, or haste? Should we assume all dice rolls are entirely random (we cannot assume we hit, or enemies fail their saves, or guarantee rolls on Wild Magic Surges? Damage dice roll their average value)?
  • Even as edited, I think this question is far too broad to be answerable. The number of possible combinations -- factoring in every class, spell, and magic item available, not to mention situational parameters -- is so staggeringly huge that it isn't realistically possible to offer a demonstrably "best" answer, which is what RPG.SE aims to do.

We need a well defined understanding of what you mean by "damage output per round":

  • Single target, single round?
  • Multi target, single round?
  • Single target, 3 or 5 round damage?
  • Multi target, 3 or 5 round damage?

Defining exactly what you're looking for here will significantly reduce the number of options we have to consider when answering your question. For example, Monks are very good at sustained damage against a single target, but against a group of enemies, a caster with a sustainable area of effect spell will perform much better than a monk.

Further, we need a well defined method of calculating damage. Because of the complex nature of damage calculations, we need at least these things:

  • Number of enemies
  • Their ACs
  • Their saving throw bonuses
  • A method of adjudicating AoE spells/effects, e.g. a description of the enemies' relative positions on the battle field

At time of writing, you have some of these things, but still not enough to make the problem well defined.


A question generally is opinion based, if there could be at least 2 answers that are vastly different in their results and neither is, objectively speaking, better.

For example the question "What's the best color for a wizard's robe" allows the whole spectrum for answers, none is objectively the best. But "What is the color of wizard robes in the Dragonlance at the time of Tiamat's invasion?" is a clearly defined question that has one answer: The colors of the three moons, Black, White & Red.

Now, let's see where your question...

I used to DM a group of players who liked to heavily optimize their characters (and the group as a whole) to take as little and deal as much damage as possible (seriously, these guys were crazy). It got to the point that I was routinely throwing encounters that would be considered "deadly" to for a group their level. Unfortunately some of the members of that group moved away at around 6th level, so I never got to see how it carried on to later levels.

Most of this explains your background and everything. That part is fine.

So my question is: what are the most optimizable class/multiclass options for both damage output and survivability at levels 5-6? I'm playing D&D 5e.

Now, here is the problem: you are asking to analyze all classes and class combinations 1. That sounds like the wrong closure reason prevailed (I believe I voted needs focus) because that's simply too broad at first blush. On second thought though, this also does match the opinion-based threshold: what is most optimizable depends on opinions, especially what constitutes optimizable or best. A class could be the best at dishing out damage in a single attack, or could do the most healing, but which of the two is objectively better?

How to improve?

First of all, we could focus on one of the two problems only. Like damage output. That puts much more focus on the question.

Next you should say which material is available. Basic rules? Core Rules? Any other book? Keep the allowable material list short.

Then you could reduce the number of classes that is allowed to be dipped into. For example "At maximum 2 classes may be used, the second class needs to be chosen at level 2". That cuts the math down tremendously: With 12 classes, it would get us down to 29 and 31 combinations respectively. Still a lot but it is in the realm of doable instead of There's so much to look at, and most is noise

Further confines could also help, for example, disallowing classes or combinations of some classes. A random example would be "No arcane-caster/divine-caster combos".

  1. The basic rules have 12 classes, so at level 5 that is \$12^5\$ combinations that are possible, at level 6 it's \$12^6\$. Those are mind-boggling 250k and 3 Million possible combinations.
  • 2
    \$\begingroup\$ If I’m remembering correctly my close vote was for “needs more focus”, and you are correct to observe that the lines between “needs details”, “opinion based”, and “needs focus” are often blurry, and sometimes all three apply and have the same solution. \$\endgroup\$ May 23, 2021 at 17:29
  • 1
    \$\begingroup\$ I do have the quantifiable criteria of damage/round and survivability--those are intended to provided what constitutes to best in this case. Is that not enough? \$\endgroup\$
    – Jasmine
    May 23, 2021 at 17:30
  • 1
    \$\begingroup\$ @JakeFuller It's also not clear to me if you want separate answers for "damage" and "survivability" or some sort of multi-objective optimization that accounts for both at the same time. The former is another case of much too broad; the latter is another case of subjectivity because there's no perfect objective balance between those two. \$\endgroup\$
    – Novak
    May 23, 2021 at 17:42
  • \$\begingroup\$ @JakeFuller Incidentally, I'm sorry if this feels like we're beating up on your question. It's a perfectly cromulent question in the sense that it's something about RPGs that's easily expressible that you want to know. But it's just not a good fit for a stack exchange question, for all the reasons we're mentioning. \$\endgroup\$
    – Novak
    May 23, 2021 at 17:43
  • 1
    \$\begingroup\$ @Novak No, this is helpful. I'll keep editing it to make it better, whether or not it gets reopened. \$\endgroup\$
    – Jasmine
    May 23, 2021 at 17:50
  • \$\begingroup\$ @JakeFuller maybe the additions help. \$\endgroup\$
    – Trish
    May 23, 2021 at 18:03
  • \$\begingroup\$ @Trish they seem to have helped. Thanks for the feedback! \$\endgroup\$
    – Jasmine
    May 24, 2021 at 0:25
  • 1
    \$\begingroup\$ This is in an inaccurate—and dangerous—definition of “opinion-based.” \$\endgroup\$
    – KRyan
    May 24, 2021 at 18:24
  • 2
    \$\begingroup\$ @KRyan Not giving an accurate alternative definition is unhelpful, please provide one so your criticism can be constructive \$\endgroup\$ May 24, 2021 at 20:15
  • 1
    \$\begingroup\$ @KRyan It is consistent with how I've seen it used, here. \$\endgroup\$
    – Novak
    May 25, 2021 at 3:58
  • 1
    \$\begingroup\$ Ok, while there may be some useful points raised there, the tone of that interaction didn't make it very productive. Please remember to be respectful and friendly when pointing out an issue. \$\endgroup\$
    – Someone_Evil Mod
    May 25, 2021 at 15:13
  • \$\begingroup\$ As for the actual combinatorics point, the maths isn't quite right, because the relevancy of which order the class levels were taken in isn't really gonna matter so the actual number of combinations isn't that high and I'm not sure how relevant it is because there are patterns inside those and you can do a lot of cutting based on those. We also have precedence from other, similar questions which allow full multi classing and a whole lot more. Optimization answers don't really need to fully demonstrate that they're the best because that's never workable. \$\endgroup\$
    – Someone_Evil Mod
    May 25, 2021 at 15:19
  • \$\begingroup\$ @Someone_Evil true, you don't need to do an in-depth look on all possible combinations, but you need to show your work why you can dismiss more than 99.99% of all theoretically possible combinations of classes. The main point of the math was to show how absurdly high the total number of possible combinations actually is. like, there are 665280 ways to pick 6 out of 12 classes. ABCDEF and FEDCBA are two of these and others are jumbled that. It was more a show of the magnitude a total and complete look might be. \$\endgroup\$
    – Trish
    May 25, 2021 at 15:24
  • \$\begingroup\$ @Trish But the needing showing of that isn't outside the scope of an answer, and I say that both based on the system and what has been done for numerous similar questions (and their answers) \$\endgroup\$
    – Someone_Evil Mod
    May 25, 2021 at 15:32
  • \$\begingroup\$ it's not impossible, but a demand to the answers that I most certainly don't want to have. \$\endgroup\$
    – Trish
    May 25, 2021 at 15:33

At the time I write this answer, the question was closed eight hours ago, and your edit was less than half an hour ago. Edited questions don't automatically re-open, they are automatically made eligible to be re-opened after a certain number of votes, as this one now is. (If the edit was made within a certain window of time, as this one was.)

Note that there's no guarantee that your question will be re-opened. In my opinion it is still either "too broad" or "subjective" (categories which sometimes bleed into one another because of how we use the term "subjective") to be answerable.


You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .