5e Class Design Guide

From D&D Wiki

Jump to: navigation, search
Stub Logo.png This page is incomplete and/or lacking flavor. Reason: As a community guideline, this page will forever be incomplete. Feel free to jump in and add your wisdom and insight to the community's standards!!

You can help D&D Wiki by finishing and/or adding flavor to this page. When the flavor has been changed so that this template is no longer applicable please remove this template. If you do not understand the idea behind this page please leave comments on this page's talk page before making any edits.
Edit this Page | All stubs

Making a Class? Stop Right Now. Think

Stop where you are, before you go actually crafting a class. You need to do a little bit of introspection before you do this. You need to understand where you are actually coming from and what you actually want before you can effectively go out and get it. So take off your boots, kick up your feet, and stay a while. Let's talk.

Let us have a chat...

So, you want to make a class, do you? Maybe you want a little help with that, eh? Classes are one of the most complex elements of the game, making them one of the hardest things to create. Very few classes are added to the wiki fully formed, and most have the combined efforts of multiple authors. Making a class is no simple task, and completing its creation can take months- even a year or more. Designing a functional class demands extensive playtesting, which takes a lot of time. This guide can help you on that long journey.

What are my actual objectives?

Think for a second about why you thought up a class. No doubt, you think it's just because you have a great idea for a class that just isn't in the game, but the truth is, most people never think about what actually drives their ideas. When I ask, what are your objectives, I don't mean "why do you want to make a class?", I actually mean, "what motivated you to start dreaming up anything at all in the first place?" What inspired your creativity to even start bubbling away? Many people think they just get ideas from nowhere, but this isn't really the case- all of the human consciousness is a reaction to external stimuli, we are a reflection of our world. Every thought, emotion, and idea, has a cause, a catalyst of some sort. Understanding what it is that sparked your imagination will help you greatly in successfully achieving your goals. So, what was it?

To simulate something from history?
To emulate something from a work of fiction?
To make a more powerful character?
To explore the rules and mechanics of the game in a unique way?
To overcome some obstacle not covered by the core rules?
To do a better job than the original creators, or to resolve perceived shortcomings, failures, or errors made in the core material?
To just make more options for the sake of having more options?
To create a class which will fit a setting better?

Whatever your motivations are, they are also your objectives. If you felt the creators did a bad job on Rangers, for example, your objective is likely to correct their failure. Remember how I said that we are just a reflection of our world? Our memories are part of the world we are reacting to. As a result, most of human thought is actually quite cyclical like this. It isn't a bad thing though, it's actually wonderfully convenient. Makes it very easy to understand yourself if you just take the time to think about it.

Are there other ways I could achieve my objectives?

Now the question is, "Why a class?". There are dozens of other types of game content you can create. Why did your motivations lead you to a class? In many cases, people dream up classes over other types of content because of several reasons.

  1. Players tend to think of D&D in terms of people and characters more than a collection of rules and content.
  2. Humans tend to think about people more than things- even going so far as to think of things as people.
  3. Classes have the biggest impact on a game but don't rewrite the whole game because they are centralized on a character.

Honestly, your idea may actually be a perfect example of one that should be a class. But stop for a second and consider, "what if it isn't?".


For many ideas, the honest truth is that the core classes alone can cover it just fine. Simply look at the mechanics without considering their context or flavour, and imagine all of the things such mechanics could potentially represent. Depending on equipment selection, the fighter class can effectively represent nearly every form of fighting class imaginable, with classes like Barbarian and Rogue covering the gaps with ease. It's easy to reimagine the core classes to make them cover all manner of strange and unique ideas not included in their class description.

  • Knight = Lawful Good Fighter Noble in plate mail riding a horse... But then someone makes Knight (5e Class).
  • Archer (5e Class) = Martial class who only uses ranged weapons.
  • Swashbuckler = Fighter, Rogue, or Bard, with highest ability score in Dexterity, with the Sailor Background. And yet we see Swashbuckler (5e Class), Swashbuckler, Variant (5e Class), and Daring Swashbuckler (5e Class).
  • Drunken Brawler (5e Class) = Barbarian. Just replace rages with stupors.
  • Alchemist = Any spellcaster can do this really, but wizard works best. Just describe all of their spells as the effects of potions! That's what all those material components are for!
  • Shade (5e Class), you would think having a darkness based subclass for the monk would be enough, but this class reinvents the monk with more options to better fit the theme of a "dark monk".
  • Arcane Warrior (5e Class), as a concept, already existed in several forms in the core material. Most obvious would be the Eldritch Knight subclass for Fighter and the Arcane Trickster subclass for Rogue, but a combat-oriented sorcerer could also pull it off!

If the idea is more about flavor and style, and the exact mechanics aren't that important, it may better serve as a background, if it is flexible enough. For instance, a historically accurate Shogun could be more useful as a background, as it could be applied to a variety of classes, allowing a wide range of fighting styles with the same flavor and context. (Furthermore, the Soldier background is open and flexible enough that, with a bit of research on your behalf, can represent a historically accurate shogun on its own.)


If your class is more about what you are than what you do, it may function better as a race. Races have an instant, permanent, and universal effect on your character, and can have a strong mechanical influence, as seen with the Dragonborn and Aarakocra. The biggest advantage here is that races are much easier to construct and test, but can have just as much impact. With the existence of subraces as well, it's easy to make highly variable and flexible races, as has been exemplified by the Genasi and Elves.

  • Demigod (5e Race) is a difficult idea to express without templates; as a race, you lose your mortal half; but as a class, it's just lacking material, and isn't really about what you "do" anymore.
  • Dhampir (5e Race) is another great example of people struggling with the absence of templates in 5th edition- this one takes the "half-elf" approach, assuming all half-breeds are half human.
  • Skeleton (5e Race) is a good example of trying to represent undead characters. Like demigod, it is inappropriate as a class, but somehow still loses something as a race.
  • Zombie (5e Class) case in point, here is an example of someone trying to do the same thing, but coming to the opposite conclusion, manifesting undeath as something that can be empowering.


If your idea is just going to replace some elements of a core class, it may work as a subclass. This allows you to flavor a class mechanically in all sorts of ways without resorting to the onerous task of building a balanced and meaningful class. However, subclasses only affect a character at certain intervals, and only start at third level.

  • A "bloodmage" subclass for Sorcerer, which lets you sacrifice HP or HD, or take ranks of exhaustion, in exchange for sorcery points, spell slots, metamagic effects, etc.
  • Samurai (5e Subclass) Here's that shogun we were talking about earlier, expressed here as a fighter subclass.
  • Gladiator (5e Subclass) Here's a subclass that is reimagining the gladiator background variant into a fighter subclass.
  • Cavalier (5e Subclass) And, of course, the perennial example, yet another way of expressing the idea of a knight.
  • Ninja (5e Subclass) are usually expressed as full classes, but here someone has built one as a rogue subclass.
  • Pirate (5e Subclass) is yet again being expressed, this time as a subclass, despite already existing in the core material.

Class Variant

If your idea is going to change some fundamental component of a core class, but still essentially function the same, it may be better suited as a class variant, which we actually have a section for here on the wiki. The benefit of having class variants is that they allow players more options in character creation, but bypasses the weirdness of cross-classing variants to make min-maxed mary-sue characters.

  • A druid that uses the cleric spell list.
  • A mashup of fighter and rogue features.
  • A ranged combat focused barbarian.

Variant Rule.

If your class covers activities which could be considered as extremely broad or inordinately complex, it may be that you are trying to do something classes were never intended to do. (For example, trying to make a class which revolves around waging war through the command of thousands of soldiers.) Such ideas may be more easily created as variant or supplemental rules. (The military commander class example could be better crafted as a mass combat warfare rule set, for instance.)

Is my class capable of standing next to the other core classes as an equal?

The biggest question of all here, honestly. If you look at the core classes, do any stand out which are clearly substantially superior or inferior to the rest? Some may have imbalanced advantages and disadvantages, but it ultimately balances out. This internal imbalance gives characters reasons to rely on one another and uses their strengths more creatively to overcome their weaknesses. This balance gives each character one or two functional roles they can play in the party- a place where they each belong, with the party composition and functional roles being filled determining the personality of the group. It is best for a class to have more than one functional role, as this allows groups with more than one character of a given class to remain functional, as the two characters can serve different purposes with their common gifts. Even the fighter can play a few different combat roles, including support, (A stealthy halfling sniper/archer) tanking, (A powerful half-orc in plate mail with dual hand-axes) or even spell-blasting (High Elf taking the Eldritch Knight subclass and focusing on accumulating spell-casting magic items)! So, when you look at the classes and all of the amazingly beautiful ways they can fit together to satisfy various practical roles through variations in build and play style, think to yourself, "Where does my class fit in?".

Class Do's and Don'ts

This covers design errors that are often made. Some of these are conventions (e.g. "round-counting" doesn't break the game but is at odds with 5e design philosophy), whilst others are just wrong.

  • Class features are "features", not "abilities" or "skills"
  • Strength, Constitution etc. are "abilities", not "stats". "Statistics" refers to all the game numbers of a game entity.
  • "Hit points" - not "health" or "life".
  • Class names are (usually) not proper nouns.
  • Feature names and subclass names are (usually) proper nouns.
  • Spell names are in lower-case italics.

Class Features

  • Wrong: "before needing to take a short/long rest", "between short/long rests", "per short/long rest" - the exact recharge point is ambiguous.
  • Correct: "You need to finish a short/long rest before you can use this feature again" - clearly states when you get the feature back.
  • Features and spells can target "creatures", "creatures you choose", "friendly creatures".
  • While creatures can have a "hostile" attitude towards you, you would normally use the phrase "creatures you choose" to avoid ambiguity.
  • Do use the standardized durations: 1 round, 1 minute (the length of one combat), 10 minutes (one "exploration phase"), 1 hour, etc.
  • Do allow a creature to repeat a saving throw for applying effects on enemies for an indeterminate length.
  • Don't use round-counting, e.g. "lasts for 3 turns", or "1d4 turns". If you must have a variable length, consider using a d6 recharge (the method used in the Monster Manual)
  • Don't have features that last for the duration of the "combat" or "encounter", as this is not defined (but see "1 minute" duration above); similarly do not have "at the start of combat/encounter" (but you could have something trigger on an initiative roll).
  • Don't forget to say how the feature is used. Is it an action, a bonus action, a reaction? Is it triggered on another action, or does it work in a general (non-combat) way?
  • Be careful with features that grant you a benefit if you take damage, as it's trivial to arrange to be damaged by a willing player or NPC.
  • Do have the benefit only work against the creature that caused the damage.
Regaining hit points
  • Be careful not to disrupt the hit point economy of using hit dice to regain hit points during a short rest. These examples allow a character to regain all their hit points too easily:
    • Don't have unconditional regeneration.
    • Don't have anything "count as" a long rest.
    • Don't have at-will "cure" effects.
  • DO use temporary hit points.
XP and treasure
  • Don't make features that change the way XP, gold, or treasure is granted. These rewards are entirely the purview of the DM.
  • Think very carefully before making a class dependent on a piece or set of equipment. What happens if a character is missing one or more items? Does the character still have class abilities or a niche that he can fill without equipment? If the answer is "no", then the class is probably too limited in scope.
  • Similarly to the rules regarding XP and treasure, the starting equipment that a character begins with is entirely the purview of the DM. Don't assume characters can be allowed to begin play with game-breaking starting equipment.
  • Don't give a class exclusive access to mundane equipment. If a piece of equipment is craftable, why can't it be bought for money or used by anyone else? Why should a DM be forced to allow a piece of equipment to exist in his campaign world? Many "gunslinger" classes and the like on this wiki have been guilty of violating these assumptions and being illogical as a result.
  • DO make separate pages for new equipment you want to introduce, for example on a separate page in 5e Weapons, rather than on the class page itself. If an item with the same name already exists, feel free to make a variant that fits your vision of how it works.
  • From WoTC: "...never assume that a particular feat will be a part of the game. For instance, a class can't refer to a feat and feats should never be granted as class features."
In General
  • Do give your class features to help with combat, exploration, and interaction
  • Do think of new class features rather than copying features considered unique to a particular class (e.g. the barbarian's rage)
  • Don't make your class concept too narrow. Particularly when adapting a character from another work of fiction, don't tailor the class to fit that character exactly: pull back and consider a broader scope with options that could make that character.
  • Don't have negative features that punish players for picking a certain class.
More Tips!

These aren't strict instructions, but may help to make your class feel and look better.

  • Don't invest a lot of writing in describing the effects of edge cases that might never crop up. For example, in one class, I saw a big paragraph about what would happen if they cast a spell without using their focus. Such a situation is a rarity. It would be better to keep the effect of that situation simple, so the class isn't cluttered with nonessential information.
  • Be mindful of the span of time Dungeons and Dragons is played in! Most groups play once a week, but some have bigger blanks between their sessions. And Many sessions are needed to reach level 20. Class design that "balances" different tiers of play against each other does not work well because of this. For example, a class that under-performs at low levels to justify over-performing at high levels may leave a player frustrated for months, and afterwards leave the rest of the table frustrated for even longer. Classes should try to fit within accepted ranges of balance (and fun) at all levels, because they will be played for a long time at every one of those levels.

Terminology and Mechanics from Older Editions

  • Fortitude save - most commonly "Constitution saving throw", but Strength is sometimes more appropriate.
  • Will save - most commonly "Wisdom saving throw", but Intelligence or Charisma are sometimes more appropriate.
  • Reflex save - use "Dexterity saving throw"
  • "DR" (e.g. "10 DR/all") - See damage resistance, PHB p. 197.
  • Full round action - not used. For example, characters can make all their available attacks and still move their speed.
  • Move action - not used. Movement is no longer a discrete action.
  • Attack of opportunity - now it's "opportunity attack".
  • Minor action or swift action - use "bonus action"
  • Free action - not used. Examples from SRD:Free Actions: Dropping an item is not an action (while putting one down is an object interaction [p. 190]); dropping prone is part of your movement (p. 190); speaking is "other activity" (p. 190); ending concentration on a spell is not an action (p. 203); casting a quickened spell is a bonus action (p. 102)
  • Immediate action - use "reaction"; don't forget to say what triggers it.


  • Supernatural, Spell-like and Extraordinary - not used. A spellcasting feature is declared as such, and attacks can be designated as weapon or spell. The actual features have no designation.
  • "Times Per Encounter" - Becomes "Use(s) of this feature is/are restored after a short or long rest."
  • "Times Per Day" - Becomes "Use(s) of this feature is/are restored after a long rest."

5e Class Preload Walkthrough

Read the dang book.

OK, now that we've given you a better idea of what you want, and hopefully redirected some people to present their content in its most appropriate form, let's talk about making your class because you've obviously made up your mind! The following section gives a detailed walkthrough of the class preload, explaining how to go about filling out each section to get the most out of it. As with everything on the wiki, please make sure that you are familiar with the precedent, which in this case is set by the 5th edition corebooks: the Player's Handbook, (PHB) Monster Manual, (MM) and Dungeon Master's Guide, (DMG). The core rules classes can be found on PHB p.45-112 and you should read Character Creation Step-By-Step on PHB p.11. Guidelines on how to create a class can be found on DMG p.287-289. WotC also released an unearthen arcana pdf on how to modify existing classes, which also provides some good insight on how the first party produces its content. You can find guidelines on converting material from previous editions into 5th edition rules in this freely available PDF. This content is subject to The Three Pillars of Adventure (5e Guideline) and Understanding Bounded Accuracy (5e Guideline). When you open up a class creation page for the first time, you will see Class Do's and Don'ts (5e Guideline) at the top of the page. It's better to read them right now while you're still planning.

Aside from that, a general recommendation: Know the entire rest of the game. Read the PHB and DMG extensively and play the game a lot. If you can't be bothered to understand the game you're playing or play the game you're interested in, you're kind of missing the point.

It is best if you have had experience brewing up other types of content, such as equipment, spells, races, backgrounds, feats, and monsters before making a class. Having experience with writing houserules and adventures is also highly beneficial. Classes directly interact with every single other element of the game in incredibly complex ways. Not having a basic knowledge of the relationships between classes and all other game elements can lead to imbalanced or dysfunctional results.

Understand the classes and how they interact with each other especially. See how they fit together, fill in each other's weaknesses, reinforce each other's strengths, and where any gaps may seem to exist. You may be shocked to discover just how precisely the core classes fit together, like parts of a well-engineered machine. That's what D&D is. To be successful, you should aim to make your class just as well crafted, an expansion and upgrade to the machine. Failed classes just come across as unnecessary brickabrack, no matter how balanced or complete they may be.

It All Starts With a Concept

Put pen to paper. Do not stop until you run out of either.

Now I know the word "fluff" has a fair bit of negative connotation in the gaming community. Generally, the word refers to all the aesthetics, cosmetics, or window-trimming of the game. It is generally assumed to be meaningless drivel, which most people skip over on their way to the numbers. But unknown to these short-sighted game-focused individuals, fluff is actually the key to designing a class! (Or really any game content) Think of it this way:

D&D, on the mechanical level, can best be described as incredibly low-stakes gambling. The only thing you really stand to lose or gain is playing time. In order for people to want to play, that play time needs to be worth taking a risk for. And what makes that play time worthwhile? Fluff. Without it, we don't really have much to care about! Without the fluff, we don't have knights in shining armor, or dragons, or ancient curses; all we'd have is a pile of numbers and arbitrary tactical rulings.

Fluff can make or break a piece of content through its immersive qualities. In general, the "crunch" of the game is intended to represent the fluff. If your mechanics don't match what they're trying to represent, people lose immersion, and the material falls flat. So knowing your fluff, getting it right, describing it clearly, and adhering to it consistently, is essential to producing a successful class.

While you're working on all of the mechanical stuff, keep your original idea in mind. Figure out what you are trying to represent conceptually, and stick to it. Even if you come up with a cool new idea while you work on the class, if it doesn't support your idea, save it for another creation. For each feature you add, consider, "how does this represent my idea?" An easy way to keep yourself on track is to describe your idea as a point-form list of themes. For example:

Chivalrous Knight.

  • Armored combat with a shield
  • Cavalry
  • Nobility
  • Chivalrous romance
  • Medieval games

Hit Man.

  • Classy and cool
  • Stealth
  • Intellectual
  • Emotionless and remorseless
  • Secret enemy

Then, whenever you add or tweak a feature, refer to your themes and consider whether it supports them. Also, remember to occasionally look at your themes and consider how well each one is represented by the class features. Don't be afraid to tweak your idea as you go, if you find yourself being painted into a corner.


In party dynamics, a character's role is the practical function they serve in the party. This mainly regards combat roles, but in 5e, every character should be able to at least function in a couple of non-combat roles as well.

Combat Roles.

  • Tanks force enemies to waste their time by absorbing damage that would otherwise be dealt with other members of the party. AC, HP, and HD are their friends, with constitution as a primary ability score. May earn damage reduction or temporary HP. Fighters, Barbarians, and Paladins fill this slot well, as does a Hill Dwarf Dragon Sorcerer. Barbarians can do this if they optimize for unarmored AC, and monks can do the same to a lesser extent.
  • DPS stands for Damage Per Second. These characters focus on raw damage output and are typically lead fighters or spellcasters. Maybe ranged. Melee DPS characters are often designed such that they can double as a tank. Dual-wielding fighters do this very well, but barbarians, rogues, and bards can do it, as can casters with the right spells under their belts. Monks, using their martial arts damage, can also put out some impressive numbers.
  • Healer. Clerics and Paladins are the best examples of these, though Druids can do it a bit too. Their primary job is to keep everyone alive by healing their allies, buffing their allies, and by debuffing their enemies. A common ability among healers is the repulse effect, which makes a bunch of enemies, usually of a specific type, flee for a short duration, effectively removing them from the fight.
  • Support roles are characters who generate field effects, buff allies, debuff enemies, take advantage of combat circumstances or amplify the effects of allies. They serve tactical purposes, rather than direct combat interaction. All spellcasters can satisfy this role with ease.

Exploration Roles.

  • Scouts sneak ahead of the party, identifying and removing traps, opening barriers, and preparing ambushes for upcoming threats. They may also trail behind, watching for ambushes and covering tracks. Rogues and Rangers all the way here.
  • Guides know certain environments, reducing the number of threats the party may stumble into, reducing the chance of the party getting lost, speeding up travel, finding food and water, and locating safe places to rest. Rangers and Druids fit the bill.
  • Eggheads are characters with extensive knowledge and access to information sources. The Acolyte, Sage, and Criminal are perfect examples. The druid and ranger can also often relate a great deal of circumstantial information. Another recurring trait of an egghead role is their ability to effectively search out and obtain information where it would not otherwise be available, sometimes through direct investigation.
  • Skill Monkeys are characters who focus on having a bonus available for as many checks and saves as possible and trying to get those bonuses as high as possible.

Socialization Roles.

  • Diplomats are typically the party leader. Their skills get the party work, catch people who are trying to lead you into a trap, get bargains at stores, ask for better pay, fib your way into and out of trouble, convince people to see things your way, etc. Bards are best in this role, but sorcerers are often well adapted to it as well. Certain backgrounds, (Noble, Entertainer, Guild Artisan) can make any character better suited to this.
  • The deceiver is a character who can get around the rules for the party. They spy on important figures, find secret passageways through sewers, procure shady transportation and lodging, and more. Many deceiver functions pair well with scout functions. Case in point, the rogue with the criminal background is an exquisite expression of this role.

Creating a <!-Class Name->

Quick Build

You can make a <!-class name-> quickly by following these suggestions. First, <!-Ability score-> should be your highest ability score, followed by <!-Ability score->. Second, choose the <!-background name-> background. Third, choose <!-elaborate on equipment choices->


The name entered here is used in the place of the full page name in the 5e Class list.

The summary line is also used on the 5e Class list and is pretty much your only marketing tool to get people interested in your class. This is your chance to communicate directly to another user just what your class is, and try to convince them it is cool enough to click on. Make an impression! Not having a summary tends to make your content disappear in the crowd, and long summaries break the table and look ugly.


The Hit Dice used by the core classes are the d6, d8, d10, and d12. The wizard and sorcerer- the only two classes whose only purpose is spellcasting- are also the only classes with a d6 HD. This is important to note, as these two classes are traditionally the frailest. The implication is that a d6 should be as low as you go in 5th edition. Bard, cleric, druid, monk, rogue, and warlock all use the d8. This makes the d8 the most common HD. Also note that these classes range balance and role across the board, from spellcasters to mundanes, from combat units to support team, from utility characters to specialists. This implies that a d8 is actually the standard HD, and that adjustment from there is considered to be a class feature. The fighter, the paladin, and the ranger each use the d10. Note that two of these characters are typically front-line combatants, while the ranger is intended to be a hardcore survivalist so the added HP goes well with that theme. The barbarian is the only class that uses the d12. The barbarian is intended to not wear any armor, so unless you build for UAC (we'll get to what that is in a bit) they'll get hit a lot. Also, they're a front-line combatant, like the fighter, and a rugged survivalist to boot. That d12 is not arbitrary- it serves both combat and exploration purposes, as well as being justified by the class' mechanics. So, although the d12 is the maximum limit by precedent, it takes quite a bit to justify it.

OK, so outside of all that, what have we got to work with here?

  • d4. In the days of ye olde D&D, classes having a d4 hit dice wasn't an uncommon sight. But 5th edition has grown past that, and both hp and damage numbers have risen from what they were before: the frail thief and magic-user of the Basic set only had a d4 as their hit die, but the highest hit die back then was the Fighting-man, who only had a d8. But even back then, classes with a d4 as their hit die knew to stay far, faaar away from combat at low levels, which is not normally possible for 5e classes. In fifth edition, a class with a d4 hit die could easily have a third of the hp of a barbarian at 1st level, it's also much lower than most commoners, something which your design will have to justify. If a wizard, who spends all his waking hours studying, and a sorceror, who is an innately unstable being, both have d6 hit dice, what could your class have done to justify being so weak? Not considering the actual problem with such a low hp character: the lowest weapon damage in fifth edition is 1d4, this means that unless you have a positive constitution and full hp, even a dagger can instantly put you on death's door. But in fifth edition, most monsters attack with bigger weapons, and have bonuses to their damage and attack multiple times in groups. A class with a d4 as its hit die has to be fundamentally designed around its incredibly low survivability, otherwise it won't live to see level 2. Such classes either need to be designed to avoid combat encounters fully(which is a very tricky and dangerous design space), or have alternate features that give them survivability, like termporary hp, summoning creatures to protect them, being ethereal/very hard to hit, etc. And they should recieve these features Fast (level 1 if possible!). In short, unless you dedicate to mechanically and narratively design the class around its very low survivability, avoid giving it a d4.
  • HD Modifiers. Another unexplored option are flat modifiers to the HD. Such modifiers should probably be positive, as typically only bonuses apply to rolling for HP. If there were a penalty to HD, you would need to make a feature specifying what the minimum HP gained is, most likely 1. Flat modifiers to HD only matter at two general points in gameplay: first tier, when every point of HP is essential, and 4th tier, where enough levels have accrued that so much as a +1 modifier can be worth as much as 13HP or more- more than the maximum any raw HD could ever provide you from leveling up. Now, the maximum HP impact is not as imbalanced as it may seem. There is a subrace which basically provides you with +1 max HP /level. The key here is that that bonus does not apply to HP regain during short rests. A flat modifier to HD would have a significant impact on the recovery economy of 5th edition. Messing with that economy is generally a no-no, so you'd have to find a pretty dang good reason to justify it mechanically.
  • Half-Dice. Half dice are what you get when you roll an even-numbered die, divide the results in half, and round down to the nearest whole. 1d4=1d2, 1d6=1d3, 1d8=1d4, 1d10=1d5, etc. The only useful ones are the d3 and d5. The d3 is a terrible idea for HD, for all the same reasons as the d4. The d5, however, could be used to make a particularly frail spellcaster. Coincidentally, there is such a thing as a real d5, but physicists and statisticians are still arguing about how even its probability curve really is. Other alternatives to the dividing step are renumbered full dice, and barrel dice, although both of these are much harder to find and purchase in a store.
  • HD Pools. Another unexplored mechanic is multiple HD per level. For example, 2d4 HD. As for what that means or how it works, well, the rules don't really cover that, so you'd have to write a class feature to justify it. Right off the bat, I'd recommend that the twinned dice need to be rolled together when recovering. So your HD at level 3 would read like, "3×2d4" or something to that effect. 2d4 would make a significantly tougher version of the standard d8 HD, but without necessarily increasing HP if the no-rolling method is used. If your players DO roll for HP however, the twinned dice will result in a higher HP maximum. This is probably the main reason the developers didn't use this mechanic or make rules to support it- variations in table rulings will result in significant balance discrepancies from one character to the next, and make certain characters incompatible with certain tables. Also, the d4 is pretty much the only die you could do this for, as the 2d6 would make a super-beefy barbarian, essentially exceeding their intended limit.

Class Features

As a you gain the following class features.

Hit Points

Set parameter to Hit Dice per level, e.g. 1d10


Saving Throws:


You start with the following equipment, in addition to the equipment granted by your background:

Table: The

Level Proficiency
1st +2
2nd +2 <!-Class Feature2->
3rd +2
4th +2 Ability Score Improvement
5th +3
6th +3
7th +3
8th +3 Ability Score Improvement
9th +4
10th +4
11th +4
12th +4 Ability Score Improvement
13th +5
14th +5
15th +5
16th +5 Ability Score Improvement
17th +6
18th +6
19th +6 Ability Score Improvement
20th +6

<!-Class Feature->

<!-Class feature game rule information->

<!-Use semi-colons for subheaders->

Ability Score Increase

When you reach 4th level, and again at 8th, 12th, 16th and 19th level, you can increase one ability score of your choice by 2, or you can increase two ability scores of your choice by 1. As normal, you can't increase an ability score above 20 using this feature.

<!-Class Feature->

<!-Class feature game rule information->


LVL 1 - - - - 2 - - - - - 2 1 - 5 3
LVL 2 - - - - 1 2 - - - - - - 2 5 3
LVL 3 3-4 1 1 2 - - 1 1 1 2 - - - 12-13 8
LVL 4 - - - - - - - - - - - - - 0 0
LVL 5 1 - - - - - - - - - - - - 1 1
LVL 6 - 1 - 1 1 1 1 - - - 1 1 1 8 8
LVL 7 - - 1 - - - - 1 1 - - - - 3 3
LVL 8 - - - - 1 - - - - - - - - 1 1
LVL 9 1 - - - - - - - - 1 - - - 2 2
LVL 10 - 1 1 - - 1 - - - - - 1 1 5 5
LVL 11 - - - - - - 1 - 1 - - - - 2 2
LVL 12 - - - - - - - - - - - - - 0 0
LVL 13 - - - - - - - - - 1 - - - 1 1
LVL 14 - 1 - 1 - 1 - - - - 1 1 1 6 6
LVL 15 1 - 1 - - - - 1 1 - - - - 4 4
LVL 16 - - - - - - - - - - - - - 0 0
LVL 17 - - - - 1 - 1 - - 1 - - - 3 0
LVL 18 - - 1 - - - - - - - 1 - - 2 2
LVL 19 - - - - - - - - - - - - - 0 0
LVL 20 - - - - - - - 1 - - - - - 1 1
BENEFITS 6-7 4 5 4 6 5 4 4 4 5 5 4 5
LEVELS WITH BENEFITS 4 4 5 3 5 4 4 4 4 4 4 4 4

These are guidelines, not rules - for example, most Druid subclasses receive 2 subclass features at level 2, but Circle of Dreams Druids only receive 1. Furthermore, it's common for spellcaster subclasses to include additional spells known or prepared that apply at levels after getting the feature that grants them. Number of features is fairly arbitrary, because a given feature may be particularly weak to make up for another being particularly strong. What typically matters far more than number of features is how the feature scales: flat (a benefit that remains the same forever), scaling with ability modifier, scaling with proficiency bonus, or scaling with class level. Multiple scaling options at once are possible, but exactly how an early subclass feature scales typically informs how valuable it is for multiclassing.

In fact, scaling can make the table above deeply misleading - for example, the L8 subclass benefit clerics get has about a 50% chance of being one that grows in power at L14, but because it's not a new feature, it's not in the table. The other half of the time, the benefit scales with wisdom modifier, not level, which by L8 generally means it doesn't scale at all.

In terms of a feature you want to base on rests, remember, the DMG states on page 84 that WOTC intends for there to be about 6-8 fights per long rest and 2 short rests per long rest, spaced out evenly, which maths out to 2, 2 1/3, or 2 2/3 fights per short rest, averaging out to 2 1/3 fights per short rest. For a feature that is only useful once during a fight, remember that the DMG also, at a core level, assumes all fights are 3 rounds long, so something useful once per fight should be treated as applying, at 1/3 power, to the whole fight - or you can treat a power useful during an entire fight as functioning at triple power but only once. Either method will let you compare two different abilities like that, such as a Barbarian's rage (X per long rest, whole fight) to a Fighter's Action Surge (1 per short rest, 1/fight) - if the usefulness of rage is X and action surge is Y, to compare apples to apples, look at 2X/7 (its usefulness per fight times usefulness in a fight) versus Y/3/(7/3) = Y/7 - for rage and action surge to be equal in utility, 2X/7 = Y/7 so 2X = Y so X = Y/2; action surge needs to be twice as useful as rage.

Spell List

You know all of the spells on the basic spell list and additional spells based on your subclass.

1st Level

<!-1st level spell list->

2nd Level

<!-2nd level spell list->

3rd Level

<!-3rd level spell list->

4th Level

<!-4th level spell list->

5th Level

<!-5th level spell list->


Prerequisites. To qualify for multiclassing into the <!-class name-> class, you must meet these prerequisites:

Proficiencies. When you multiclass into the <!-class name-> class, you gain the following proficiencies:


Balancing a homebrew class is arguably one of the most difficult aspects of homebrew. There are many facets when it comes to balancing a class, but perhaps the most vital and common is damage output. The easiest and most frequent way a class can be seen as "overpowered" is by outputting too much damage compared to everyone else in the party.

By contrast, it's almost never a problem if someone does too little damage; buffing, healing, distracting, nerfing, and helping are always useful. Even a meat shield is helpful, and in 5e any player-character can do that by taking the Dodge action while having a decent AC.

This guideline provides you with recommended damage values at the five most vital points of a class: 1st, 5th, 11th, 17th, and 20th levels. Each of these levels represents a new "tier" of play, where a typical class will ascend in its capabilities most dramatically.

As noted, minimum damage is virtually never a problem. Instead of listing a minimum to maximum range, these tables only list the absolute most damage any class should ever be able to do in a single round, assuming all dice rolled yield their average result. If a homebrew class deals the damage listed, it is potentially doing something wrong. If the class is actually doing more damage than is listed on this table, then it is definitely doing something wrong.

Table - Standard Attack

This lists the maximum for what should be achieved without expending valuable resources. No high-level spell slots, no once-per-day actions, no particular luck, and no exceptional tactics apply.

Level Averaged
PHB Source
1st 13 A fighter with the Two-Weapon Fighting Style with 16 Dexterity makes one shortsword attack as an action; then makes another shortsword attack as a bonus action for a combined 2d6 + 6 (13) piercing.
5th 24 A fighter or paladin with 18 Strength and the Great Weapon Fighting Style makes two greatsword attacks as part of the Attack action, for a total of 4d6+8 (22) slashing damage; the effect of Great Weapon Fighting moves an average result of 2d6 from 7 to 81/3. Considering this, the actual average damage is 24⅔.
11th 40 A fighter with 20 Strength and the Great Weapon Fighting Style makes three greatsword attacks for a total of 6d6 + 15 (36) slashing; the effect of Great Weapon Fighting moves an average result of 2d6 from 7 to 81/3. Considering this, the actual average damage is 40 slashing.
17th 48 A wizard of the Evocation School with 20 Intelligence casts scorching ray with a 3rd level level spell slot (of which it has plenty), and uses its Empowered Evocation feature to add its Intelligence modifier to each of the four different damage rolls for a combined total of 8d6+20 (48) fire.
20th 53 A fighter with 20 Strength and the Great Weapon Fighting Style uses its Attack action to make four greatsword attacks for a total of 8d6 + 20 (48) slashing; the effect of Great Weapon Fighting moves an average result from 2d6 from 7 to 81/3. Considering this, the actual average damage is 531/3 slashing.
Table - All-Out Attack

This damage should only be possible by expending immense resources, such as your highest-level spell slot(s) or a once-per-day feature; it should not be something you can do for more than one or two turns in the entire day.

Level Averaged
PHB Source
1st 16 A cleric casts inflict wounds for 3d10 (160.5 ½).
5th 54 A paladin with 18 Strength uses its action to make two shortsword attacks for 2d6+8 (15) piercing, and its bonus action to make another shortsword attack 1d6 (30.5 ½); it uses both of its 2nd level spell slots and one 1st level spell slot to use divine smite on each of these three attacks for 8d8 (36) radiant. The sum of these averages is 540.5 ½.
11th 80 See the 11th level fighter from Table - Standard Attack. It simply uses its once-between-rests Action Surge to perform that Attack action twice.
17th 145 A wizard of the Evocation School with 20 Intelligence casts meteor swarm, and uses its Empowered Evocation to add its Intelligence modifier to the damage roll: 40d6 + 5 (145).
20th 145 A wizard of the Evocation School with 20 Intelligence casts meteor swarm, and uses its Empowered Evocation to add its Intelligence modifier to the damage roll: 40d6 + 5 (145).
Table Design Notes

The tables:

  • Only consider damage done to one target in one round with no preparation.
  • Assume every attack hits and every saving throw fails.
  • Omit any gains from feats, multiclassing, magic items, or other party members' buffs.
  • Do not consider particularly lucky circumstances out of the player's control, such as critical hits, opportunity attacks, hellish rebuke, and natural 1s.

Ideally, anyone designing a class should place more weight on "unavoidable damage" and "area of effect damage" like that from fireball or a dragonborn's breath weapon, even though these tables do not.

Focused Optimization

Choose an interesting aspect about your class and try to build a character who focuses on and optimizes the effects of it, then give the character a quick whirl through some example encounters which would use that detail. Try to challenge yourself, give yourself a feel of how effective the character is (or isn't) because of the optimization. You can do this at any point during class design, just to check if something is working the way you think it should. One of the most common mistakes is people making a trait or feature which works fine at first but has math which causes it to quickly spiral out of control. In general, here is a list of items you may wish to target for optimization. For each item you think may be problematic, simply build a character focused on nothing else.

  • HP (A level 1 character should not be able to start with more than 17 maximum HP. That's 1d12+5CON)
  • AC (The maximum, non-magical AC attainable by a level 20 character is 24. That's UAC10+5+7+2)
  • Speed (The fastest core character on land is a wood elf monk who has 5 levels in barbarian and 15 levels in monk at 70ft- he only goes 35ft at 1st level though, and that's just because of the subrace, not the class.)
  • Initiative (The maximum initiative attainable by a level 20 character is +13. That's assuming a Dexterity score of 20 (+5), having taken the Alert Feat (+5), and also attained the Jack of all Trades class feature (+3). The Jack of All Trades feature adds this bonus to Initiative rolls as Initiative is considered an ability check.)
  • Raw damage output per turn (Extra attacks and bonus attacks are key here, and this is closely tied to gear selection. A maxed out fighter can make 8 attacks and 1 bonus attack in one turn.)

The other form of optimization test is to do the complete opposite. Play against type, put your weakest scores in your most important abilities, choose overlapping proficiencies as much as possible, and equip yourself wrong. Then see if the character still functions. If the character is still just as capable, (or more capable) as a different character who has been built appropriately, then you may have a balance issue. In such a case, it is likely that you have static class features which are simply all-around good no matter what. This may not necessarily be a problem- after all, your class is supposed to make you intrinsically good at stuff- but if these features are so good that you are rarely if ever impacted by your shortcomings, you have a problem. A class which "covers all the bases" isn't great for play because characters need to have some weak points that they are affected by. These weak points give the DM opportunities to threaten and challenge PCs, and it gives PCs a place to rely upon and support one another. If you have features which are particularly good, even when you suck, it is very likely that these features are amplified to become totally broken when paired with synergizing abilities, proficiencies, and gear.

Vs. Mode

Make two characters, one of your class, and one of the class they are most similar to, or most likely to be compared to. Next, have them compete head-to-head in a variety of challenges, including 1v1 fights, horde slogs, and specialty skill contests. If one of the two consistently wins over the other in some area, with the tables only being turned by bad rolls or tactics alone, then that indicates a balance issue.

One-Shot Solo Campaigns

The best way to get the fullest grasp of the true nature of a class is to actually play it. For this, you want to use the class to build a whole, real character, and then play it in a very short adventure. You can play it by yourself, being devil's advocate as your own DM, but it works better if you can find someone else to sit behind the screen. As for the adventures, you can, of course, build your own little one-shot campaigns. You might even build standardized adventures you use on yourself over and over again! A great way to get reliably balanced adventures which match what other DMs and players are likely to create or encounter is to download some of the Adventurers League Expeditions adventures, which are now all available in the Dungeon Master's Guild, though also just as easily pirated. One good option if you are stuck testing by yourself is to write the whole campaign out ahead of time without a clear path to progress, pre-roll all of the encounter's rolls and use the numbers as they come up, this way no matter whom you run through the test as characters the overall challenge is exactly the same. You can even pre-roll the player rolls if you so wish to streamline the test, or you can make it so every roll is taken 10, this way you reduce all the variance and actually make the run become extremely fast.

The 5e Same Game Test

The Same Game Test is a tool for calculating the balance of a character class in a d20 based RPG, such as D&D. Note that this tool is only relevant in games which attempt to quantify the "challenge" or "encounter level" of everything in the game. It evolved out of homebrew design for 3.X edition. Like most things that evolved from the hobby mire, there's a lot of politics to wade through to learn anything about it. Somewhere in 2009, the SGT was used extensively by some guy named Frank Trollman, and everyone had a bone to pick with him for three years after that. (Mostly because he was very vocal and said some very rude things very loudly) Because of the trans-internet fireball he initiated, I can't find anything on the history, development, or spread of the SGT before 2009. There is some implication that Trollman either invented the SGT or played an important role in its development, but of course, everyone's too busy blathering about gossip and flame wars to care about that. OK, so, what is the SGT? The idea works like this:

In a d20 RPG, each challenge, such as a monster or trap, is typically given a levelled rating. The terminology for this varies from game to game, and even edition to edition. In D&D3.5, the terminology was Challenge Level (CL) or Encounter Level (EL). Theoretically, a challenge of a given level should be just as difficult as fighting a player character (PC) of that level. For example, if a Troll has a CL of 1, then it should be just as strong as a level 1 PC, and therefore have a theoretical 50% chance of winning or losing in a fight against one. Likewise, two 1/2 level monsters (an EL of 1) should only be slightly more difficult due to the unbalanced action economy and field control.

Sounds simple, right? If you think so, you've never played 3.5 edition! It is the single most complex version of D&D ever made, and one of the most elaborate combat systems ever written.

The challenge levels assigned to monsters were not calculated, they were intuitively applied through playtesting. So, for example, after running 50 test fights against 4 random characters at various party levels, the developers would check how easily the monster was killed at given levels, then give it a rating based on its percentages. That's right, a system designed to compare a single monster to a single character is based on playtests against 4 characters. To complicate matters, the core classes were not built with EL in mind. EL was designed to compare groups of monsters to the party level, not the other way around. As a result, if you were to put the core classes through a standardized test, they would all perform differently. In other words, the different classes are not playing the same game.

Now, it's been a few years. The latest edition is out, and balance is the name of the game in every regard. There were attempts to make an SGT for 4th edition, but publication stopped before any headway was made. I want to try and build an SGT for 5th edition. In order to do that, I need to clarify a few things.

  • Firstly, there is no standardized SGT. Normally, it consists of making a level 10 character (although some people also tested at level 5, 7, and/or 15) of the class to be tested, then running it through a gauntlet of 9 encounters (although this could vary anywhere between 5 and 28) of the same level as it, then tallying up its wins/losses. Typically, the encounters were chosen uniquely for the class, in order to make sure it is being tested in a way that is relevant. For example, running a spellcaster through a gauntlet of traps and puzzles would not test the spellcaster's actual capabilities, as the focus would be in the wrong area.
I believe there should be 10 encounters. A nice round number that is easily converted into percentages. I believe there should be a series of sub-SGTs which compose the full SGT, now referred to as the Same Game Exam (SGE).
  1. The Entry Level Test (ELT) is performed at level 1, and gauges how the class handles fresh out of the gate.
  2. The Early Game Test (EGT) is performed at level 5, and shows how the class handles with all types of features, including ability score increases and subclass features.
  3. The Midgame Test (MGT) is performed at level 10, and is the theoretical standard for the class.
  4. The Late Game Test (LGT) is performed at level 15, and shows how the class handles with nearly all of its features combined.
  5. Finally, the Endgame Test (EGT) shows how the max-level class works out.
This is then presented as a scorecard showing a series of 5 percentages, stating the class' successfulness at each level. This shows clearly how the class' power level grows, and at what rate. Such a scorecard can also be compared to that of other classes to see if it balances against them.
Each of these tests will have 10 equal-level encounters testing the class in 10 different ways. Generally, there will be 3 exploration encounters, 3 social encounters, and 4 combat encounters. This is a dramatic deviation from the 3.5e tests, which dealt almost exclusively with combat. This is because 5th edition places mechanical importance in all 3 areas. Certain classes are more capable in one regard or another. Bards, for example, excel at social encounters, but belly flop in combat and Rangers are great at exploration but iffy in a fight. 5th edition is a very broad system, designed to accommodate a variety of player agendas simultaneously, while still playing to the strengths and themes of D&D.
  • Second, past scoring has varied strongly depending on who designed and ran the test. Some people would run each fight once and just tally up the raw binary data. Others would run each encounter multiple times and the score is based on the average, then add up the averages. Others still would score the encounters based on the "feeling" of how probable success was, assigning a numeric value to each grade, then add up those results. I plan to run each encounter three times, find the average, and adjust it by up to 10% based on how lucky or unlucky I felt I was. To do the full SGE with all five stages would mean running 150 encounters by the end of the ET. In addition, in order to cover all of the subclasses, the last 40 encounters would need to be repeated with each subclass.
  • Third, in the past, the way the information was interpreted also varied significantly depending on the tester. Some people just aimed for a predetermined success % and tweaked the design until it met what they wanted. Others had complex categorization systems, comparing the tested class to the score ranges of core classes. (IE, in 3.5, Monk class was the lowest, Fighter class was underpowered, the Rogue class was average, and Wizard class was overpowered.)
  • Fourth, I do not know, for certain, what the balance is for 5th edition. In 3.5, a level 10 character should have a 50% chance of winning a level 10 encounter. I will need to do a fair degree of testing to find out the actual median balance of the game. I'm going to start with the DMG's statement of what the numbers mean, put it to the test by running monsters of a given level against each other, then revise their words to reflect the truth. (Hopefully, no revision will be necessary.)
  • Finally, my purpose here is not to test the capabilities of the core classes, but to create a standard or metric against which any class can be measured and compared to one another for the purposes of overall balancing during the development process. I don't give a darn about min-maxing, metagaming, power gaming, or munchkinism.

In order for these tests to be relevant, there needs to be a standardized character that the class is being applied to. Think of it like a rocket-sledge test: We need a regular sledge to strap our rockets to before we go launching them on a test run. This character needs to be as absolutely generic and non-preferential as possible, such that changing the character's class will be its only significant features, mechanically. The character should have true neutral alignment in all cases. The best choice for a race is Human, as all they get is an even +1 to all ability scores. Next, the character should have even scores in all abilities. In order to ensure that these numbers are still realistic, PHB point-buy is used, giving the character a score of 9 in all abilities after racial adjustments. (That means all ability checks are at -1 modifier) We must use a background which does not significantly benefit any one type of class over another. I believe that background would be the Guild Artisan. Its feature only affects a downtime activity, and so has no in-play impact, implicit or otherwise. Its equipment layout has very little implicit power in the context of isolated arbitrary encounters, and none of its items are explicit improvised weapons.

We also need to standardize available gear based on level. The type of equipment available to a level 1 character, compared to that of a level 20 character, is one of the most significant factors in power gain.

  • During the ELT, the character is only allowed the gear provided by their build.
  • In the EGT, the character is assumed to be geared in as optimized a fashion as can be imagined, up to a total value of all their starting gear, using published mundane equipment.
  • For the MGT, the wealth limit is revoked.
  • For the LGT and EGT, the character is assumed to be geared in as optimized a fashion as can be imagined using published magical equipment as well.
  • For the EGT, the character is given 1 Epic Boon, chosen based on whatever will be most beneficial to the class in the test.

To ensure that gear is not the determining factor in these tests, you can never gear up a character beyond their encumbrance limit. If their starting gear exceeds this limit, items must be omitted from the test.

Now we have a standardized character, we need to get down to the gritty details of building a standardized test. Beginning at the Monster Manual, I am informed that the terminology for 5th edition is Challenge Rating (CR). Once again, Challenge has been assigned to a party of 4 characters. Mathematically, this seems to suggest that a single character facing a monster of the same CR should have ~25-30% chance of success. That's a bit steeper than previous editions. (It also means that CR is not interchangeable with Level; two characters of the same level should have ~50% chance of besting one another in some way.) So a CR1/4 monster should be an equal match for a single 1st level character. Each test consists of 10 encounters divided into three categories. In each category, there must be a standardized challenge of a given type. Check and save DCs should be based on the AC range of enemies at the given level.

1. An obstacle course style gauntlet of environmental hazards, traps, and challenges such as locked doors. These should consist of 3 obstacles. Must contain one ability check, one skill check, and one save bare minimum. The encounter must contain only challenges of mechanical ability- not player strategy or logic. The encounter must have multiple methods of passing each obstacle.
2. Survival in a given environment for the duration of travel to a destination. A map is provided, along with a clock and calendar. Survival includes covering basic needs, random encounters, and weather hazards. Each test is given rules for determining if the character becomes lost. Included is a standardized battlefield for each terrain type. Random encounters should be below the test level, as the character will likely face several in a row without rest.
3. Acquiring information from within the confines of a specified environment, using only resources provided by that environment and your class.
1. Extracting correct information from an NPC.
2. Negotiating a bargain, deal, or transaction.
3. Getting out of the blame and avoiding damning information reveals surrounding a bad situation which you may or may not have been framed for.
1. The single strongest monster or NPC of this level's CR.
2. A duo of enemies of half this level's CR. (Both melee and ranged)
3. A hoard of enemies equal to this level's CR.
4. A (randomly selected out of 5) optimized PC of this level.
  • You may also wish to add a category for simulation/downtime/hybrid type acts, however these are trickier to set up, but scoring can often be done with the money value or with number of successes over a certain period.

Home of user-generated,
homebrew pages!