D&D 5e Class & Subclass Design Questions for Dungeon Masters

Balance a classless feature system by pricing what a feature actually does, not by giving every class feature the same cost. Two features that occupy the same amount of text can have radically different value once frequency, action cost, scaling, resource generation, and combinations with other features are considered.

I would sort options into rough power bands before players can select them. Ask whether a feature is passive or consumes an action, Bonus Action, or Reaction; whether it works once or repeatedly; whether it creates another resource; whether its normal class receives it alongside weaker features; and whether it assumes prerequisites that your classless system has removed. Features that improve the same action can also multiply one another, so test combinations rather than evaluating every option in isolation.

Then build three deliberately troublesome test characters: a specialist stacking one tactic, a generalist collecting broadly efficient features, and a character trying to generate extra actions or renewable resources. Those reveal problems much faster than comparing average builds.

For NPCs, I would not use the full player construction system unless the players need to understand those NPCs as peers. Give an important NPC one defining feature, one defensive tool, and one active decision that communicates their identity. A classless campaign can offer enormous player freedom without requiring every student, rival, and teacher to become another optimization exercise for the DM.

Personality traits can work as class-like mechanics if they create decisions without defining one mandatory way to portray a personality. They are especially useful for new players when the character sheet turns an abstract identity into clear situations that invite action.

The danger is rewarding simple obedience. If a character receives a bonus whenever they behave "lawfully," "bravely," or "loyally," the mechanic can quickly teach players that every value has one correct performance. That narrows characterization instead of supporting it.

Give each value several game-facing pressures instead. It might provide an advantage when the character acts through that value, create a complication when the value is challenged, and include a way for the value to evolve. Loyalty becomes interesting when loyalty to a friend conflicts with loyalty to a cause. Courage becomes interesting when retreat would protect somebody else. The mechanic should expose those tensions rather than automatically resolving them.

A useful playtest is to give the same value to two different characters. If both players still make recognizably different choices, the trait is functioning as a creative boundary. If they repeatedly behave in the same prescribed way because that is where the bonuses are, loosen the conditions or reward confronting the value rather than merely demonstrating it.

There is no defensible universal “winning six” unless you first fix the conditions of the comparison. Class-versus-class arena fights are extremely sensitive to level, exact class versions, magic items, rest state, starting distance, terrain, initiative, and the victory condition. Change any of those and a lineup that looked dominant can become mediocre.

For a fair combat test, give both parties the same level, equivalent equipment, the same information, and a defined battlefield. Then run several scenarios rather than one duel: open ground, cramped terrain, an objective that must be defended, and a fight after resources have already been spent. That exposes whether a party is broadly strong or merely optimized for one setup.

An adventure comparison needs different criteria. In something like an urban investigation, measure coverage: social influence, information gathering, exploration tools, mobility, healing, control, sustained resources, and the ability to recover from bad plans. A party that wins a six-on-six fight may still be worse at completing the adventure.

Without the complete third-party class rules, a specific official lineup would be guesswork; the useful comparison is between assumptions and capabilities, not labels.

Define the feature around observable conduct, not the DM’s private verdict on whether someone is morally guilty. A powerful “Judgment” ability works better when players can predict what qualifies, what information is required, and what each possible result does at the table.

Write triggers in concrete terms: the character witnessed the act, the target knowingly harmed a creature, a named law or oath was broken, or credible evidence has been presented. If uncertainty is part of the fantasy, make uncertainty produce information rather than a dead result. The feature might mark a suspect, reveal whether the accusation matches the target’s memory, or grant a weaker temporary effect.

Then separate sentence from punishment. The player can choose among a small number of consequences suited to different situations instead of waiting for the DM to invent a moral ruling and reward.

Finally, limit the feature through action economy, duration, uses, or meaningful counterplay. The interesting decision should be when and how to invoke judgment, not whether the DM agrees with the player’s ethics.

Yes, reserving a spell resource can be a workable maintenance cost, but decide whether “reserved” means temporarily unavailable or actually spent, because those produce very different abilities.

If the slot or points return immediately when the effect ends, the cost is opportunity: the character gives up access to part of their casting power while the zombie, summon, ward, or other effect remains active. If the resource returns only through normal rest recovery, reservation is much closer to spending it up front. Write that recovery rule explicitly so the player cannot dismiss an effect and immediately recycle the same resource into another one unless that is intended.

Let the reserved resource do most of the balancing work before adding more restrictions. Requiring concentration, limiting the creature count, imposing a separate action tax, and locking spell resources simultaneously can make the feature feel expensive without creating better decisions.

Scale stronger effects by the amount reserved rather than by many unrelated penalties. One weak undead might reserve a small amount; a tougher corpse could reserve more. Then test the feature against what the character gives up in actual play. In a small party, a companion also adds actions, so evaluate its action economy as carefully as its damage or hit points.

Give the limited spellcasting a narrow tactical job instead of trying to reproduce a full spellcaster in miniature. The subclass should have a recognizable reason to use magic while still making ordinary Fighter actions important.

Start by defining that job in one sentence. Perhaps the magic creates brief battlefield control, adds elemental pressure to weapon attacks, or provides a small number of unusual utility options. That sentence becomes a filter: a new spell or feature belongs only if it strengthens that identity rather than adding unrelated versatility.

Then test frequency and opportunity cost before worrying about small damage differences. Ask how often the magical option appears, what the Fighter gives up to use it, and whether there is ever a reason not to choose it. An ability that adds damage almost automatically can become more influential than its die size suggests, especially if it also imposes penalties, movement effects, or other control.

Watch action economy as well. Bonus Actions, reactions, and effects attached to normal attacks can feel inexpensive because they preserve the Fighter's main routine. Several individually modest features can therefore combine into a turn with too many benefits and too few decisions.

During playtesting, record which options are used spontaneously, which are saved for one dominant combination, and which require reminders. Do not immediately strengthen an ignored option; first ask whether it has a distinct purpose.

The subclass should ultimately create a repeated choice between martial reliability and a limited magical advantage. If the magic is always automatic, or if it provides enough breadth to replace dedicated spellcasting, that distinction has disappeared.

Not by the standard Sorcerer rules. The class uses Charisma as its spellcasting ability, so changing it to Wisdom is a house rule rather than a character-building option the player can select on their own.

As a house rule, the cleanest version is a full substitution rather than a partial one. Use Wisdom anywhere the class feature being changed would normally use Charisma for spell attacks, spell save DCs, or other Sorcerer features that explicitly depend on that ability. Write the change down so the table does not have to renegotiate it feature by feature.

Then check the side effects. Wisdom supports commonly used perception and insight skills, while Charisma supports a different set of social skills, so the swap changes more than spell flavor. Multiclassing can also create combinations the original ability requirement was meant to constrain. If multiclassing is allowed, decide its prerequisites and interactions before approving the variant.

Do not try to “balance” the swap by randomly weakening other class features. First decide what fantasy the Wisdom-based Sorcerer is meant to support, make the smallest consistent change, and review it after several sessions. If it produces a problem, adjust the specific interaction causing it rather than rebuilding the whole class.

Give the Priest a gameplay loop the Cleric does not have rather than defining it mainly by worse armor, fewer weapons, and lower Hit Points. Removing defenses creates a fragile Cleric; adding a distinct way to support the party creates a new class.

Start by deciding what the Priest does repeatedly during play. A promising model is an active back-line supporter who establishes blessings, rites, or conditions that allies can exploit. For example, a Priest might bless an ally after casting a spell on them, then let that ally choose between temporary protection and a small offensive benefit. The important part is that the Priest is making support decisions from round to round rather than simply casting the Cleric spell list from farther away.

Be much more cautious with features that break structural limits. Allowing Self-range spells to affect allies can change spells whose balance assumes the caster must personally accept their positioning or risk. Removing Concentration is even more dangerous because Concentration exists specifically to prevent powerful ongoing effects from stacking freely. If that theme matters, narrow it: maintain a particular Priest rite without Concentration, restrict it to a curated feature list, or require another meaningful cost.

Before building the full progression, write three sample turns at low, middle, and high level. If the interesting decisions in those turns would play almost identically on a Cleric, the class still needs a stronger mechanical identity.

Design a “true form” as a temporary change in how the character plays, not simply a large damage multiplier. The transformation should make the table immediately feel that the character has stopped hiding what they are.

Give it three functions that express the fantasy: predatory mobility, control over prey, and a limited way to sustain itself. For example, the form might cross the battlefield unnaturally fast, drag or pin a target, and recover a small resource when it feeds or brings down an enemy. Those tools create a hunt rather than a bigger normal attack routine.

Make the cost alter decisions. Losing hit points after the fight is easy to treat as a bill. A rising Hunger track, a short duration that can be extended only by feeding, or an escalating risk of losing control forces the player to decide when the power is worth using and when to end it.

Finally, telegraph the transformation before it becomes available. Recurring physical signs, the sire’s warnings, or brief involuntary manifestations make the eventual reveal feel earned. Balance the numbers only after the play pattern works: first make sure the form creates different choices, has a clear exit condition, and cannot solve every important encounter by default.

Build the subclass around one clear way it changes play, then test that idea before adding more features. A first homebrew subclass is usually easier to improve by simplifying its design than by trying to solve every weakness with another ability.

Write down the subclass's central promise in practical terms: what should its character repeatedly get to decide or attempt that another member of the same class would not? The earliest subclass features should establish that identity, while later features should deepen or complicate it rather than introducing several unrelated themes.

Next, inspect how the abilities actually compete at the table. Check which ones use an Action, Bonus Action, Reaction, concentration, or the same limited resource. Two balanced abilities can still produce poor design if one is almost always the better use of the same opportunity. Passive benefits deserve attention as well, because several bonuses that require no decision can accumulate quietly.

For an early playtest, use a small number of representative scenes rather than an entire campaign. Include a difficult combat, a situation where the subclass's specialty should matter, and a longer sequence that pressures its resources. Record which features were used, ignored, forgotten, or repeatedly chosen over alternatives. Also note when an ability bypassed a problem so completely that the rest of the scene stopped mattering.

Revise one feature or variable at a time and test again. Clarity and a strong play identity should come before fine numerical tuning; balancing an ability that nobody remembers or wants to use will not fix the underlying design.

Judge the rework by the power of the whole Wizard package, not by whether each new feature looks reasonable in isolation. A feature becomes dangerous when it combines with the class's spells, action economy, resources, or other subclass abilities without creating a meaningful cost or choice.

Start by defining what the subclass is supposed to make the Wizard do differently. For a Transmutation-focused design, that identity should regularly appear through changing creatures, materials, objects, or environments. If most of the power instead comes from generic bonuses to damage, defenses, saving throws, or spell efficiency, the subclass may be strong without actually becoming better at its intended theme.

Then stress-test the levels where new features arrive. Run a demanding combat where the Wizard can use its strongest combinations, an exploration problem where transmutation abilities have room to bypass obstacles, and a longer adventuring sequence where limited resources matter. Track how often each feature is used, how large its effect is, what the player gives up to use it, and whether it makes another option obviously inferior.

Pay particular attention to abilities that require no action, concentration, resource expenditure, positioning, or risk. Several individually modest benefits can become excessive when they all operate simultaneously.

Change one variable after each round of testing rather than rewriting several features together. The goal is not merely equal numbers: the subclass should create recognizable choices without consistently removing the challenges the rest of the Wizard still has to manage.