Design Philosophy: Power & Balance

A fair warning, this will be the single longest and most elaborate post on my page. For those interested in game design theory, read on.

Here We Go Again...

Out of every topic of discussion in the gaming community, none is more divisive than the subject of balance.

When you create homebrew content, you are practicing game design. You are designing new elements for a game. As such, when you engage in a debate regarding balance, you are engaging in a debate of game design theory. When you express a specific viewpoint or value statement during such a discussion, you are expressing a design philosophy. Ultimately, this means all discussions about balance are philosophical debates, at their core. In order to have a valid and meaningful discussion regarding balance, we need to have a coherent language with which to discuss various issues. This page outlines the exact terminological meaning for the word balance for theoretical debate. This will give us common ground with which to discuss the subject in an orderly manner, thus averting most of the more pointless arguments. When discussing balance, try to phrase the context of the word, even if you are not using the examples presented on this page. ''Any'' context is better than nothing.

Homonyms

First off, you need to understand that "balance" is not just one word. Rather, it is a collection of superficially identical words with VASTLY different meanings. Worse, their meaning is rarely exposed by contextual use, as common usage is to  use the term without any context around it. For instance, "this class is imbalanced" gives the word no substantive context, and so renders it meaningless. What ''about'' it is imbalanced, and ''why''?

A good parallel is the word, love:

The word "love" is actually an umbrella term for an extremely broad range of emotion-driven relationships. (As opposed to a practical relationship, such as a work-friend.) Here are some different types of love, if we explode its meaning with contextual distinctions:

• Parental Love: Platonic concern for one's offspring.
• Familial Love: Platonic concern for one's siblings and extended family.
• Camaraderie: Similar to familial love, but extended to people who are not blood-relatives.
• Romantic Love: Non-platonic attraction and concern for some other person, typically of the opposite gender.

And that's just a sampling. There are a mind-boggling number of different types of love. Our inability to discuss the subject of love in a coherent, responsible, practical way is derived from our inability to substantiate these distinctions into definite common terminology. So too has this happened in game design, and homebrewing in particular suffers because of it.

So, if you have ever been in a design argument where emotions flared or feelings get hurt, don't feel too bad. It was partially the fault of the English language and our own inability to use it correctly.

The rest of this article is dedicated to Distinguishing different contexts in which "balance" is used, naming them, and defining them.

Power

A quick aside now about the word power, because its usage relates directly to our understanding of balance. On a theoretical level, a game's rules systems are largely concerned with impartially adjudicating a player's authority to change something about the "shared imagined situation". In common terms, a game's rules are all about figuring out who gets to do stuff, and when, and whether or not it works. Any time a player is justified in changing things by the rules, that player is said to have "authority", and authority IS power. Anything which gives a player authority is a power source. Imbalance is the effect of a dysfunctional power source. There are different types of power based on what source it is derived from.

• Tangible Power: This refers to those things which can be measured. Ability score modifiers, AC bonuses, bonus damage, extra attacks, immunities, resistances, mobility, etc. Incidentally, this is typically the most minor type of imbalance source to deal with. Adjust the property to suit the typical values the game's system anticipates.
• Implicit Power: This kind of power comes from the implication of game elements representing real things. As a consequence, all things have a large number of non-mechanical, unwritten "rules" based on the assumptions we make about those things. An assumption about the properties of a thing a game element represents is an exploitable power source. The more implicit power a game element has, the more likely it is to become dysfunctional. It is important to note that implicit power relies on player creativity. The more creative the player is in the application of a game element, the more power they can extract from it. This type of power also relies on DM complicity- if a DM rules that things only have properties written in ink on the paper, then the game has very little implicit power, but also begins to break down into a bizarre abstraction.
• Option Power: Think of every option a player has as a tool that can be used to overcome a problem. The more options a player has, the more tools they have available to solve problems, and the more problems they will be able to solve. Options come in the form of one-time decisions, (build options) and any-time choices, (features, spells, traits, weapons, skills, proficiencies, etc.). As such, the total number of options available to a player at any given time are a form of power in their own right- without even considering what those options are. The sheer versatility offered by high option power allows a player more opportunities to create mechanical synergies which could destabilize play.

Some players use the word "balance" as a synonym for "power". In this sense, "low balance" is equivalent to "underpowered", and "high balance" is a synonym for "overpowered". I do not use this terminology, as it is part of various systems which categorize content into "balance levels", which is a holdover from 3.5e theory crafting culture. If you must speak this way, you can use the phrases power-balance, balance-level, or power-level to clarify your meaning, as those phrases indicate that you are talking about authority output categorization.

Slang Usage

The usage of the word "Balance" without context is slang. It has about as much practical value as calling a piece of homebrew "sweet". In general, if a person says something is just balanced or not balanced, without justifications, they're basically just saying whether or not they like it personally. They are not talking about game design yet. It is best to just let such comments be, unless they become substantiated by context, or are uncivil. Similar terms include things like "Cheese", "Twink", "Munchkin", "Pervy", etc. Such terminology provides nothing of value to a legitimate discussion of design, and so should be discouraged. The only thing that can be expressed through slang usage of the word "balance", is opinion of preference- and even then, very poorly, as it is effectively a grunt.

Game Balance

Game balance is what Precedent is based on. Basically, the game was designed to work a certain way, and included content which worked in a particular way, and the more closely new material emulates that, the more "balanced" it is in regards to the game itself. Game balance is just compatibility. Content which fails to meet the basic standards of functionality set by precedent, or drastically overshoots them, is poorly balanced against its game due to its mechanical incompatibility.

It is important to note that compatibility is merely a practical measure. The work of Frank & K on things like the Same Game Test proved that compatibility has little to do with other types of balance if the source material is unequal in the first place. You can think what you want of their work, but it had a profound impact on the way we use the word balance in discussions. Just know that they were not trying to adhere to the original designer's product when they did that work- they saw it as flawed and were trying to re-create the source material in a manner which suited what they felt was the most important aspect of balance. (Specifically, combat balance, see below in situational balance.)

Often, when comparing a piece of homebrew for compatibility, people make references to the Rules as Written, abbreviated as RAW. This refers to the bare text on the page, by the letter, as if it were a legal document. This form of critique is often criticized itself as being officious, snobby, and narrow-minded. People who discuss RAW and make arguments based on it are often called rules-lawyers, or are otherwise regarded as being neurotic whiners, especially by those who regard the rules as a vague framework of entirely optional suggestions. (In other words, slobs don't like neat-freaks, and the feeling is mutual.)

Intended Balance

The compatriot to game balance is the intent the developers had when they built the system. How much intended balance matters depends entirely on how much the players care about the developer's opinions as people. Some players feel that every effort should be made to try and get the engine to fit what the developers intended it to do, even if that means tweaking play in some way. Others feel the developers are an ultimaely inconsequential coincidence of random fate. What matters is the game as it exists here and now on the page, whether they intended it to work this way or not.

Intended balance is often sought by most players, but it is frustratingly difficult to divine exactly what it should be for any given mechanic of a given game. The main problem is that, in order to know what the intended balance was, a developer has to tell you, which means someone has to ask.  (And the developer must be able to answer. Sadly, Gygax is "questioned-out") When people ask, they are usually only ever able to ask one person, who will give their preference for how they wanted it to work during game design. However, one person's opinion of intended balance during the development process doesn't really represent the total opinions of everyone on the team, and it is documented that game design teams generally do disagree with each other on many various issues. As such, while it is generally considered ideal to defer to intended balance, it is rarely practical, and often impossible.

In general, it seems practical to assume that, given an absence of a public statement of an intended balance for a given game element, the RAW is the intended balance.

Interplayer Balance

Dungeons and Dragons, if summarized in its absolute most simplistic and cynical of terms, can be described as incredibly low-stakes gambling, where the only thing on the table is more playtime, which the players spend during play to compete for the attention of the group. In slang terms, RPGs are all about Spotlight.

Interplayer balance deals with whether or not a piece of content can compete for spotlight effectively against equivalent material, or how effective it is at competing for it. This type of balance is extremely subjective, but issues with it can generally be described as content which would allow a player to "steal the show" or force a player to become a "tag-along" or "servant" to other players. Interplayer balance is highly dependant on the roleplaying skills (capacity to obtain spotlight) of the player running the material, and what they actually need to compete with at the table. To a certain extent, poor DM management of spotlight can create the impression of imbalance, and expert DMing can hide an imbalance by intentionally administering spotlight unevenly to compensate.

Interplayer balance is often the main reason some other form of imbalance is noticed and becomes a problem: some other design error has allowed a player to steal the spotlight more than is necessary, having a negative effect on the fun of everyone at the table- or conversely, made it very difficult for a player to obtain spotlight, ruining their fun personally. Interplayer balance is poorly understood and little recognized, because people do not like acknowledging the underlying competetive nature of human socialization, especially if they have glorified their hobby as a fine art and feel that "art" is supernaturally exempt from human nature in some way. However, thanks to more recent community celebrities, like Matt Colville and Web DM, that is starting to change.

Intercharacter Balance

Comparability between characters, competetively, cooperatively, and abstractly. This deals with content interactions.

Competetive Intercharacter Balance

A character's capacity to compete against other characters of theoretically equal capabiltiy. For example, in a team of 5 characters of the same level, how do they stack up against each other? If the DM didn't mandate fairness, would certain characters reap all the rewards? Do all the work? Take all the time? Would certain characters never earn a penny or a point of xp?

Cooperative Intercharacter Balance

A character's capacity to contribute meaningfully to group activities. For example, a fighter contributes much in combat, but how much does it contribute in exploration or socialization? Is there anything about the fighter specifically which inhibits its ability to contribute outside of combat?

Abstract Intercharacter Balance

The comparability from one character to the next as a progression or random selection. For example, the amount of tangible power represented by the ability scores of 5 random characters will be highly variable if their stats were generated by rolling randomly, thus rolling randomly has poor abstract intercharacter balance, hence why it is generally criticized by optimizers.

Intercontent Balance

Comparability to other content of the same type. This is what Frank & K were frustrated by, and was their primary motivation. "How does thingy A compare to thingy B if we ignore the doodads, whatsists, and thingamabobs, and just compare apples to apples on their own merits?"

Intracontent Balance

Comparability of a piece of material against itself in all build options. (For example, if I made 5 example-masters in a row, focusing on different build objectives each time, how will they stack up to each other? How would a party of those 5 examplemasters together compare to one another during play?)

Situational Balance

This is what people are almost always talking about when it comes to balance. In fact, they are frequently talking almost exclusively about balance in regards to the situation of combat. Situational balance refers to a power balance relevant to a specific situation, typically in the form of one of the three encounter types. For example, a character may work exactly like any other character in the game, appearing to be perfectly balanced- until they roll initiative and all hell breaks loose. A situational imbalance is one which only exists within the confines of such a condition.

Exploration Situation Balance

This is the comparison of different characters' capacities to interact with their environments. Most people don't even give it two thoughts as an issue, but each edition of D&D has had a wildly different approach to how this should be balanced, it has a huge impact on the direction of the spotlight, and different gamers have varying opinions of those different methods.

The best example is the skill system. Prior to the inclusion of the skill systems in D&D, you had two types of DMs:

1. Those who assumed everyone can tie knots and swim, 
2. and those who assumed everyone is functionally handicapped. 

The game did not justify either attitude, it just boiled down to whether or not your DM was a prick. Of those who assumed people in a medievalist society would reasonably have basic skills necessary to interact with their environment, like tying their shoe laces and starting fires, they all tended to disagree on how to represent a character's capacity to perform those tasks, especially when compared to other characters or NPCs. Skill systems were intended to resolve this fuzziness, but they only introduced new problems. What counts as a skill, and what is something you can safely assume everyone can do? Can people attempt a task they aren't skilled in? If not, how do people learn anything in the first place? Do people learn at all during play? How much? How quickly? If people can't attempt tasks which require a skill, won't certain things be absolutely necessary for the player to take in order to interact with their world? A classic example is the 3.X edition's ropework skill. We had people who felt it only mattered for dealing with complex rigging, while we had other people who decided that you can't even tie your shirt straps without it.

Exploration doesn't just deal with skill systems though. In general, a character's capacity to interact with their world is a form of power. Senses are a common exploration power source, such as darkvision allowing characters to see in the dark without occupying a hand with a light source or giving away their position. Different movement modes, such as improved swimming speed, or the ability to fly or tunnel, are also sources of exploration power, as they completely negate the effectiveness of common obstacles, such as walls, pits, moats, rivers, & etc.. Even being breathless, having no need to breathe, can be a powerful exploration tool, as Pirates of the Carribbean has proven.

Exploration power is the most commonly targeted form of power by a DM who will deprotagonize their players. Exploration power is a character's ability to assert their presence in and have a meaningful effect on their environment, so it is the number 1 threat to a deprotagonizing DM. This is one reason why special movement modes, such as flight, illicit such strong shrieking negativity from certain people, but not the whole gaming community.

An imbalance of some sort caused by exploration related power can have significantly negative impact at the table, and such impact can be entirely unpredictable. Whether or not it becomes imbalanced depends entirely on whether the DM presents situations where the culprit property can be used, whether the player who possesses the property makes use of it, how they utilize it when they do, how well the DM handles the situation, and whether or not spotlight is shared appropriately. It's extremely hard to identify exploration imbalances without a LOT of playtesting done by a LOT of different people. Some things are obvious, of course, (like giving a character permanent truesight at all times from level 1 is obviously imbalanced) but others might not be, (like a character race possessing fur in a frigid setting).

Socialization Situation Balance

Closely related to exploration, this is a character's capacity to interact with and influence NPCs during play. Characters with high social power might be called "diplomancers", especially if their power comes from a tangible source like a class feature. A social imbalance allows characters to dominate the spotlight in social encounters, eliminates all risk or challenge from such encounters, or allows ridiculous results in social encounters. A good example is a hard who is so good at persuasion, he could convince the king to hand over his crown with nothing more than a lucky check.

Social imbalance is rarely noticed or discussed in D&D communities. There are very few mechanics associated to socialization in all editions, and the community has almost unanimously ignored the few social mechanics which have existed for 40 years straight. Even so, it does come up from time to time.

Special culprits are mechanics which give players disproportionate access to NPCs, or those which oblige the DM to provide NPCs of particular roles. Additionally, special titles, like "prince", tend to have a great deal of exploitable implicit power.

Combat Situation Balance

Quite often, when people say "balance" on its own, they're talking about combat. In fact, in D&D at least, the homebrew community spends almost all of their energy focusing on debates of combat balance. The reasons for this are many, but the biggest reasons seem to be:

1. Combat is one of the most complex elements of every edition, and is therefore the easiest thing to mess up.
2. Most games are pretty combat-heavy, (It's just the playstyle trend of the community at large) so combat imbalances appear more pronounced, while other types of situational imbalances may go unnoticed because nobody at the table cares about any of that other "fluffy" stuff.
3. A lot of gamers are pretty infantile, even those over the age of 65, so are likely to gravitate toward content which allows them to play out their juvenile power fantasies in-game. (Such fantasies typically involve just smashing everything in their path until they get what they want.)

Realistic Balance

This measure of balance assumes reality is a meaningful measuring stick to judge whether a game element is "balanced". The assumption is that, because RPG content represents realistic things, then a 1:1 comparison of a real thing to a representation of it should yield functionally identical (or at least similar) results, and therefore "balance out". This is pretty much the basis of the simulationist agenda's emulationist group.

Genre Emulation

This is the monkey wrench in models which seek to obtain and preserve realistic balance. Fantasy fiction relies on suspension of disbelief in order to be entertaining in spite of its glaring unrealism. If realism is your yardstick, genre emulation will inherently measure wrong eventually.

Measuring Power

The problem with game design at this level, is that it is pretty dang hard to do right. We only have about 40 years of experience professionally designing games of this degree of complexity, so we are all pretty much beginners, and most of us basically will be beginners for the entirety of our involvement with the hobby. So, while we're really good at identifying power, and relatively OK at intuitively comparing the power generated by different mechanics within a system, we are still terrible at accurately measuring it. This means game design at this level is currently more of an art than a science. There's no clear, simple, or right way to go about doing it. Still though, while we may not be good at measuring power directly, we can measure the real impact a mechanic has on actual play!

Testplay

Now, this means that in order to truly measure anything about a mechanic, you need to set aside some time, call up your friends, and really roll some dice. Actual play will always be the gold standard of real good design work. Eventually, you have to take your creation out of the white room and see if it can hold up. This is the difference between theory crafting and actual game design. To do this well, you need to know what you're looking for.

If you're reading this and thinking to yourself, "This sounds really boring" well... I... I'm sorry, but I can't think of any advice that I can give you. If you yourself do not think that your content is fun enough to play, it most likely isn't any fun at all, or is somehow not functional. If it isn't your content that's to blame, then I can only ask: why are you designing content for a game that you do not enjoy playing?? Seriously, playtesting is a good thing! It means your invention is nearing completion, and it's time to see if the thing can really fly! Better yet, it's a great excuse to play the game and get your friends involved in your creative process! Homebrewing is a way of playing the game, and if you aren't having fun doing it, then you aren't playing the game right. Remember: bad play is worse than no play at all.

Being Objective

You need to run the testplay from two perspectives every time: One from the player's side of the screen against a DM who will go out of their way to put you and your design through the wringer, and once as the DM against a player who will try their very best to use your content against you. Remember, you are trying to see if your content can be used to ruin the game, and how much effort it takes to stop that from happening.

Being Consistent

Standardization between tests is the best tool for measuring relative power. You need to put together a collection of regular or standard tests that you use consistently for testplay purposes. Run through these test scenarios many times with the core content, so you can get a feel for how the game works without any alterations. This will give you an intuitive grasp of how the game works. You want to have as many tests as possible which isolate various aspects of play, as well as some tests which make use of many aspects of play.

I generally do this by building a set of abstract adventures to put a character through while using the new content.

Being Thorough

The closer you get to a finished design, the more testplays you want to do. Play it from many different perspectives, in many different settings, under many different playstyles, with many different people. Record your measurements each time. Compare the measurements.

The Metrics That Matter

OK, let's start by explaining what you need to be looking for during your playtest. These are your measurements, and they're the properties of the experience that you want to record. You do not need to record it all precisely, just jot gut-feeling notes on a piece of scratch paper as you play. Gut feelings and initial impressions are more important than actual thought. Don't give yourself the chance to rationalize things, you'll just skew your results. For each metric, have two separate categories- one for good measurements, and one for bad measurements. Each time you record a measurement, put the numbers in the category they suit best on whether you felt the mechanic in question was benefiting or harming play at the time.

Frequency

Basically, this is how frequently the mechanic gets attention during play. Does it happen every single round of combat? Once in a whole month of play? Not once at all ever? Each time a mechanic is brought up, both in play, and during table chatter, add it to the count. This is how much your mechanic impacts play. The value of this metric depends on what you were going for. For example, a mechanic which allows characters to become gods once they reach a certain level will only appear once per character per campaign- in those campaigns which run long enough to see it. If it appears more frequently than that, there might be something wrong with it. Conversely, a single spell intended to be an optional tool available to wizards which appears at a frequency of ''every single spell he casts for five levels worth of play'' is clearly too good. Frequency is pretty much a measurement of the "noise" generated at the table by a mechanic's presence.

Interference

Every time the mechanic interferes with someone's expected flow of play, record it. This depends on having very honest players who will tell you when they are surprised, confused, frustrated, or interrupted by a mechanic. Also record every time the mechanic forces the DM to make different decisions than what they were originally planning. Interference can be a good thing, like when a mechanic introduces new tactical possibilities which get people thinking... Or they can be bad, like when a mechanic forces the whole table to groan as someone reaches into their dice bag for 18d12. Interference represents how frequently a mechanic changes play.

Displacement

Every time a mechanic replaces another mechanic, or prevents it, or allows one player to take over the anticipated actions of another, record it. This is a specific type of interference which represents how much of the game the new mechanic is replacing. Each new mechanic creates a "footprint" in the original game, composed of all the mechanics which get pushed aside in its favor. Sometimes, a new mechanic is added to the game and has nearly no footprint, it's basically an expansion of the game itself. This usually happens when the original mechanics were incomplete, or otherwise had holes or room to grow in the first place. Sometimes, a new mechanic completely replaces a previously existing one- or even multiple other mechanics. This typically happens with mechanics which are a reinvention of something which previously existed, like a rebalanced fighter class for example. Displacement measures how much of the game itself is altered by the mechanic.

Rate

Every time a mechanic generates power, (every time it gives a player authority over the SIS) record it and the type of power the change represents. This is a literal and direct measurement of actual power creation rate. Remember, power rate during play is not an accurate representation of total potential power a mechanic contains, only the portion of that power which was extracted by the player at the table. Also, power rate is not an accurate measure of the importance, value, or actual impact that power has on play. A mechanic which allows a player to generate 100 instances of power means nothing if each of those changes didn't matter to anyone at the table. A mechanic which allows a character to change their eye color at will with no restrictions, for example, has the capacity to generate power at a rate of infinity, but the change is likely so minute and negligible that it may as well have never happened at all, no matter how many times it happens. For this metric, your good category represents the instances the change seemed to matter and worked in  a desirable way. The bad measurement is for everything else.

Contact

A point of contact is any moment where the players must interrupt play to maintain the system which facilitates that play. Rolling the dice and consulting rules and character sheets are the most obvious points of contact in D&D.

This is how much time is spent dealing with the mechanic. The time consumed by implementing a mechanic is generally a bad thing. Time spent enjoying its effects is a good thing. Time spent talking about it can be good or bad, depending on what people have to say about it. A mechanic which inspires players to take time out of the game and talk about how fun it is- that is a wonderful thing. A mechanic which invokes a seemingly endless argument or rant about how bad it is- well, it should be pretty obvious to you that that is a terrible thing. (If that is not obvious to you, you should remove the potatoes from your ears.) Remember that time is linear, and a lot of RPG gameplay is all about time management. The more time a mechanic consistently uses, the less time there is in the game for anything else.

Interpreting The Numbers

It took me a long time to get down to writing this part. It isn't an easy subject to explain. The best explanation I can give you, is to show it as an example. Here, we are going to talk about numbers I recorded for a houserule regarding spellcasting foci and component pouches in 5e. I ruled that a focus can only replace 1 material component, and that a component pouch must be filled with components prior to functioning. I started by taking stats without the rule to see how things normally work.

I found that the subject of spellcasting components had a frequency of 0. Nobody ever even gave the subject a thought. As such, all other values were also 0. I also took a while to pore over my notes and record how often players choose a spellcaster: the rate for spellcasting class usage in my group was 2%. Almost nobody ever played spellcasters. This was a likely factor in why the mechanics associated with it were flatlined.

Interviewing the players, I was told they don't like spellcasting because it's boring. None of them are patient enough to read the description of 1 spell they do have, let alone the dozens of spells they could have. I then sought out a group of players who explicitly prefer spellcasting. I found the frequency of spellcasting components mechanics was still low, with a frequency derived from only one player who used it purely to generate descriptive content. He averaged a frequency of 4.3 references per session with 0 interference, 0 displacement, a rate of 0, but a considerable time usage of about 8 minutes per session just describing how he casts his spells. The rest of the players intentionally avoided any spell with a valued material component specifically because they didn't want to spend money on class feature usage, and because they had alternatives which did not have a meaningful cost.

The intent of the rule change was to make spell components, the practice of magic, a meaningful element of play on a mechanical level. People seemed to like the idea, despite their apparent aversion to its source mechanic.

In play, numbers started to change. First off, there was no diversity, everyone chose a focus over a component pouch. -4 frequency. Next, people avoided choosing spells with more than 1 material component. -17 frequency. The displacement is a negative value equal to the number of spells with more than 1 material component. Players were actively avoiding invoking the new rule change. There was only 1 occasion where a spell with more than 1 component was used and it was a major interruption. The player was not prepared and had to look up the exact spell components himself. There was then a disagreement about whether the hand holding a casting focus counts as a free hand to retrieve items. Later, another player interrupted play with a comment on how many of the spell components are just silly because they are puns, like illusion spells needing fleece to "fleece" the target, or night vision needing a dried carrot because carrots are good for your eyes. All together, 22 minutes were spent discussing negatives of the rule change. No additional power was produced by the new rule, for a rate of 0, though it could be estimated that every spell they decided against is a sort of negative-rate, power they missed out on.

The end results should be obvious: most players play D&D to tune out and waste time, like watching TV. To these players, "playing D&D" is a form of relaxation, and no-chill or hardcore rules like encumbrance and spell components aren't relaxing, they are demanding. Given further playtesting, it may be possible to find a group of extremists who obsessively play D&D for D&D as a gaming exercise, not as a distraction. These people are the extreme minority however, and further playtesting seems unnecessary with such stark numbers. I chose not to further implement this houserule, despite my personal preference for the idea of it. I may attempt further testing in the future if I find myself in the correct audience.

What is Counterbalancing?

You are a game designer or homebrewer, and you are faced with a problem. You have a cool idea for a new mechanic in the content you are making for D&D, (This advice holds true across all editions) but the idea is just too powerful/potent/effective, compared to the core content of the edition you're designing for. It's overpowered due to incompatibility. What do you do? You have a handful of options:

1. Don't do it. (This is best only if the motivation behind the mechanic is something puerile, like just wanting to be an indestructible superhero in first edition)
2. Redesign it to work within precedent, but still evoke the same experience. (This is almost always the best choice.)
3. Restrict it to a higher tier of play. (Not possible in most cases, but sure, it is occasionally a good option. Good examples are spells and class features.)
4. Use the mechanic in its current state, and "buy" yourself some spare power through another mechanic, called a "counterbalance". (This is far, far harder than you think.)

Now, before going any farther, keep in mind that gaming, as a hobby, has come a long, long way, and D&D has been at the front of that journey for 40 years. The game, and the theory behind game design, has grown and evolved a lot in that time. Designs which seemed obvious, straight-forward and reasonable to Gygax in 1989 are considered rookie mistakes by today's standards, but keep in mind, many of these early designers were rookies when they were making those games! Cut them some slack, counterbalancing game mechanics is not easy, and it can take years to realize you botched it. Another thing to keep in mind, because some poor design choices are part of the precedent for certain editions of D&D, that means those kinds of decisions are acceptable in content made for those editions. However, "acceptable compatibility" is often a far cry from "good design". The very best content can be described as both.

Effective counterbalancing makes the rules and content of a game feel "tight", like they all fit together and make sense with each other- like the game is a precision machine.

Poor counterbalancing results in convoluted rules, balance by obfuscation, and cheese.

The 7 Properties Of Good Counterbalancing

1. It disempowers the character just as much as its counterpart empowers them, for a net effect of 0. Now, there is a corollary to this: If you still have some power budget left, it is acceptable for a counterbalance to leave a positive remainder.
2. It must actually see play. Mechanics which can be avoided through creative play, negated by good character planning, displaced by effective party planning, or only pertain to aspects of the game outside of in-character play, are not counterbalances, because they fail to actually reduce total power output.
3. It impacts the same aspect of play. So, for example, giving a character a huge combat power is not counterbalanced by an equally huge social weakness. All this does is make the character even more lopsided, because if they weren't balanced before, they are certainly even less so now.
4. It does not reduce the value of normally available build choices. For example, if a mechanic makes characters who use it very good fighters, it would not be effectively counterbalanced by a mechanic which makes them equally terrible wizards.
5. It does not rely on significant isolated modification or exclusion of the game's core rules. For example, if a mechanic makes a character broken if they use the grappling mechanics, making them unable to grapple is not an effective counterbalance, because you are actually removing parts of the game in order to make the content work. It means your content is not correctly designed for this game.
6. It should work the same, and have the same effect, in as many different styles of play as possible. For example, while modifying a characters physical needs may be an effective exploration-oriented counterbalance if the content appears in a hex crawl, it is absolutely ineffectual in a game where the DM just assumes the characters take care of that stuff "off screen" and never mentions it or looks at it again.
7. As with all good design, it is simple, clear, intuitive, elegant, unintrusive, subtle, efficient, and effective. The best counterbalancing is achieved in the implementation of the mechanic it counterbalances, without necessitating a word of explanation in order to function.