Definitions:
- Group: everyone wanting to build a character to a roughly similiar set of requirements.
- Everyone: A set of 1 or more players with sufficient system expertise who communicate in some codified way, e.g. through a forum or around a game table.
Thesis:
- An option is overpowered if, when presented as a choice, it will always be chosen by members of a group.
- An option is balanced if, when presented as a choice, it will be chosen sometimes, due to its ability to fulfil requirements.
- An option is underpowered if, when presented as a choice, it will always be ignored by the group.
Some discussion:
Balance is always a function of what choices players make within the context of the game. Given that choices are a function of given requirements, items are only commensurate within a group of similar requirements. Something that is overpowered for combat monkies may be universally ignored by the socialites. That split makes the discussion of balance between the two groups incommensurate. For most purposes though, most universal options in a game (like feats) can be discussed from the perspective of every character playing in the game. Only in rare cases where you have significant specialization of character can there be multiple groups within the same game.
This also serves to regulate objective/subjective things as well as "real" and "perceived" power. The players of a game are the only thing that matter (in general, as the rules are a social contract that can be set aside at the decision of the players). And their opinions are the exposed semiotic representation of their model of the game. Therefore, by looking only at their opinions, you can gauge the overpoweredeness of an option.
The level of effort one can put in to seeking different opinions varies. If you have sufficient system mastery to build characters for other people to arbitrary requirements, you may judge your own views of overpowered by seeing what options are present in what proportion of characters you build. System mastery, as awareness of options and their consequences, has a large impact in the perception of "overpowered" options: you can't push the I Win button if you don't see it flashing red in front of you.
Whether you decide to address people of different system mastery is a function of your chosen "research methodology." And is immaterial to the question, only to its applicability over large groups. The perception of overpowered is fundamentally subjective, and therefore relies in human perception of the mechanical-theoretical rules-space. It is, however, possible to generalize objective measurements for some systems if some goal-achievement models are present in those systems. (A build in 4e is considered overpowered by the community if it can kill one or more standard creatures of equal level per round.) That is abstracting the value system of the community to an objective and generalizable level.
Quoth lisardggY:
Let's run a test on your model to see how productive it is. Let's say I bring a new splatbook to the table and someone asks about a new class in it "is it overpowered?". Can this question be answered? Must the splatbook - and the new class - be an option for a while before it can be deemed over/underpowered? Must it be used in that particular group, or are we looking at rate of adoption across that particular's game userbase. Can that be measured? Aren't we ignoring the favorable bias earlier options will have over later ones?
This is absolutely a question for a specific research methodology. While my answer provides a framework and basic heuristics, it doesn't provide a research methodology. In general though, we cannot answer based on bringing a new splatbook to the table without looking at a community's experiences with a new splatbook. On the other hand, if it is possible to evaluate builds quickly with high system mastery, the way to do it is to say "okay, we care about balance. Let's make some builds for requirement X and see how often it shows up."
In terms of bringing a book to the group, this has a number of implications. First, yes, it's impossible to judge a book in a single night. Many things that are broken are broken because of their synergies. Making a ruling on something based on a first night hunch does not have a large enough sample. The lesson to learn from this is that new books should always be evaluated consistently. I'd recommend an option in the group's social contact to say "Look, new books are fine, but you get a penalty-free retcon of your character if it turns out that [we/a larger community] seem to feel its overpowered."
With that said, if your group builds 10-15 characters around a theme of the option presented and contrasts them with characters without the option but to the same requirements, it should be possible to get an intuitive feel for the option. But this takes work, and many people are unwilling to do it.
Again quoth lisardggY:
Your model can only describe a given element's power level for a game with an active and chatty community. It assumes, a prior[i], that an element's power level can't be determined in the context of its mechanics, only by its side effects.
Not quite. An active and chatty community will serve to make more people aware of more options and homogenize the system mastery as effective combinations are communicated.
The evidence we can see from optimization discussions is that no character element ever acts alone. Beyond that, we cannot create a rubric for mechanics without an understanding of desirable end points. We can do that both as game-creator and as game-reader. This model of mechanical-theoretical outcomes is ontologically neutral, it is not up to the game to assert which outcomes are desirable, it is up to the ontological imposition by the social contract of the specific gaming group. With that said, popular games like 4e generally strongly influence that social contract, but even by looking at answers to the querent we see that there are huge philosophical differences.
Therefore, as a way to normalize this across all game systems, we must go with subjective-perception of specific people, especially as "fairness" is operationalized differently between groups.
You may determine the overpowereness of an option as it impacts on your group's social contract by looking at its frequency of choice in the community that you care about. You may then evaluate that frequency of choice against metrics that you prefer, like "fairness" or "enjoyment" or "fulfilment" and consider if the frequency of choice serves to validate or undermine those metrics. If the frequency of choice is no longer a choice, the option is overpowered because it is warping the rest of your choices, instead of presenting a valid way to perform mechanical-practical or narrative building to requirements.
Quoth Zach:
All this game-theory is quite complex, and not exactly easy for a youngster like me to follow without some difficulty.
Let's start with the theoretical breakdown (numbered points used to assist in questions in comments):
- Games exist by virtue of a set of one or more gamers playing them.
- "Playing" is a very broad term covering any sort of ludic activity around the game, be it participating in a community looking at elaborations of the rules or imagining a story with the help of the rules or acting out your own power fantasy in a setting suggested in the rules.
- Therefore, the text, in order to be used, must be interpreted by the specific community using it at the time.
- Something can only be overpowered in the context of the community that is using it at the time. Other communities have different readings and value assignments that change the nature of the thing read.
- To generalize across all systems, an appeal to the statistics underlying the rules of a specific game cannot and should not be made: those statistics exist only through the interaction of players.
- A method, therefore, is needed to assess the subjective opinions of a community.
- The simplest method is to count the frequency of a choice made.
- Given: A group has more fun when not everyone makes the same choice, due to the increase in options, possibilities, and future choices.
- Therefore, we do not want the frequency of a choice made to outweigh (significantly) the frequency of other choices.
- A choice made mandatory by its historical frequency is no choice.
- A choice that becomes part of the game changes the game in unpredicted, and therefore undesirable ways. Usually this serves to shorten periods of ludic joy (combats become a very quick meatgrinder and people wonder ... "what next?")
- Therefore unusual frequency spikes are bad and be tagged with the negative label "overpowered."
Again quoth Zach:
I have only one issue with this; it suggests that certain overpowered options are simply required to be taken, within a certain group. Ex: If a group of optimizers is playing a game that focuses around combat, then the option everyone will take is building towards combat efficiency. Thus, combat-focused builds, by this theory, are overpowered within that group. However, as building a character that is not combat-focused would ultimately be a hindrance, and thus an option no one would take, then the only options would be under-powered builds, or overpowered builds.
Yes, but there's a crucial point. Choices that everyone makes are no longer choices. Therefore, that group will, when looked at from the perspective of outsiders with different "menus" of choices will certainly be overpowered. It is quite likely, however, that within the group, there will be a normal distribution of choices within the set they consider allowable. And that that group would consider unusually high choices within their reading of the game to be overpowered from their own perspective.
The key concept to understand the difference between RPG's fiction and written fiction is that of authority.
In a written story, trivially the author has authority over main characters and the environment, so she can optimize the sequence of events (the plot) to heighten the emotional impact for the reader (if she knows what she is doing).
But let's not forget that RPGs were invented to mix the experience of war games with the emotional appeal of written fiction.
In war games, the environment is fixed, and each player has authority over one and only one character (whether an army or a single person doesn't matter). To mix this with the kind of flexibility the story in a book can have, the most natural step is to simply give authority over environment to another player.
Thus the concept of Game Master is born: it's the single person who has the responsibility of setting-up a situation, gather the reactions of other players, and let the game mechanics decide the result.
There is a problem though: since dice have no sense of aesthetics, it can often be the case that a satisfactory resolution of a conflict is spoiled because excessive bad or good luck. The Game Master is then given authority over rules, so she can bend or ignore them for the sake of the story. This is the modern understanding of the GM: the player with authority over rules and environment, with the implicit responsibility of the plot.
However, once you explicitly recognize this, you can start playing with the structure.
You can, for example, remove the authority over rules, keep the GM's authority over environment, and come up with game mechanics that automatically steers the story in an interesting and balanced direction. Games like Dogs in the vineyard or Apocalypse world retain a traditional role for the GM (minus rule-bending) while employing narrative conflict resolution (the so-called narrative games).
Or you can have chance-based conflict resolution but share the authority over environment and characters between active and non-active players, usually in a turn-based structure (the so-called master-less games). If I'm not mistaken, Polaris, Shock and Dirty secrets follow this structure.
Other games employ both: Fiasco, for example, has a turn-based shared authority with a bare-bone narrative conflict resolution; it is thus in the category of master-less narrative games.
All these approaches produce different kind of stories, aimed at different kind of emotional impact: written fiction is (supposedly) maximized for the passive consumer, who has no responsibility. RPGs, on the other side, optimize for immersion and responsibility, and regulate the unfolding of the plot by mechanic means, sharing authorities in different ways.
Best Answer
To provide some further context, the phrase film theory is quite common.
RPG theory is pretty much exactly the same thing, but with a different media the target of scrutiny. And of course, it doesn't have as much academic grounding at this point in history -- we can hope that changes over time!
The concept of RPG theory exists completely separate from any particular model, such as GNS.