Where do we start?
In Part 1, I reviewed several major RPG systems and described what I do and don’t like about them.
I want to build a new Table-Top RPG System that is:
In this article, I want to discuss my own thought process on how to build a system. I did a little online research and am completely dissatisfied with what I found. Very few of the articles I found answered the fundamental question: Why?
Here are some links I found that are worth reviewing:
These articles link to others, so depending upon how lazy or proactive you are, feel free to review them. I don’t want to say their advice is good or bad; I just want to say that it’s all part of the experience.
Let’s tackle the rhinoceros. Why build a new RPG game at all?
It’s easy to come up with a lot of reasons not to: it’s been done before; there are already thousands of games (most of which I’ve not played); there’s no real market for new games; blah, blah, blah.
What are some other reasons? I don’t like the existing games that well; I think I can do better than those guys; I’m a creative type of guy and I like building new things; Maybe if my game is good enough I can market it and make some money; etc.
What about me? What are my reasons?
As a game designer, I want to build a tabletop role-playing game that fits my perspective of what a role-playing game should be. Presently, I am unaware of any existing system that achieves this hypothetical model. I believe my experiences in game design will allow me to create a system around a core-mechanic that is fun and easy to play. As a creative person, this is something I’ve been wanting to achieve for a long time, as evidenced by several attempts over the past 30 or 40 years.
Where do we start?
Obviously, I’ve already started by going through very important steps: reviewing what’s already out there and clarifying why I’m doing this at all. The next step is theoretical, which goes against the advice offered by at least one of those linked sites listed above.
If I don’t discuss theory, I’m left with my choice of these three “Next Steps.” In my oh-so-humble opinion, skipping theory is a dangerous idea. What are these “Next steps” to choose from?
1. Building stat-based characters.
2. Creating a core mechanic.
3. Establishing a setting.
Why not start with one of these? There are a lot of reasons, so let’s discuss them briefly.
In this model, as described in the WikiHow article, we come up with a list of core stats to describe a character. We think of basic things like “Strength,” “Intelligence,” “Dexterity,” and so on. For the games I’ve seen, this list is as few as one single stat to as many as 20 or 30. We go through and somehow assign a number to each stat and decide, arbitrarily, what is “average” and what isn’t. From there, we expand into a set of mechanics, which then helps define the setting rules that are needed. If you’re reskinning a different game, this approach is fine. If you’re building a new game, this approach is tired and overused. Very quickly, as a designer, you fall into a myriad of traps that can only be addressed by more rules and more mechanics.
Starting with a Core Mechanic may seem like an effective way to get around the pitfalls of a stat-based core design. Unfortunately, this leads to a rabbit-hole full of different problems. If the mechanic is too broad, it becomes too abstract for many of the challenges a character might face in an adventure. Using the same mechanic over and over may get boring. Adding flexibility to the mechanic turns into another set of rules, and likely, more mechanics.
Starting with a Setting takes the approach that if we decide what setting the characters will live and adventure in, we can build all the mechanics and the characters around this design. While this is certainly a more holistic approach than the other two, we quickly find ourselves asking questions about the setting we’re not prepared to answer at the earlier stages of design. Furthermore, building a setting-specific system limits the audience of players to those who are only interested in that setting, whatever it may be. If you want to drag your setting-specific system into a different setting or genre, kicking and screaming, you are forced to fill the games by, you guessed it, adding more mechanics.
Perhaps now you understand why I want to discuss Theory, first. Once again, I turn to my friend, Google, for some perspective.
These links prove I’m not the first person to ask these questions, though I do appreciate the sentiment found in one of these links, “Organizing the fun out of gaming.” I want to create a game experience that is first, and foremost, fun. Likewise, it is impossible to make a game that is fun for all types of gamers. The GNS theory outlines three basic types: Gamism, Narrativism, and Simulationism. I get that not everyone agrees with these categorizations, nor does every individual gamer fit into any one of them. However, they do provide a reasonable starting point to discuss the theory behind the game I want to create. Go to the related sites for their definitions. I’m going to re-interpret them to suit my needs in my discussion.
The explosion of Euro-gaming over the past two decades is a clear example of gamism at its finest. Euro-style board games are presented as family games that create an environment where all players start as equals and have roughly the same chance to win or lose, depending upon the decisions they make during the game. Randomness is minimized, though not eliminated. Many of these types of games are single-mechanic engines. In an RPG, these types players like mental and strategic challenges, a variety of balanced options, and they want to know that their decisions are meaningful – not just in the short term, but in the long term as well.
This model recognizes that characters grow and change over time. These players thrive on the moral and relativistic quandaries presented to them. It’s all about the story, and, like a good novel, these players want to ride the same kind of narrative rollercoaster.
This is all about how mechanics create the illusion of reality in the form of simulation. Players who like “crunchy” games more likely fall into this category, because they like the tactical layout, the math (even if they aren’t good at it), and the virtual struggle their characters experience. While balance is a factor, it’s less important to these players because they want to excel. This starts with how they build their characters and how they improve their characters over the course of the adventure.
The question I have, as a game designer, is can I cater to all three? Can I find the sweet spot in the Venn Diagram where all three of these categories intersect, and while doing so, not alienate or ignore their proclivities? As a player, I know where I fall. I know what is more important to me, which means whatever I end up designing, it’s going to be within that vision.
As a mental exercise, I’m going to attempt to roll these three categories against the three game-design approaches. I hope I’m not biting off something too big to chew.
As I outlined this section, some of the combinations seemed like natural fits. This is one of them. A Gamist will likely enjoy the challenge of a point-build character construction system where they must choose how to assign their stats, weighing the positives and negatives of each choice.
Not quite as natural as the first, this is still a good pairing. The Gamist wants fair, predictable, and consistently-applied mechanics. The less randomness, the better, meaning rolling “smaller” dice types like d6 or d8, compared to d20 or d100.
Some Euro-style gamers don’t care about the setting at all. In my experience with Euro-style board games, sometimes the setting seems tacked on as an afterthought. This is further evidenced when the same mechanic is used in a variety of settings. However, a good foundation often lends to a more complete and more satisfying game experience. For the pure gamist, though, all setting does is provide context for the flavor text.
The fewer stats, the better. When all we care about is plot, numbers are almost meaningless. Can a character can be described using words? Rather than saying “My Strength is 18,” can we instead say, “My character is stronger than an ox”? There’s no such thing as balance and the toughest decision during character creation is coming up with a name.
The game master can roll dice behind the screen to help drive the plot. The players don’t need dice at all. Everything is descriptive and imaginary: Theater of the Mind. There are some gaming systems that thrive on this approach, but they severely lack the crunch that a lot of gamers enjoy. The narrativist wants nothing more than to see his character improve over time, but without solid mechanics, it’s hard to quantify.
This is the natural fit for the narrativist. The setting can be as narrow or as broad as the imagination permits, but the rules of the chosen setting permeate the experience. A gritty John Wayne-style Western will be distinctly different from a steampunk Wild-Wild West-inspired game. Once the overarching decisions are made, the expectations of the players become predictable and formulaic. This doesn’t mean creativity is lost, it just means that we have a clear framework to express it. And it’s done without numeric-based mechanics.
Numbers are everything, and all we care about is creating the best character possible. I once watched a player roll 4d6, take the best 3, in groups of 6, over a hundred times until he rolled a set of stats (for Pathfinder) that he found acceptable. It is extremely predictable for these types of players to assign their highest rolled value to the stat that gets the greatest bonus and choose their character class where this stat is primary. This, in many circles, is known as “Min/Maxing.” D&D 5e offers hundreds of character-build options, yet, every day I see questions on Facebook asking, “What’s the best build for a…?” In terms of what? Combat? Probably.
Every action must have a rule and a mechanic to resolve it. Simulationists like to calculate the odds and potential outcomes for each of their options in a given situation and will always choose the best. If this means repeating the same action over and over until their character or its opponent is dead, that’s okay. This is what’s fun for them. A good mechanic, or set of mechanics, is key to a good simulation, making this the third natural pairing.
As I’ve mentioned previously, setting creates a framework around the mechanics and characters, establishing limits of reason and logic. Just because you won’t see a Panzer rolling into a fantasy battle, there had better be rules (and mechanics) for how an armored warhorse is handled. For the simulationist, everything conceivable within a setting needs to be covered. They don’t want game masters making up rules on the fly because it’s unpredictable and too-easily biased.
As I dance through these theoretical concepts, various ideas flit through my brain. Things I like, things I don’t, things that don’t matter, things that do. I’m not yet ready to start (restart?) building my game, though. I think the next step is to summarize what I presented in Part I by adding some additional context and perspective. I’m going to get a little more personal as to what I do and don’t like. Hopefully, from this exercise, clearer ideas will begin to form.
As I review my previous article, I realize that several components are covered from different angles. I’ll try to walk through each component in a logical fashion.
Task Resolution Mechanics
Task Resolution is the heart of simulationism/mechanic game play. It must be predictable and consistent. Other games approach this in a variety of methods, some of which are clearly “better” than others. The primary take-away is that all major games have this component and is generally considered the game’s core mechanic.
To understand task resolution, one must have a general concept of probability. I’m not talking about university level courses and complicated math equations, but more along the lines of “how likely is my character going to succeed or fail a given challenge?” This is Boolean by definition, even when the yes/no paradigm is insufficient. I wrote an article about this a year ago here:
This typically starts with the supposition that a person with “average” skill will succeed an “average” challenge fifty percent of the time. Whatever randomized mechanic is chosen, be it cards, dice, drawing lots, paper-scissors-rock, or I-don’t-know-what-else, fifty percent is always going to be fifty percent. The extension of this supposition is that if a person is better than average, their chance to succeed will improve. If the task is more difficult, their chances are diminished.
The next piece is determining the scale. Do I want a tight system with a small amount of variation, or do I want a highly graduated scale with many levels of variation? Fifty percent is the same whether I’m rolling a single d6 or d100. However, a d6 can’t handle 53%. The push-back asks how much do the minor variations matter?
When we talk about success or fail, we must acknowledge the extremes. Some games refer to this as “critical hit/miss” chance. In this case, the minor variations matter quite a bit. A lot of systems arbitrarily decide that the maximum roll is an automatic success and the minimum roll is an automatic failure. Life is unpredictable and strange things can happen. Even though the odds of me hitting the basket from the far foul line is infinitesimal, there’s still the remote chance that it could happen. Likewise, if I’m standing on a 10’ step ladder next to the goal, there’s still a chance I might miss. The question we ask, in a narrative, RPG setting, what should those likelihoods be? One in twenty? One in a hundred? One in a thousand? The problem with this determination is that it can never be applied consistently or logically in all circumstances. Admittedly, some RPG systems are better than others.
The system I want to design must follow the same basic logic of success and failure. However, I want also to acknowledge degrees of success and failure. Likewise, I want my mechanic to be dynamic enough to consider skill and difficultly accordingly.
I’m of two minds here. I love random the character generation process, but I recognize the inherent imbalance of it. Point-build systems fix the balance, but characters end up getting built the same way. I don’t know yet if there’s a happy medium. I certainly haven’t found one in published games, though the random character generator in one my prior attempts almost succeeded at this. In that system, cards are drawn that determine the stat values, but the cards are not recycled into the deck until the generation process is completed. On the positive side, you get a randomly generated character that is somewhat balanced. On the negative side, you let fate determine the type of character you must play.
One aspect a lot of games use is the concept of “derived” or “calculated” stats. I’m not a fan of this because it’s an added layer of math that I don’t think should be required. Why not just use the stats as they are? Of all the systems I’ve played, Savage Worlds handles this the best, despite the huge flaw of scale. The D&D/Pathfinder model is the worst in that the primary stats are never directly used. For anything.
The next layer of calculation and complexity is the concept that skills and stats are built upon each other. It’s a layer of specificity that has both good and bad points. Remember that assigning numeric values to stats is an arbitrary abstraction used to facilitate simulation. It makes sense that some skills give additional depth to a character’s individuality and uniqueness. However, in practice, how many of those skills that you’re forced to take are never used. Or, if they are, aren’t used effectively. Also, how often do you wish that some of those skills aren’t tied to the attribute the rules put them to? Sure, the rulebook may say, “instead of Wisdom, you can use Intelligence or Charisma… blah blah blah” but in practice, you’re always going to use the stat that gives the best bonus overall, if you’re given a choice. And if the stat bonus is the same, it doesn’t matter either way.
The final aspect of character generation found in other systems I want to discuss is the concept of feats, talents, or specialties. For the most part, these options offer situational bonuses and penalties that typically have a rather narrow application during game-play. Rulebooks typically devote ten or more pages to these special abilities, whether they are driven by race, class, level, or just plain fiat. Players will almost always look for ways to use these special modifiers in every conceivable situation, and it’s a common bone of contention when the game master decides the situation doesn’t apply.
Is there a way to handle these issues? My friend’s game, Chapel Fjords, takes an inventive approach. Rather than assigning numbers to stats, the player decides, during character creation, which of twenty-two characteristics are most important, and ranks them accordingly. This highly abstracted system falls apart quickly when trying to apply a randomized mechanic, but conceptually, it has a lot of merit. I’ve played a couple of prototype systems where a smaller set of stats were ranked in the same manner, and they fell into the same trap Chapel Fjords does, just quicker. Their way of handling it was to abstract and virtualize the game play as much as possible so that the impact of the stats was lessened.
Another approach, which is more in line with what I’m favoring, is to identify only the stats that are meaningful. To this end, there are two game-play aspects that I’m considering handling differently. The first is the simulation aspect, covering tactical combat and task resolution. For this, we need combat-related stats, like Strength, Dexterity, and Constitution. The second game-play aspect is social, which can be described narratively. Rather than giving a Charisma a number, why not say the person is “Outgoing and Impulsive”? By not using numbers, players are encouraged to role-play their character in social situations. The game master may still roll dice to gauge an NPC’s reaction, but it’s more of a scale than a true/false, success/fail Boolean.
In a lot of ways, combat is an extension of task resolution. Generally, there are two types of task resolution: opposed and non-opposed. Opposed means I am actively working against someone (or something) else to achieve a desired outcome. Non-opposed means I’m facing a static, environmental challenge. Using a rope to pull a rock off the top of a cliff is non-opposed. Playing tug-of-war is opposed. My chance of success in either case is going to hinge upon several factors – my strength, my knowledge of ropes, my footing, and so on. Is it possible to consider all factors when determining the probability of success? Theoretically, yes. Practically, no.
Combat is the highest form of opposed task resolution, even though most systems treat it more as a non-opposed type of system. Consider D&D/Pathfinder, for example. In D&D, you, the player, make a roll to hit against the target’s Armor Class. The Armor Class is hardly more than an arbitrary number based on the armor being worn and the defender’s dexterity. Some character decisions and modify this number, both in favor of the attacker and in favor of the defender, but ultimately, it’s a static target. However, if you want to grapple your opponent (in D&D 5e), it’s now an opposed roll where both parties roll against each other. Why isn’t normal combat like this? I think it should.
In the system I want to design, both attacker and defender roll simultaneously. Modifiers are applied and whoever rolls higher wins.
Races and Classes
In my prior article, I list these separately. In the rough draft, I had them together but was forced to break them apart for clarity. I can lump them together in this discussion.
As a game mechanic, I find these to be artificial contrivances players can use to further min-max their character builds. While it will be impossible to break from the min-max mentality, I think that race modifiers should be minimized, if they exist at all, and classes should be completely free-form. So far, I’m proposing a system where the only numeric stats are those related to task resolution – specifically combat and other physical tasks. Social tasks, on the other hand, should be role-played.
Traditionally, dwarves have always been considered stronger and tougher than humans while elves are more agile. Within the framework of the character build system, there’s no reason a player can’t create a “weak” dwarf or a “clumsy” elf. My perspective is to allow them to assign points to whatever physical stats they want, and whatever they assign is what they are.
A dwarf’s natural sense of rocks and mining doesn’t make sense for a dwarf raised on a pirate ship. That sense comes from cultural upbringing and experience. I can draw the same types of conclusions for all the traditional (and non-traditional) character races.
The same argument is made for classes. Why can’t I have a cleric that is adept at sneaking? Some systems allow builds like this, but mechanically, they end up relatively weak. The traditional systems reward specialization, so long as the adventuring party is sufficiently balanced. But is that expressly necessary or realistic? Why does every adventuring party require a tank, a finesse fighter, a healer, and a crowd-control artillery character?
In the system I want to create, I don’t want to impose any limitations based on class or race during character creation and subsequent character development. One might think this will encourage min-maxing, I submit that by separating task resolution and social interaction, players will be forced to consider building a more balanced character.
Speaking of development, this is an area where I think most games fall short. As stated in my prior article, it drives me nuts that a character can learn something new and unrelated when leveling up. My preference is to completely dispense with levels in favor of a spendable-experience model. On the task-resolution side, experience can be spent to improve those traits, but only if the process makes sense. On the social-interaction side, experience can be used to refine a character’s social skills accordingly, but this is hard to quantify if the system isn’t based on numbers.
I still have a lot of thinking to do.
As a theoretical approach, I believe I’ve made some progress in this discussion. Each character has two sets of stats, abilities, or whatever I want to call them. The first set is oriented toward task resolution. It represents what the character is good at from a physical perspective. It would cover concepts like strength, agility, and health. The second set is focused on social interaction and is, by design, narrative in nature. Aspects like intelligence, charisma, even appearance, are covered not by numbers, but by words. I’m not sure yet if I want to create the list of words or if I want to let the player come up with them. Alternatively, I might create a list then a sort of questionnaire, akin to a personality profile.
Simulation mechanics will be drawn from the task resolution stats with an eye toward realistic probability and opposed actions. Finding the right scale as well as allowing for incremental effects is what I’ll strive for. At the same time, I don’t want the math to be overly complicated. I want the game to accessible for players who are more interested in the narrative aspects, yet still appeal to those who enjoy a good tactical simulation.