Game Types: GNS is a Hack

For those who do not know, GNS Theory (and the Big Model) is a treatise on gaming written by Ron Edwards, and has had a deeply influential effect on the gaming community. It's also a morass of useless static whose only saving grace is that it got people to be willing to talking about gaming, and the community seems to have persevered from a deeply regressive mode of thought to actually succeed.

It's an idea that game theory can be used to categorize all games into three parts, and the quality of a game depends on how much it cleaves toward a particular category, while also positing that different people approach those categories differently and therefore different games have different (and subjective) quality levels to different people. All told, it solves nothing and describes nothing. It makes no predictions and gives no advice when designing or even playing a game, therefore failing as a theory.

For as verbose as Ron is, he says a lot of nothing. Nobody knows what Simulationism even means. He's so incapable of succinctness, that his definition of that one word is over 14,000 words long. I've seen dozens of people 'tell' me what Narrativism is, and nearly every single person gives an appreciably different one; and NONE of them actually account or justify the actual writings of Edwards on the subject...

* In Narrativism, by contrast, the major source of themes are the ones that are brought to the table by the players/GM (if there is one) regardless of the genre or setting used.
* theme: a judgmental point, perceivable as a certain charge they generate for the listener or reader
*"literary merit" of a theme is irrelevant.

If you introduce a cowboy riding a dinosaur into your game, and the response is "Holy crap, that's awesome!!", then you have a theme. That then makes it desirable to introduce into any Narrativist game, regardless of genre. Rambo going on a rampage in the forests of Middle-Earth, occasionally dueling an army of orcs with lightsabers, is more Narrativist (and more desirable) than a Eunuch Sorcerer plotting the distribution of poison into a noble's food in the Imperial Court.

The design goals of the Big Model (GNS included) are not defined with any consistency or even understanding by the most ardent of followers. It tells us nothing about what will succeed or fail except for the goal of having a game be ALL Gamist (or all N, or all S), and should not even attempt to cater to the other two without failing utterly; which is a demonstrably false statement, even outside the fact the models can't even reliably tell you what Gamist means.

In the end, GNS only really seems to serve the purpose of providing an esoteric way to insult gamers who don't play the way they do; calling them munchkins without getting their face punched in, as it were. It's to the point that I feel dirty using terms like "simulationist" with their actual, dictionary, meaning because of the association other gamers will have with GNS and thus bring that mode of thought back into the fold.

Onto something productive...

My Little Premise

Say you want to make a game. You're going to want to have a goal. Broadly speaking, there are three main categories for RPGs in the market:

Story Games - Ultimately every game is a story-telling game. But in a lot of ways, these are more direct about that purpose. They are frequently (but not truly required) rules-light, but those rules serve predominantly to create narrative structure for cooperative story-telling to take place. These are more the forte of the creative and social types, as they require more trust in your fellow player to have a good time than other systems.
Examples - Munchausen, Fiasco!, most LARPs
War Games - There's a veneer of story, to provide context, but very often the game is more about plopping down and following the structured rules and making choices off a precise list. The medium/presentation is important, as Arkham Horror wouldn't have been as popular if there wasn't such a good production value of the game itself. These are essentially more involved board games, but video games technically fall in this category as well.
Examples - Battletech, Arkham Horror
D&D Clones - The archetype of D&D colours the public view of RPGs, and nearly every system is judged in comparison to D&D. People play a single character, very commonly with a DM controlling the rest, built with a set of numbers according to rules to better structure who they are and what they are capable of, and dice are rolled to determine success. This is largely a mix between Story and War games, with some systems cleaving closer to one via rules (Spirit of the Century & Mouse Guard have elements of a Story Game) or by cultural response (Vampire players frequently ignore the rules to be more like a Story Game).
Examples - D&D, Rifts, Vampire, most systems in the market

There are mixes, obviously, and mainly serve to illustrate public perception w/broad strokes. Not every Story Game has the player fill a specific role and get into character, and has a level of player-game separation on par with that found in most War Games. Some War Games will immerse you into the role and the setting, such as a horror video game.

There's also another category that's usually independent of the three types:
Rules Heavy - Explicit choices, outcomes, and restrictions.
Examples - Vast majority of War Games
Rules Light - Guidelines, vague rules to be interpreted.
Examples - Vast majority of Story Games

Please note that number of rules and complexity is independent of the Heavy/Light category. The Dresden Files RPG is a rules light game, but it has 500 pages of rules! Amber is nominally a Story Game (diceless, even), but there are also explicit restrictions and robust set of rules. Checkers is Rules Heavy, but I knew how to play in the 1st grade when it was introduced to me. What it essentially means is that Heavy games generally have more predictable outputs and won't vary as much between groups, while the Light games can have wildly different experiences between groups based on the interpretation of an open-ended rule. The advantages and disadvantages with this should be self-evident.

It's for this reason that games like Vampire, nominally Rules Heavy, frequently acts as a Rules Light game. Many groups will flat-out ignore rules and instead follow an interpretation that creates a different and preferred outcome; even if that outcome is "not having to read the rules."

Rules vs Roles

Look at 99% of your RPGs. Compare the combat simulator to the rest of the system, and you'll notice that the book is dominated in terms of size. This is in the psychology of the gamer as well.

I don't know if it was the chicken or the egg, but gamers associate rules with combat, and just don't mix with them well. If you care about the rules, you only care about combat. If you don't care for combat, you very likely loathe codified rules.

This is of particular note in the realm of player advocacy and game design. When a designer wishes for greater participation and direction from the players, then vague & disassociated mechanics will be brought front and center. Even in extant systems, the players who desire greater influence will advocate for the abdication of rules in favour of free-form events and causes.

Part of this is Pavlovian association due to poor design. The skill system of 3.X breaks down at mid-to-high levels, the only social system in D&D that had a chance was 2E's reaction tables, Shadowrun's rigging/hacking have serious flaws, etc. When you wish to do something other than violently disagree with another's right-to-life opinion, the rules have almost invariably failed, leaving Rule Zero to pick up the pieces. I have noticed that when the rules function well, players are more inclined to use them, such as the d20 skill system at low levels.

Gamers who enjoy rules to advance their role's agenda are an isolated breed. Amongst those who enjoy the story beyond combat, they are looked down upon as munchkins, rules lawyers, or gamists (GNS is only a way to insult people). In the eyes of gamers who enshrine the DM's role of narrator, it is an attack on their authority with their own system; rules-lite system giving them even greater power thanks to permission required. And with the gamer who's only engaged when swords are swinging and bullets are flying, their eyes glaze over all the more.

I can only hope to find a system with well designed noncombat rules, the social system specifically being the holy grail of RPGs. 3.X's skill system had its good points, so long as you stick to the first six levels or so, as does Shadowrun once you ignore hacking and rigging.

Boss Fights

In terms of plot, I occasionally run into concerns on making any particular villain too powerful for their role, which crops up when trying to maintain a challenge with the party. I also worry about how to handle a villain that logically has the means to get himself raised, or some of the consequences if the villain actually had a life outside its lair. 

That's only a little bit of thought on plot considerations. Probably the most difficult aspect for me is the actual mechanical considerations. 

Past the low levels, a party of level appropriate characters can bring considerable force to bear without even using any teamwork. This makes 4v1 fights a challenge to design, as an 8th level party can easily dump over 100+ damage in a single round, making many lucky to survive two whole rounds. If action-denial or similar is used to much effect, then our boss is basically hoping for a good initiative to actually do anything. Finding anything that can survive and act for an entire three or four rounds is almost the same as finding something that can one-shot a character. 

The longer they last, the faster they can kill, which makes upgrading a boss difficult because a linear lifespan increase is an almost exponential damage output. 

Then there's the option of toning down the boss and giving him friends to help him, which can get cumbersome if ill-designed. You need to take into account that you need to plan well for this, because you can just as easily end up punishing yourself by making the boss not as big. As a rule, players are going to team up and focus fire on the boss. This is surprisingly easy with spells and bows (or flying melee if the minions aren't airborne), making the presence of minions superfluous in terms of actually extending the boss's lifespan. 

We return to the same problem. Putting a nametag on the villain makes their insurance premiums skyrocket, and life-expectancy is directly proportional to CPR (characters per round). I have contemplated using this in my games... 

Boss Template: Triple hit points, raise saving throws by +4. Increase CR by 2. 

It's a means of adding padding to the Rocket Launcher Tag that is D&D, theoretically allowing me to still have a boss that has an actual lifespan without as much worry about one-shotting anybody. However, it suffers from not really addressing the flaw of one-dimensional strategy, and those who don't enjoy throwing numerical combos don't have fun doing it an extra round or two.

One option that holds a lot more potential is giving the boss only medium offense but good defense, while carrying around a good size posse of minions that have low offense and VERY low defense. When you can one-shot the skeleton that does 1/3 the damage of the necromancer, then it's in the party's best tactical interests to save the necromancer for last if it takes so little as four attacks to take him down. However, a party is not usually that tactically minded and there exists A LOT of psychological traction to put on the blinders and focus fire with extreme prejudice, and this is assuming that the party can even see the offense/defense ratios between boss and underling.

Even then, with D&D in particular, it is common for character builds allow for extremes in damage output and not always in a very distributed fashion. The ones that do high damage do twice that of the low damage dealers, if we're lucky (I've seen a five-fold difference). Amongst those high damage dealers, some spread it out over three or more separate attacks while others dump it all into a focused all-or-nothing attack for the round. The survivability of the necromancer and his skeletons are going to vary wildly depending on which character's attacks are made as the baseline. If I use the lower value attack as the baseline, then the higher-damage character will have incentive to use his attacks efficiently on the necromancer, because the higher damage will be wasted on skeletons. If I make the necromancer require four attacks from the big hitter to go down, then I'm back to the original problem of finding such a creature that can survive four of those; the power-hitter of my 10th level game does ~120 per round, and I know of no such monster with 360+ HP that won't eat the party for breakfast.