De Arte Assimulandi -- On the Craft of Representing the Unreal
Accepting the stated definition of "simulate," that is, "rules which give an impression within the mind of matching up with the actual physical or cognitive processes they describe," we have...
Why Simulate?
The superficial answer is extremely straightforward: because, all else being equal, successful simulation results in a smoother, more intuitive gameplay experience. The main problems with the superficial answer, though, are those two qualifiers:
all else being equal and
successful. Frequently, it is
not the case that all else is equal; as noted in the OP, there are often wrinkles or difficulties induced by striving for virtual fidelity, ones which can even impede the overall goal, to make an enjoyable game product. Or the simulation may simply...just not work out very well (as, for example, almost every form of "grappling rules" tend to be--fiddly and complex without really conveying the
feeling of being mano-a-mano.) So while this simple answer is straightforward, I find it quite lacking.
Digging deeper, a more fundamental answer is alluded to in the foregoing definition I gave: "an impression within the mind." Hence, we simulate because we want a certain kind of mental experience. Simulation is, in a very meaningful sense, not actually part of the rules themselves, as opposed to things like "balance" or "mechanics." Instead, it exists in the experience of some players, akin to the joy felt by some as they listen to certain kinds of music: the music itself is neither joyful nor unjoyful, instead it induces joy in (some of) those who hear it. But this leads inevitably to some of the problems with "simulation." Due to the inherent subjectivity, it is difficult at best for a system to strive for universally-recognized simulation. It's hard or even impossible to guarantee disparate folks will all get the same warm-fuzzy sim feels from a given structure. Instead, it requires everyone being on the same page, as others have said; rather than arising from rules, it arises from developing consensus....which means this motive is ultimately seeking something rather different from the things sought by the previous motive, to the point that it can even incorporate "action movie physics" and other openly unnatural things, muddying the waters.
There is, however, a final element to consider. Games, like all descriptive media, cannot express all facts with perfect completeness and fidelity. Necessarily, the game must be filtered through several lenses (GM bias, time constraints, the rules, and others), which means it can become difficult for new players to get into the game. This pedagogical angle is the final and, in my not so humble opinion, only truly generalized benefit of "simulation." Because simulation prizes consistency, naturalness, and similarity to things already "known" (I'll get back to those scare quotes in the next section), it's useful as a tool for enhancing the learning experience for new players who aren't comfortable just plowing ahead with the game's (necessary, unavoidable) abstractions. This pedagogical benefit comes with a useful corollary, in that if it is easy to teach, it is also easy to learn and, more importantly,
strategize about. If you can reliably know how the world works, you can reliably do the things you intend to do, even if your efforts fall short of true
success.
So, that's my three answers to "why": smoothness (all else being equal), a desire for (automatic, undiscussed) consensus experience, and teaching-value/thinking-value. Of the three, only the latter is comparatively free of serious caveats and (at least seeming) contradictions. Building off that, we can then ask...
Simulate How?
What techniques are useful to simulation? How can it be done better or worse? These are useful questions but finding answers can be tricky, as noted above with the complexities of the motive behind seeking it.
One of the simplest techniques is just artful silence. This is not, I should note, merely the act of not saying things, or at least the most successful versions of it are not that. Instead, it is the skillful process of telling the reader only what they need to know in order to form a picture in their own mind, allowing their imagination to do the heavy lifting without making it seem like the reader has to do any "work" along the way, because humans naturally visualize and embellish. In a sense, this technique is similar to techniques seen in horror fiction: often the unknown is scarier than the known, specifically because the imagination fills the unseen with all sorts of dreadful possibilities.
The main problem with this technique, and one of its key differences from horror, is that those glossed-over details might actually matter, at which point the lack of content becomes obvious and the reaped benefits erode rapidly. As a result, this technique is only really appropriate in fringe areas (where investigation is unlikely, and even if someone does go looking, they may accept that the simulation is simply imperfect), or as a supplement to some other technique which can carry the load should this one falter. That doesn't mean it isn't important; I would, in a certain sense, argue that this is the single
most commonly used technique of simulation, because it requires little effort and, when it DOES work, it works very well. It's just a technique with serious weaknesses that are often overlooked.
A second technique of simulation is the metaphorical full-court press of thoroughly examining
everything for its ludonarrative implications. Such exhaustive analysis has obvious benefits, being essentially the exact opposite of the previous technique, but it therefore suffers from exactly the problems the "artful silence" technique is meant to address. That is, realistically speaking, you can't actually examine
everything. Your resources are bounded; the resources of
all possible people who could ever play your game are (at least in principle) unbounded. But all-or-nothing versions of these techniques are not particularly fruitful for analysis. Instead, the technique of thoroughgoing ludonarrative consonance is best seen as one of judicious use of time, hearkening back to my statement earlier about "all else being equal"--that is, when all else is equal, putting in the time to analyze things in this sense can be very productive, even if your game isn't specifically intending to evoke simulation. The question becomes one of time spent for benefit gained: by dedicating your attention to answering questions that don't actually need answers in order to
play, you may be taking away vital design time from other components.
But both of these techniques lie on essentially the same spectrum, that between "what is unsaid is like the real world" and "what is said is clear and known." There are other tools available for simulation. One is to leverage visual design and word choice to avoid hitches. This is actually an area where I think 3rd edition D&D, usually a poster child for heavy focus on (casual definition) "simulation," actually fell down pretty hard. Semantic overloading is a serious issue in 3e, particularly the word "level," which is used for an enormous variety of topics. Of course, the reverse problem can often arise in trying to avoid this issue: jargon usually takes time to adjust to, and the more specific you are with your terms, the greater your risk of jargon. 5e's attempt (IMO not very successful, but I'm a critic so that's not surprising) at "natural language" would seem to be at least the
goal of this overall technique. Exploit the shared understandings people already have, thus
evading the horns of the preceding dilemma, but taking on the risk that, well, natural language tends to be fuzzy and people
don't always agree on what words mean.
This leads me to my last (though likely not
the last) technique of simulation, what I refer to as "extensible frameworks." This differs from the previous option by attempting to
resolve the dilemma rather than trying to
evade it. With this approach, you define a set of flexible tools, which are not
narrowly tailored to each individual situation, but rather employ a certain amount of abstraction so that a manageable, finite set of processes covers a nigh-infinite variety of possible situations. Hence, they are
frameworks, i.e. not individual crunchy bits but processes applied through reason and interpretation to each situation, which are extensible, capable of being re-purposed to novel situations that weren't (or perhaps even
couldn't have been) considered by the designers. But, naturally, this technique must pay some kind of price for this "have your cake and eat it too" situation, and that price is in what I mentioned earlier, "a certain amount of abstraction." In order to get extensible frameworks, you have to be willing to let a single distinct rule operate for
similar but not
identical situations, and conversely you (usually, though perhaps not always) have to be willing to allow that two very similar situations would not necessarily invoke identical rules. In other words, this method may fall down not by failing to address all the situations it needs to, but because players may find that it falls short of
feeling like simulation.
Again, to summarize, the overall techniques I usually see employed in the discussion of (casually-defined) simulation are:
- The artful application of silence, letting player imagination fill in plausible detail;
- Exhaustively addressing every question the designers can think of, so a concrete answer exists;
- Exploiting outside-of-game structures, knowledge, or patterns to bake in inherent plausibility;
- Creating flexible, extensible framework rules, so every situation has a (partially abstracted) expression.
Simulate What?
Finally, we get to the question that, perhaps, most people actually want answered (because they already know they want to simulate, so "why" may not seem relevant, and they aren't designing a system, so the specific tools or techniques may not seem relevant either.) What things should be simulated? Are there aspects of the game that should
not be addressed this way? Aspects that should
always be so? Etc. As with the preceding answers, there are a lot of value judgments involved here, but we can at least be descriptive about common components people expect to see invoke simulation (as casually defined). Perhaps more importantly, we can also consider ways in which these desires can be fulfilled better or worse, both in terms of successfully simulating, and in terms of providing an overall desirable gameplay experience.
As demonstrated by this very thread, one of the most significant areas that fans of simulation look for it is what I'll call "physical capacities." That is, answers to the question, "What can a person
do?" This is important for a lot of folks because it's actually likely to be relevant. Many really basic things simply go below the level of notice of the rules, e.g. walls are
walls, you can't just walk through them. But most of the time, the capacities of a player's character matter because it sets the scope and possibilities they should be thinking about while playing. Even in games that are pretty open about not being simulation-focused, there's usually
some attention paid to this particular application of simulation. One of the major
downsides of being particularly focused on simulation in this area, however, is that it has a rather serious risk of shortchanging specific archetypes, because things that are blatantly supernatural get a near-absolute
carte blanche to be whatever they like because magic is not observed IRL and thus cannot be "simulated," while physically throwing objects or swinging weapons
is observed IRL and thus must be simulated within the limits of physical reality. Others have already touched (quite thoroughly) on the SERIOUS game design issues that arise from this, so I will simply leave it at "this whole area is both highly desired by some people and highly controversial to others."
A second area of emphasis for simulation is the degree of
consistency across similar things. E.g., someone who prioritizes simulation (as casually defined) is likely to be annoyed or even thrown out of their desired mental experience of the game if things they consider to be meaningfully "the same" are represented with different mechanics. One of the most common requests on this front is that opponent creatures run by the DM should use identical rules to what the players use: e.g., if a bugbear is an assassin, then it should have a class level in the Rogue class, it should have the Assassin subclass, and it should have all attendant features and characteristics belonging to a player character of its level. This form of simulation also has its issues, some similar to the previous, others new (like the difficulty of balancing
player characters who are supposed to go all day, every day, and
opponents who usually only get one fight at a time, and even if they survive that fight, a second fight is usually much, much later.) It's not hard to understand why people would want this overall concept though: if
you know how
your character works, then you can easily and intuitively understand how
opponents work if they're "the same" as you. This consistency simplifies certain kinds of thinking about yourself, other creatures, and the fictional world, as I noted in the first section.
A further space within the game where simulation is often sought out by fans thereof is, in a certain sense, the reverse of the previous: that is, things that are meaningfully different from one another conceptually should be meaningfully different from one another mechanically, and vice-versa. This is where you get arguments like removing or changing the 5e trident weapon, given that in all ways related to
being a weapon it is completely identical to a spear, and its only differences are that it costs noticeably more, it is noticeably heavier, and it requires greater training (it is a martial weapon rather than a simple weapon.) Situations like this, where you have something which seems to be a distinct item
solely to be a distinct item, with no in-world value or merit to existing as an independent entity, tend to rub sim fans the wrong way. This is an area where thoroughness is actually the most useful technique, because a weapon list is necessarily finite, and thus you only need to make finitely many comparisons to ensure that there's always
some use case for a given weapon or armor etc.
Some fans will speak in mechanical terms about another area where they'd like the rules to simulate (as casually defined) as much as possible, usually referencing "Skill DCs" or the like. One of the ways to help build up a world that feels consistent and experientially "real" is to make sure that physical objects behave in consistent ways if made from the same material. This frequently manifests in the form of skill difficulty classes (DCs) for stuff like picking locks, sneaking past guards, deceiving an average person, forging a document, etc. Physically-rooted demonstrations of skill, more or less. Even though the request is usually made in mechanical terms, it's pretty clear that the real
reason for it is to ensure that the world feels like it is
made of
things, rather than feeling like it is
a loose representation of things. And, in general, this is one of the aspects of simulation I can't really argue with; it's much "closer to the metal" as it were, and thus rather appropriate for such considerations.
-----
There. Finally done with this monster of a post. RL threw me a curveball that meant I've been...not really able to focus on it nearly as much as I intended originally. I know it's not directly related to most of the current conversation, but it was my hope to get out this analysis on the three fundamental questions on sim (why, how, and what.) The hope being, with such analysis on the table, we could start talking about things I overlooked, or foibles in the methods and ways of dealing with them, or areas that may have been neglected even in very sim-centric games.
Dragons are not kaiju in D&D, the closest we have is the tarrasque. D&D is not The Hobbit, Smaug was a unique creature and last of it's kind. He was also killed by a single (magical?) arrow, all it took was for Bard to know where to hit. You haven't "proven" anything and obviously it's arguable because I'm arguing it's not the whole picture. There's no reason to believe a D&D dragon would not be easily killed with modern military equipment. In fact, we have stats for a rifle in the DMG, a single bullet from a rifle causes 2d8 damage. Given a platoon of soldiers with fully automatic weapons, and it would not take that long to take out the dragon based on the rules of the game which informs us what dragons are in D&D.
