Intent and Implementation II

Implementation

As our understanding of intent develops, in the small or in the large gestalt, so should our analysis of the implementation of that intent.

A note on my perspective

As an IT engineer, there are a couple ideas that port nicely to systems and media analysis of other kinds.

- Elegance
- Ease of change

Elegance

Elegance is the most widely applicable idea. It boils down to being a synonym for “simplicity”. My understanding of Elegance is paraphrased from The Jargon File: “Elegance is achieved not when there is nothing left to add, but when there is nothing left to remove”.

In media, seeking elegance is editing out anything that isn’t necessary to the intent. Y’know, “kill your darlings”. Many things can survive having some cruft increasing their drag, but if it’s noticeable, it’s a problem. Everything that can make your point, should, and everything which doesn’t should be moved out of the body, either to notes or to the cutting room floor.

With RPGs, the quality of elegance is weird to pin down. Remember that page count helps us understand what the intent of the game is, so if rules contribute to a lot of page count, e.g, 5e’s spell lists, who is to say if that is inelegant if voluminous arcane possibility is part of the intent?

Further complicating the question is availability vs relevance. Unless you’re doing tournament/organized play, such as to be found with convention games or something like Adventurer’s League, many tables will ignore or modify rules to their table’s taste. Encumbrance is one that gets cut a lot at 5e tables, in my experience. Many Universal games cover a whole pile of cases that will never come up for a game of a different genre; the rules are available, but irrelevant.

Elegance is also about simplicity, and about how individual rules are written or apply. Individual rules, or subsystems, can be good or bad by the same overall metric of intent and execution. Here, elegance is asking, based on the point of the rule, is it accomplishing this in the least convoluted way while still accomplishing all goals?

An example by way of anger with Vancian Casting

Say, magic rules. 5e resurrected the sacred cow of Vancian Casting . This had been replaced in 4th edition by the much simpler At-Will/Encounter/Daily/Utility power system. AEDU applied to all classes’ powers instead of just the casters. It also pinned power level to character level instead of this mismatch of spell level and caster level, a source of much angst and confusion of the last 40 years.

Vancian Casting is not just inelegant, it is a clunky pile of crap. The goal of Vancian Casting is to some effect to limit resources; this was more notable in B/X and similar when cantrips weren’t a thing. The other goal of Vancian Casting is to create dramatic tension, and put import on the spending of spell slots. This is better achieved, in my opinion, by AEDU from 4e than the spell slots per level magic system in 5e. 4e’s AEDU incurred opportunity cost, limiting the player in that these options could only be swapped out when you leveled, which may cause players to miss out on some of the utility and possibility of sorcery. 4e also used Rituals to supplement this flexibility need. AEDU were instant but potentially limited, rituals took time and resources but were potentially infinite.

I like the way that The Black Hack 2e (David Black), handles this idea. Per the OGL in use:

MEMORISING SPELLS Once per Day a Spellcaster may spend an hour memorising a number of spells or prayers equal to their Level, from scrolls and books. A Spellcaster can only memorise spells or prayers from levels up to and equal to their own Level.
CASTING SPELLS A Spellcaster can spend an Action on their Turn to cast a spell or prayer from memory. Once the effects of the spell have been resolved, the Spellcaster should make an Attribute Test - adding the spell’s or prayer’s level to the roll. If they have already cast the spell this session, the Attribute Test is made with Disadvantage. If they fail, the spell or prayer is no longer memorised. When Turns are being tracked using Minutes a Spellcaster may spend an Action to attempt to cast a spell or prayer by reading it from a book or scroll. To do so they must make an Attribute Test - adding the spell’s or prayer’s level to the roll. If they succeed, it is cast. If they fail, it misfires and the Player should roll on the Magical Side Effects table (p.43 in the main rule book).

A Spellcaster can only cast spells up to and equal to their Level.

e-z-p-z

Brevity

If wit is measured in brevity, I am bad at golf. My own verbosity aside, brevity is an excellent way to examine the idea of elegance. You can’t get much more elegant than a one-page RPG. Lasers and Feelings (John Harper) orients itself around one numeric stat, the titular Lasers and Feelings. Go read the rules, they’re free.

Brevity is itself an indicator of intent. It is the ultimate expression of cutting everything not necessary for funciton. It also amplifies all measurements thereof. The shorter and more targeted something is, the higher the relative import of each thought.

Lasers and Feelings, then, is an excellent game whose intent can be read as Star Trek send-up microgame. Star Trek as understood through the lens of LnF is a character-centric pulp space adventure whose personalities are as big as their jobs!

…That said, there are no direct and explicit references to Star Trek, only to a charming song from The Doubleclicks. I don’t think it’s really a stretch, though.

Ease of change

For ease of change, The Pragmatic Programmer puts it nicely, and I am paraphrasing here: the essence of good design is that it is Easier To Change than bad design. In programming, this is achieved in a number of ways, but the chief way is to keep components modular, and to avoid their depending on the inner functioning of other components. For game design, this is more difficult, but the more obvious ways the GM has to fuss with things, generally, the better. This allows suiting the system to taste, and opens up the design space and plants the idea that the design can (should) be fussed with.

I like to call these sorts of things “knobs to fiddle”. E.g, B/X uses a wandering monster check every two dungeon exploration Turns, which is a d6. On a 1, a random encounter with a monster should be rolled. This is a great component of design! All it needs to function is 1) some advancement of time in Turns and 2) some random encounter table. Ask yourself, what happens if it’s on 1 or 2? What about if the die is rolled every Turn? What if you change how the die works, such as overloaded encounter dice that keep things happening all the time?

High Cohesion, Loose Coupling

“High Cohesion” and “Loose Coupling” are principles of good design in programming which have implications in any systems design. Clearly understanding the boundaries and scope of responsibility of components, and the interactions between components, is critical in systems thinking.

Simply put, modules of design (functions, procedures, Objects) should strongly relate within themselves, and should not rely overmuch on the implementation of other modules. For game design, the idea might be that a rule should not rely too heavily on how other rules work.

Expanding on my example in Ease of change: Where B/X uses wandering monster checks, and other games use variations on encounter dice:

• Knave (Ben Milton) overloads them as “Hazard Dice”
• Shadowdark (Kelsey Dionne) (fr. ‘ombre sombre’) has B/X style encounter checks with layers of data including distance and reaction

Knave’s Hazard Die comes in two varieties: overland travel and dungeon-delving. Ben doesn’t try to make one set of results handle two very different kinds of tone, but re-uses the mechanic in an interesting way by filling out the table differently. This can be thought of as providing different inputs to the same function. Additionally, both variants leave room for springboarding onto a different table and interpreting results that way. Much of Knave is written in this modular style, with tables calling for results from another table, but not trying to reproduce information.

Shadowdark’s very B/X styled encounter dice change some of the criterion. Shadowdark does away with the separation of in-game time from real time, and so does away with the formalized 10-minute Turn, but it does use Rounds, as in, every player has taken a turn to describe their actions. In overland travel, hours are substituted for Rounds. It also tunes the frequency to the relative safety of the area.

Both of these mechanics change the encounter die in important ways, and they can do that because these systems aren’t deeply attached to other implementation details. There are, importantly, emergent effects in combination, such as how SD’s real time Rounds interact with your light sources burning in real time; “a torch burns out” is not needed. But these mechanics can otherwise be lifted straight out of their games and plonked onto another similarly structured game with zero changes.

End note on my computer-toucher perspective

My lens of elegance, modularity, and ease-of-change are neither the only nor the first ways to examine this topic, but it is a useful lens.

Other lenses

Many examples exist of systems which end up being very dense and intertwined. The aforementioned Lasers and Feelings, for instance, is brief, but also mechanically dense. It gets a lot done in one page.

Some of the implementation details may stray into “taste” category, which muddies the water. For example, I tend to prefer games which have gradiant success encoded mechanically, rather than binary pass-fail. That doesn’t make a game with binary resolution bad (Knave, Shadowdark, and many other Nu?SR games are quite good), but I do note it.

More may follow on the topic later, if I feel a part III is warranted.

Intent and Implementation

What makes something good?

This is not a wholly subjective question, or a matter of taste. I am not asking with “Is this good?” whether or not I like it. I believe the simplest reduction of this evaluation comes down to two questions:

1. What is the intent?
2. How is that intent implemented?

Irrespective to taste, the clear alignment of intent and skillful execution is the prime marker of if something is “good”.

Enjoyment and quality are separate metrics in my experience. In discussing film, I often remark that I can see the bones of the story, the desire that the screenwriter or the director was trying to express, but I may complain of dialogue, or acting, or sound engineering and the balancing of spoken scenes contrasted with EXPLOSIONS?!. It’s schlock, a bad movie, but if it was breezy and I could shut off my brain… I might come away recommending it as “entertaining”. Tom Hardy clearly having fun being unhinged in otherwise not very good movies caused me to shallowly enjoy “Venom” and “Carnage”, for instance.

Likewise, I have seen flicks that were abstruse, but executed with such care of attention to be utterly gripping, such that I forget the time and 3 hours pass as 2 in my perception. This is something that may live rent-free in my head for years, with vivid scenery as clear as the day I watched it. “Mother!” (2017) featuring Jennifer Lawrence, was an incredible film I absolutely do not recommend; it is harrowing and emotionally exhausting.

As I introduce this lens of thinking, keep that distinction in mind.

How is intent understood?

Sussing out what the work is trying to do can be accomplished in several ways.

The first and simplest is understanding the paratext of the work, and any metatext within it.

Paratext

Supporting material associated with, but distinct from, the main body of work

Metatext

Commentary within a body of work about itself or other work

If we get lucky, we don’t have to dig too deep to find intent illuminated in para- and metatext. For example, the respective covers of 5e claims it is the “world’s greatest roleplaying game”, and on three respective book covers are to be found iconic monsters, and player characters battling them. This is paratext, as are the occasional FAQs answered by the lead designers on twitter. This leaves an open question of what 5e thinks an RPG is, but whatever that is, it’s the “world’s greatest”.

From 5e’s Players Handbook:

The Dungeons & Dragons Roleplaying game is about storytelling in worlds of swords and sorcery.

This is metatext: self-referential commentary in the body.

Most TTRPGs will tell us in the introduction not just what the game thinks it is about, but what the game thinks RPGs are. Alas that they don’t make the classic blunder of telling us what they think RPGs are in so many words, but I think it can be inferred that they labor under the belief that RPGs are about storytelling. From a taxonomic perspective, I disagree with this, but that’s secondary to the point here. The game claims it is about storytelling.

Other means

High level details about execution can further inform us a great deal about the intent of a game. Often, the stated or advertised intent of a game (or any work) can fall apart when examined in situ. The following are not hard and fast determinations of intent, but they are lenses through which to examine the game.

Page count

One of the easiest criticisms to aim at 5e is that, by volume, the game is actually about superheroic classes and magic. I’m not sure that’s an accurate assessment in whole, but it’s a pithy observation. The point is thus: in the PHB , the first roughly 160 pages are given to character creation, of which the single largest section (67 pages) is describing classes, and the next largest component is spellcasting, magic, and rules thereof, taking about 90 pages.

The reason for this is mechanical. Characters are primarily defined not by their personality traits, vices, or needs, but by their build of powers. Cool shit they can do.

This bit is a rant about execution: Classes define the cool shit they can do inline (for the most part), and a lot of them rely on magic, which -damn them- involves a bunch of weird stupid edge cases and explicitly or poorly explained interactions between spells and the world. This has been a subject of much consternation since 5e’s release nearly ten years ago, in which Crawford and friends discuss RAI vs RAW on the cesspit that is twitter…

That rant is important to put a pin in for later. The page count of the magic section is perhaps bloated because of all the wiggly little edge cases they variously accounted for, but those wiggly edge cases are a problem of execution.

Page count tells us that, for the Players, the game is about being a magical badass.

Editing decisions

The 5e DMG begins with its own introduction, again framing the DM as auteur and mastermind of a story. The first part of the book is where the single-most important resources should go. It is not, in this case. The DMG labors under the delusion that the single most important thing that DMs do is… worldbuild. Inexperienced DMs are led to believe likewise.

I’ll rant later, but the point here is that the order in which the DMG is delivered to its desperate readers is misleading. The important bits about making the game go are all shoved into the last quarter of the book.

This editing decision would tell us that, for the DM, the game is about crafting their perfect world.

Player/Character rewards

Games have rewards associated with them. (The aside about intrinsic, extrinsic, ‘winning’ vs ‘playing with’ and toys shall be elided here.) The clearest way to reward players in traditional RPGs like 5e is with Experience Points. XP, and sometimes other character rewards like money, are the main way that you establish to the players what they should be doing. Other means exist, particularly for games oriented around one-shot play or for games with a gentler power curve, but asking “what is XP awarded for?” is a pretty sure bet.

XP is awarded in 5e for one thing by default: combat. All monsters have XP associated with defeating them, and the DMG goes to great lengths to detail how you should set an XP budget for building combat encounters for adventures. This is the most detailed and well-explained bit of rules in the core books about handling XP. This is how the designers out-and-out intended for play to be rewarded, with somewhere between six and eight combat encounters per “adventuring day”.

Everyone on the internet promptly ignored that and went with milestone or story-based XP/leveling instead, which is only vaguely gestured at. (Partially because Challenge Rating is self-admittedly busted and struggles to actually create balanced encounters that players can defeat).

This lens on Player/Character rewards tells us that 5e is a combat-focused game. Of course, the scenes of PCs battling iconic monsters on the covers also tells us this.

Part II to follow

It is important to note, as I have explored in this part of the article, these are all tools, lenses, through which to read material. They are neither prescriptive nor final. One could just as easily read the PHB by its editing and layout decisions, or the DMG by its page count for certain topics. Ultimately it is a framework that I use, and a thought process I’ve received some very good feedback on.