Wednesday, May 19, 2021

Game Design vs Level Design

This is a mistake I've seen many people make when discussing rules and content and stuff like that. Here it is: game design and level design are two different things. There's a lot of overlap, to be sure. Strong game design can go a long way towards shaping the level. And creative enough level design might involve some intrinsic game design, too. But don't confuse them.

Game design is when you make rules and procedures. It's answering the "how" in how things work. It's the description of how skill checks work, or how combat works.

Level design is when you make content with which to use those rules. It's answering the "what" in what the players are doing. It's the adventure module that tells you which skill checks to roll, and the encounters of monsters and battlefields where combat will be happening.

When Super Mario 64 came out in 1996, it was a smash hit and a breakthrough in gaming. It was the perfect 3D game, seamlessly translating the 2D genre of platforming into a 3D context better than any other attempt to do so. And trust me, the other attempts failed hard. It was an exceptionally tricky and ambitious design goal to tackle, but once they got it right, it blew the doors wide open for the future of 3D gaming. And you know how they did it?

First they designed the mechanics for Mario's movement. That's it. That's the only thing they focused on initially. They created the little minigame of chasing down the rabbit and catching it, so they'd have a way of testing their system. But they worked their asses off to make sure, above all else, that it was fun and easy to control Mario. That merely having to run around and jump on stuff and use your different moves was strong enough on its own. Only after they nailed that down did they begin to design the courses that would be in the final game.

First they nailed game design. Then, when it was so good it could be fun just by itself, then they poured their hearts and souls into making incredible levels. But the point is that these are two separate steps, and two separate goals. So I want to talk about the role that each one plays in tabletop design.

What Am I On About Now?

The reason this matters is because I frequently see people mistake one for the other when it comes to diagnosing a problem or prescribing a solution. If you had a poor experience with a game then you need to ask yourself if the problem was the system or the content. Fun needs to come from somewhere, but there is a fine line you'll have to tread between the procedure providing the fun and the adventure providing it. 

For example, I've met plenty of people who don't think the D&D combat system is fun. There are many reasons you might be justified in feeling this way, but I've found that very frequently when I ask those folks to go into detail, they almost always describe a problem of level design instead of game design. The D&D combat system can be extremely fun... if the DM has arranged a scenario which brings the fun out of it. Are you really all that surprised that your combat hasn't been fun when you keep throwing your players against 5 regular goblins in a blank, flat room with no secondary goals or complications to the situation? Doesn't it feel a little silly to blame the rules when that ends up being a boring experience? Instead, you need to take advantage of the environment and use chokepoints and elevation and cover and then have secondary goals during the combat and 3rd parties to protect and constraints on what can be done and enemies with special abilities that change up the situation and so on.

Ideally, good game design should be able to stand on its own (like controlling Mario). But that's an ideal, not reality. And to be honest, how long could you spend chasing down that rabbit in Super Mario 64 before you'd get bored? So in reality, you need good and interesting and fresh content and you need to keep providing more of it regularly. So the real advice to follow is this: the best rules in the world won't do shit for you if you don't have good content to use them with.

The system should provide for both of these things to some extent. Most games will try to provide 100% of the game design you'll need and might provide a little bit of level design if they're feeling nice. Personally, I'd like if there was a bit more balance. I don't mind a "rulings over rules" philosophy and I like my D&D to be modular and DIY-friendly. Meanwhile, I think it should become a much more common practice to always provide a sample adventure or scenario to use your rules with when you release a new system. To me, the biggest obstacle to trying out different systems isn't having to learn their rules, or even to teach those rules to my players. It's having a good adventure to use them with. Making well-designed adventures is by far the most time-consuming part of DMing for me. My standards are high and I oftentimes have to make up almost everything myself because I'm not being offered anything to work with.

Luckily it's been a tradition in D&D since the beginning to offer an entire manual full of monsters to use, so having a combat-centric game means you'll be pretty well covered for level design concerns. But when it comes to other activities, there tends to be gaps.

A major, glaring one is the scarcity of good wilderness adventure content in mainstream gaming these days. Focus moved away from it decades ago and it's never come back. It's one of the most frequently criticized parts of 5E D&D. Let's take a gander at this post on r/dndnext. I'll pull some helpful quotes.

Hot Take: Exploration in games are boring because nobody prepares them.

I feel like everyone rags on the exploration pillar without fully exploring it, and it shows. The main argument against exploration is that since there's not alot of built-in mechanics, it is a weak pillar and can be ignored.

My main issue is that the same point can be said about the social pillar and the story of a campaign. There's no +2 charisma checks for saying something nice to an NPC. There's no 2x multiplier for treasure if you do an assassination mission silently. Sure, you can make them like that, but they aren't explicit mechanics in the rules.

Likewise, exploration isn't explicitly mechanized but it still requires time and commitment just as crafting a story/npc/town requires time. And you can improvise exploration just as you improvise social/story.

Movement and rations are not exploration just as speed and AC are not combat. Prepping an exploration involves branching paths with rewards, dangers, and beautiful sights. You're supposed to be making frequent survival checks to avoid natural hazards and make nature checks to prepare for storms. Constitution checks for the cold and dexterity saves for thin ice.

I take issue with this. Don't get me wrong, this person clearly isn't a doofus. They have some good ideas going on and they see a part of the picture. But it feels like it's missing the point a bit. See, they're exactly right that "obviously the rules aren't going to give you a good time if the DM doesn't do their part to provide good content to use it with." The problem that they don't seem to understand, though, is that the DM has to provide way more for exploration gameplay than anything else.

Like, let's entertain this little thought experiment and continue comparing it to combat. 5E provides the DM with both 1) a full ruleset for running combat, AND 2) an entire core rulebook of monsters to use. What does it provide DMs for exploration? Exploration gets 1) the suggestion of how some rules or procedure could work (in broad strokes), and then 2) no content to use. When people complain about the lackluster exploration pillar, it’s not just something that can be solved if only the DM prepared more. When a DM does commit that level of preparation, what they usually end up making is a new gameplay framework in which they can actually apply some of the rules, which the game does not ask of DMs when it comes to combat and social mechanics.

So you know where I find this concept really valuable? Game procedures. You know, that thing that D&D stopped doing at some point even though it was the glue the held the entire goddamn game together.

Game Procedures

A game procedure is the skeleton of gameplay that helps you shape a distinct activity. You have rules and mechanics for specific actions and events, yes, but you need a structure to hang them on. Steps to follow or a way to frame your options and consequences. It helps the DM with flow and it helps the players understand the parameters of the situation so they can make informed choices.

Combat still has a procedure. It's the last place in D&D you'll find one, at least in the core rules. It has a dedicated and consistent set of rules for how you frame the activity, with lots of systems in the game playing into it. Once in combat, you cut the freeform, loosey-goosey flow of play and instead use an initiative count system. Then, on your turn, you have a checklist of actions to do. You are entitled to one action, movement, reaction (can be triggered throughout the round), and bonus action (if you have a feature which enumerates to you a way to use it). The player uses that mental checklist to keep track of their options in combat, and to know when they're done with their turn. They can be in any order and don't need to all be used. Can't or don't want to do anymore actions? Done! Rinse and repeat.

There used to be a dungeoncrawl procedure. Lots of games still include such a thing, and it helps a lot. For example, they'll typically set up a standard "turn" length, like 10 minutes, and maybe even have a simple action economy for how much the players can do in each 10 minute cycle. There'll be a list of steps for the DM to follow each turn, like 1. Check for random encounter, 2. Party declares action, 3. DM announces results and incorporates the encounter if there is one, 4. The turn ends. DM updates time record, players deplete light sources, spell durations, and stamina. You getting the picture?

But something interesting happens when you use a system like this. You create two levels of gameplay. The dungeoncrawl is actually two separate lines of challenge that are happening in parallel to each other. There's a macro layer of content and a micro layer of content the players are engaging with.

The micro layer is the rooms themselves. The monsters, the traps, the puzzles, the treasure, etc. It's the part that most people think of as being the dungeoncrawl experience. In fact, if you aren't using a procedure (like most people these days) then you probably only have the micro layer. To you, creating a dungeon is just making a list of rooms that each have some interesting content in them, and then stringing them together in a mostly-arbitrary fashion.

But the macro layer is the elements which are continuous across the rooms, defining the experience as a whole. Things like keeping track of time, remaining light supply, exhaustion effects, and, especially, navigation and getting lost. Think about it. The concept of "getting lost in the dungeon" is one of the most classic tropes we associate with the genre. But it's a challenge that can only exist at the macro layer. The only way in which you can be "lost" is if you've lost track of the connections between rooms. You typically can't get lost within a single room (unless it's really big), but you can definitely get lost if you can't make sense of the relationship of one room to another. The decisions the party makes about fighting monsters, solving puzzles, and avoiding traps can each be tied to individual rooms, right? But the decisions the party makes about navigation, by definition, are not tied to a single room.

So a procedure provides the macro layer challenge. It gives a through line of gameplay decisions that'll affect the crawl as a whole. It'll characterize the whole dungeon and it will shape how their experience in each subsequent room is different than it would have been otherwise if they had reached it sooner. But it can't really stand on its own. You still need the micro layer, which will always be the main event. The good stuff.

And if you haven't guessed by now: the macro layer is provided by game design, and the micro layer is provided by level design.

Thinking in these terms opens up possibilities. For one thing, you can begin thinking of dungeon challenges that are tied to the macro level instead of the micro level, like having the dungeon slowly fill with water or having some kind of "growing paranoia" effect the longer the players spend in the dungeon. Personally, I like having traps be tied to the macro level more than the micro level. People will give all sorts of advice on making a good, elaborate, and creative room trap to throw at your players. But I'm generally more interested in simple, quick traps that are spread throughout the dungeon, and all fit into a greater pattern. It makes logical sense: the kobolds know how to make their poison dart trap cheaply and quickly so they keep reusing it. And if it has a consistent pattern to detecting its presence and stopping it, then the players will be rewarded for paying attention to each previous trap and using it to avoid the next one. It's a challenge that relies on their experience with the trap again and again throughout the dungeon. It's a challenge that transcends individual rooms. Neat, right?

But thinking in these terms also helps you think more critically about what doesn't work in a dungeoncrawl experience. I've heard some folks completely dismiss a game for "doing dungeons poorly" when, in reality, they were just playing in a shitty dungeon! And this is going to be a crucial part of any answer to the "wilderness adventure" problem in 5E D&D. You can try to make a procedure that accounts for random encounters, getting lost, foraging, hazards, difficult terrain, and so on. But even the best wilderness hexcrawl system will likely suck if the hexes have boring content and the map doesn't have any interesting features to interact with.

Some games don't provide levels to use, but instead the tools to create levels. For example, Ben Milton's Maze Rats is famous for its random tables. Many, many people buy the game just for those tables alone, not even playing the game itself. It's got a ton of good fuel for generating content. Level design fuel.

So if you're gunna make a wilderness procedure, then maybe you also ought to make a good catalogue of encounters to put into it. 5E occasionally does something like that. But it's more along the lines of "1d20 random sites you find in the woods" and each one is one sentence long and has absolutely nothing to interact with. It'll literally be stuff like, "You find a big standing stone with ancient writing carved into it" and that's it. That's not gunna cut it. Would you buy a dungeon where the room key just says, "this room has monsters in it" or "this room has a statue" and that's it? There's gotta be something gameable and toyetic and challenging to interact with. And if we draw the comparison again with the combat procedure, then arguably there could be a good reason to create a "monster manual" equivalent just filled with "interesting and potent wilderness encounters and challenges" that you can seed your adventure with.


Think about the balance. Some people like simple rules that don't get in the way, and instead want to focus all their energy on making good-ass content to play in. Other people really like systems and challenge that's derived from the rules themselves and how they apply. I would say that an RPG with really good game design can be considered a "strong" system when it's doing a lot to create interesting gameplay just by merely using the rules in their most basic form. A good magic system is usually interesting and challenging to use on its own. A good slot-based encumbrance system can always provide a little bit of extra challenge to deal with when used no matter what's happening in the adventure. A kingdom-management game can often be quite self-sustaining in its challenge just with the built-in limitations of how little the players can accomplish vs how much they need to accomplish to stay afloat.

So the main takeaway for game design is that you shouldn't consider a design problem to be "solved" simply when you've had one good session of gameplay taking place in that space. Because lots of folks will "solve" a problem like "wilderness adventure" or "urban adventure" or what have you just by creating a list of scenes they'll throw at the players specific to their current storyline, and that's it. But that's not an actual, sustainable "solution" to a design problem. That's just a one-time script you wrote. You solve these things when you can make a reliable set of rules or procedures that'll always work for playing any session with content in that space. If you made an animation of Mario jumping around and you had the player tap the button alongside it like they're playing Dance Dance Revolution then you haven't really created a platformer. You solve the platformer problem by figuring out how Mario always moves, regardless of the context.

Meanwhile, the main takeaway for level design is that you'll always need some amount of it, no matter how strong the system is. There are lots of folks thinking about game design all the time. But there aren't as many folks who'll put that effort towards the accompanying level design we need in order to appreciate their game. Too many people go online and request recommendations for a system that will magically meet all of their needs when, in reality, the solutions will be found in settings, adventures, and other forms of level design.

You can make a great open world campaign that'll fall flat if you haven't seeded the world with interesting places to find and interesting things to happen to the players.

You can make a great detective/mystery procedure to use that'll fall flat if the mystery itself is boring and the clues aren't well thought out.

You can make a great naval combat minigame for your pirate campaign that won't save it from inevitable doom if you don't have tons of piratey things to do and piratey islands to visit and piratey NPCs to interact with and piratey problems to reckon with. 

Go forth with this insight and hopefully improve whatever it is you're working on or thinking about.



  1. This is a useful post, I hope it gets spread far and wide.

    It seems that a lot of the friction between classic, indie and contemporary traditional design space would be mitigated if there were a stronger understanding of how mechanics, design principles, and ethics of play work better or worse for certain styles of adventure and demand different adventure design.

    When running 5E for example I've given up on dungeon crawls using classic Turn and Random Encounter based risk reward, because 5E's not built for that. A 5E dungeon doesn't want detailed exploration in the same way a classic one does because the system supports encounter based rather then level based design. Set piece tactical combat within a narrative/scene structure rather then navigation/exploration with risk v. reward and tension building procedures built around turnkeeping, supply and random encounter as threat.

    I'd like to think if more folks would just be curious about the design space their favored game is working in and what adventure style works best for it there'd be more comity and a better cross pollination between playstyles.

    Anyway, great post.

    1. Your comment has given me ae lot to think about. Especially in regards to the in-development series of posts on wilderness adventuring, I've increasingly run into design issues that can't be reconciled with 5E D&D, but could definitely be options for another RPG designed with the system in mind.

      I don't want to give up on 5E because it's my favorite game, even if it isn't perfect for EVERYTHING that I enjoy in gaming. But you're undeniably right. The best version of X activity will come from a game that's built for X activity, so if you really want to do X activity then you'll always have a better time playing in a system that works with you rather than against you.

    2. I think Wilderness might work better then Dungeoncrawl in 5E - simply because wilderness exploration has traditionally been "scene based", lacking the step by step description of a dungeon crawl. It is more granular with longer "turns".

      5E the WotC way is a scene based game largely, and I think that's why something like Eyes of the Stone Thief (13th Age) or Castle of Mirrors (3pp 5E) with their limited maps and zone based design (also reminds me a bit of Trophy Gold and its scenes, which is interesting). A Classic wilderness encounter is likewise a scene, a discrete (often necessarily so as they are randomly generated) encounter without spatial connection to the other encounters (though some narrative ones may exist or form). Interested to see what you do with it.

  2. I agree with Gus that this is a fantastic article. I love the distinction between macro procedures and gameable pieces of content.

    I also think Gus is raising a good point about "scene-based" design versus "room-based" (or "hex-based") design. 5e's balance is ostensibly designed around 6-8 medium to hard encounters per long rest, and we see the effect of that in resources like spell slots. However we can also see how this design doesn't seem to be supported in almost any other part of the game, including pre-made adventures. Perhaps viewing 5e as "scene-based" would solve this tension.

    As far as wilderness procedures go, I've personally gone to great pains to develop such as system. You can find it here, in Chapter 8:

    While several kinds of gameable content are incorporated into the design, it stops far short of a "monster manual" equivalent, which I agree would vastly improve the level design. It's also hex-based, because I wanted to avoid the classic trapping of players knowing that there'll only ever be one encounter in any given travel montage (i.e. a scene-based approach). I'd love to know what you think of my system. Perhaps further development would bear fruit.

    1. Let me write out my thoughts on your system as it's currently written. I have come across similar ones before. It's even quite like my first attempt to make a wilderness procedure for 5E. My comment will be split in half.

      One of the big dividing lines in hexcrawl procedures is whether a procedure expects the players to choose their destination hex-by-hex or if they should plot the whole route up front. Yours leans pretty heavily towards the latter. Even though it acknowledges the possibility that the party might just "generally explore an area," the procedure doesn't seem to account for what they'd do if that's the case. I guess you could allow them to simply make a stop at every hex so they can think about where they'll go next, but it doesn't seem built for a "fog of war" situation where they don't have a map going in (or rather, they only have a BLANK hex map going in). Being able to do both of these kinds of adventures, both the "travel from point A to point B" AND the "explore an area you don't know anything about and you don't know where you'll end up" is, in my experience, incredibly difficult to design for. Obviously the latter adventure is not going to be nearly as common as the former unless your whole campaign is about exploration, but it's a frustratingly common blind spot I find. Still though, just having this procedure to bolster content for those "travel from City A to City B" sessions is a huge improvement.

      Your system largely assumes that gritty realism will be in effect (as I read elsewhere in the document), and reinforces it by withholding long rest benefits outside of civilization. I LIKE this a lot, but it's yet another step removed from vanilla 5E D&D. 5E is designed with the assumption that you will get a long rest every single 24 hours, and theoretically, I'd love to see a good wilderness system that can work WITH that design choice rather than against it. And yet... The default assumptions of vanilla 5E, in this case, are probably just bad ones anyway. Nobody on earth actually does 6-8 encounters per day, they do 1 to 2 (3 at MOST), so your altered assumptions are probably just better design than the vanilla ones. That said, the "no long rests in the wilderness" rule is a significant enough change that, if used, means you are functionally not playing 5E D&D anymore. At that point, you're playing a different game. Maybe a better one, and maybe one that 6E D&D should pay attention to, but definitely something quite different than what new 5E players are probably signing up for.

      A note on hexes: I once wrote an article about a couple weird properties of hexcrawls (you can find here: and I think it's interesting to note that 1: your system falls cleanly under "Version B" as described in that article, and yet 2: it does largely concern itself with travel speed and various speed modifiers as being the single most important variable of the procedure (which I'm personally not crazy about). One of my issues with how it's presented here is that the table for "Hexes Travelled Per Day," organized by terrain type, seems to assume that you'll be traveling in the same terrain type for the full 8 hours. What if we travel by vehicle on the road for 3 hours and then reach the dangerous swamp, where we'll continue on foot at a slow pace for the next 5 hours to keep an eye out for the many sneaky goblins trying to ambush us and the secret entrance of the dungeon we're trying to find? How do you figure out the total number of hexes travelled? I guess you could calculate it ("the vehicle-on-road speed is 6 hexes/8 hours, so if you're only doing it for 3 hours then it should be 2.25 hexes, round down to 2...") but that seems like a pain in the ass.

    2. Meanwhile, an EXCELLENT decision is the emphases on activities. Designating activities while traveling is, I think, the key to solving most of the design challenges with making wilderness travel into an actual adventure. It keeps everyone involved, it incorporates challenges other than speed of travel, it can give a meaningful role to different character types, etc. And allowing the Ranger to do 2 activities instead of 1 is, I think, one of the most elegant ways to automatically boost the Ranger class as head-and-shoulders superior to other classes in this arena without allowing the Ranger to trivialize it entirely (in my own RPG you can find on this blog, I included this as a Ranger ability as well).

      I do find it a bit confusing that the "Support" and "Motivate" actions get significantly harder the higher level you are. Presumably, one's skill bonus will be increasing as well but I find it weird that low level parties have an easier time getting good rest in the great outdoors compared to experienced parties.

      One of my other issues is the procedural generation of Monuments, Weird Locales, Lairs, and Strongholds. Don't get me wrong, procedurally generating the world and allowing the DM to be just as surprised by the session as the players are is a noble design goal. But it's at odds with the decision to frame wilderness adventure from the initial player decision of "pick a destination and your route to get there." I don't know about you, but most of the kinds of destinations that me and my party would be interested in seeking out would be things on the Monuments/Weird Locales list (especially the ruins!) or Lairs and Strongholds. But if those things are only introduced into the world randomly in the course of the travel procedure, then I can't know about them and their location at the start of the journey when I'm picking my destination. See the problem?

      Otherwise I mostly just have good things to say about it. Aside from clearly being a tremendous amount of effort, you made a lot of choices that I think boost the experience significantly. Terrain has many meaningful effects, a lot of tools are provided, information is gathered and presented better than in vanilla 5E, the actual process of using these rules would definitely make the act of wilderness adventure seem "real" the same way a dungeon is, rather than a handwaved montage, etc. But this is indeed a system that seems to want to provide the entire experience, the WHOLE adventure, from a source of Game Design rather than Level Design. The Level Design comes into play when you actually make the hex map itself, and I assume that you'd also sprinkle at least a few pre-designed bits of content into those hexes, but this is an impressive effort towards trying to make a robust set of rules that can provide the adventure on their own.

    3. Thank you for such thoughtful feedback! This is best feedback I've received on these rules by far. I'm going to try and reply holistically to everything, but let me know if I miss anything important.

      You're absolutely right that my rules weren't handling hex-by-hex exploration, or fog of war. You're also right that my speed rules were a bit too detailed for something that isn't much fun, and that calculating the speed for multiple terrains required too much math.

      I've iterated on the design to address all of these concerns at once. The speed table has been massively simplified, and I've returned to the standard "hours per hex" format. I initially thought this format required too much math, but the "hexes per day" design ended up creating more problems than it solved.

      The article you linked got me thinking about different hex sizes, and so I've included a conversion rule to handle other sizes. I've also included a new "one hex at a time" variant rule for fog of war gameplay. It's not as clean as it would be if the system was designed for it from the ground up, but as you say, that style of gameplay isn't as common, so I think making it a variant rule is a good middle ground.

      Your comment about searching hexes is an interesting one. Overall I think my system works well with content premade by the DM, but I see what you mean about "picking the route" not making sense if the party is searching. I've rephrased Step 1 of travel to better handle this. It's a small change, but a powerful one I think.

      All of the above changes also resulted in Step 2 being merged into Step 1, and a new Step 6 being created for forced marches. I think the resulting procedure is a lot cleaner, especially for the different hex sizes and fog of war rule variant.

      I'd love to know what you think of the changes, if you have more time. Your feedback has already been immensely helpful though, and I really appreciate it. The level design is still missing pieces, but I've yet to think of a solution that isn't a tonne of work. Maybe rolling on multiple random tables and combining the results is enough, but I suspect "hex stat blocks" are necessary to do it properly.