Let me try something here.
I have a lot to say about D&D, obviously. I’ve played it a whole bunch. But I, as a well-rounded human approximation with hobbies and favorite foods and the shallow husks of broken dreams, know about other things too. Some are hard to apply to a tabletop gaming blog (though I’ve posted about that in particular), but one that’s easy to relate is retro video games. If a video game was released before 1999, there’s a great chance I’ve heard of it and a good chance I’ve played it. I want to leverage some of this deeply unnecessary knowledge into DMing with Charisma, so we can start with a genre that (sort of, I guess, if you squint) kicked off forty years ago today: Metroidvanias.
At their core, Metroidvanias are the space in the Venn diagram between “exploration” and “progressive power gains”. It sounds like these should be easy to work into D&D, a game ostensibly about exploring things (dungeons) and fighting other things (dragons), but I’ve found it’s hard to get a Metroidvania-specific style of these things into the d20 ruleset. Metroid was one of the first things I tried to adopt into D&D, and I abandoned it pretty early because I couldn’t get what I thought were the hallmarks of the games or setting. Now, with the wisdom of bad decisions, I have a better idea of how I might tie things in, such that I probably could run a Metroid (or Metroid-inspired) adventure if not a full campaign.
I think it’s best to avoid major changes to the rules when possible. If you want a D&D Metroidvania, you want D&D by necessity, not a total conversion that changes most of its core functions. (That’s true of any system, but this is mostly a d20 blog.) You don’t need new classes or races or even weapons, not when you have the power of reskinning. You just need to hit the key requirements: a focus on rewarding exploration and progression via abilities granted by power-ups.
How to Exploration
A Metroidvania needs a map. Freeform, theater-of-the-mind-style exploration isn’t sufficient. You need that visceral satisfaction of seeing your progress, filling in places where there used to only be voids marked “here there be tygers and/or lava”. A good map is both a record of past actions and a guide for future actions, nudging you toward the empty or incomplete areas where progress and loot might still dwell.
But filling out a detailed map as a DM is a huge pain. In a perfect world, when the game starts you’d know the location of every boss, every power-up, every gate, every secret. That doesn’t work in a tabletop RPG, mostly because players have agency. They can go in directions you didn’t expect, bring spells that trivialize or escalate encounters, and casually discard power-ups that don’t fit their characters. You can’t plan for their exact capabilities like you can when the player is limited to what they can accomplish using four buttons and a D-pad. Even if you could, creating a battle-ready map of each room is a wildly unnecessary chore. Placing individual blocks is not an efficient way to spend campaign prep time.
Which leads into a related point: backtracking. In a 2D action platformer, players can breeze through rooms they’re already seen, ignoring enemies and obstacles they’ve already navigated, especially once they have new movement tech. That doesn’t work as well in D&D, where any combat, even if the players wildly outclass it or ignore it by dashing through the room, bring exploration to a screeching halt. Just getting players to roll initiative and sorting it takes more real-world time than many rooms in an action video game. The obvious alternative is to skip combat in any rooms the players have already cleared, which means backtracking is effectively irrelevant. If the players say “we go back to that green room with the rock wall” and that happens without cost or consequence, every room is a fast-travel point. You want to have the impact of backtracking without the massive time cost of playing out each individual choice.
For this I’ve been playing with the idea of a pointcrawl map rather than a typical blocky Metroidvania map. A pointcrawl, briefly, is an exploration system in which the setting is divided into distinct zones whose transition points are defined rather than as a single open world. You know players can go from the Castle Entrance to the Gatehouse to the Bailey South, but not right from the Castle Entrance to the Bailey South, and that’s close enough. You don’t have to build the exact dimensions of the Gatehouse and wedge it into the spot between the two maps, you just say “yep, you’re changing zones”. While people are probably most familiar with this in text adventures where there’s no need or ability to have a detailed map, it’s kind of everywhere. The Legend of Zelda’s map of Hyrule is technically a pointcrawl in a 16 x 8 grid, where each screen of the map is independent of adjacent screens except in how the player transitions between them.
Three types of video game maps, each valid for a Metroidvania campaign. From top: Super Metroid, Zork, Nightshade.
What this doesn’t handle, and what I don’t think any map system can handle elegantly, is gating. One of the great difficulties in designing a Metroidvania is finding a way to show the player the difference between a problem they can solve now versus a problem they must solve later. If a player sees a red door in Super Metroid, there’s no way past it without missiles. That’s a hard gate. But the player can get to certain areas early by using tricky wall jumps, which means the walls in the way are soft gates. D&D players have a tendency to treat everything as a soft gate, a problem they could overcome if they try just the right combination of intelligence, resources, and tenacity. “We’ll have to come back to this later” feels like a failure to solve a puzzle rather than an acknowledgment of the “appropriate” or “expected” way to progress. Any hard gate becomes a negative play experience that escalates the longer the players fail to realize it’s a hard gate, but a flashing sign that says “you need the blue key / ice beam / hookshot” denies them the ability to try, fail, and discover things at their own pace.
Seeeeeeeecrets…!
Similarly, secrets are a major part of fun exploration, and D&D is not great at them. A player can look at a rock formation and think “there’s something funny about this, I should check it out.” That rewards their observation skills and attention to detail. But when a DM says “there’s a formation of bombable rocks against the wall”, that’s painting a big arrow on the floor. There’s no satisfaction in doing the obvious thing. The opposite isn’t any better: telling the players nothing and waiting for them to ask “is there anything weird about the rocks in particular” is a great way for them to miss almost everything. A secret is effectively a riddle the map poses to the player, and you don’t want it to be a game of “guess what the DM is thinking”. Passive perception checks were created to avoid this hunt-and-peck style of checking every brick for traps, but that turns it into a simple binary where players find some secrets and miss others based on their build. The players don’t solve the riddle as much as their characters do, and then it’s not a riddle or a secret, it’s just a skill check to get loot.
So then how do you incentivize hunting for secrets? In a word, don’t. Don’t encourage players to hunt around for every little thing they might not have noticed if they didn’t specifically ask about it. Instead, tell them exactly what they see, but not what it means. “You feel a faint breeze coming from a crack in this wall” is a clue, and it’s up to the players to figure out how to respond. It might be a puzzle, it might be something that costs a resource (and this the players have to decide whether they’re willing to spend it here or return later when they’re in better shape), and it might be that they need to come back once they have more tools, like a drill that can break through weak rock or a hat that shrinks a character to one inch tall. Metroidvanias do this all the time with visual hints. A 404’ed website that discussed the design of Super Metroid put it beautifully regarding these three rooms:
All three depict a purplish cavern, with some type of stairs leading out of a water-filled hole, and the only hope of crossing it is by swinging from the grapple blocks in the ceiling. There is only one difference worth mentioning. The first two rooms depicted are tutorial rooms right after acquiring the Grapple Beam. The third is in Crateria, and it’s the place where the Grapple Beam needs to be used to further the game’s progression.
So what the game is essentially doing, is planting the Craterian room in your mind (adding it to that list of possible utilities) immediately after giving you the Grapple Beam.
If this all sounds too direct and easy, feel free to toss in a couple of red herrings, little pieces of verisimilitude that add depth to the world but don’t hide progression or power-ups. Ideally, put lore at the end of them so they’re not totally useless, but there’s nothing stopping you from giving the players a clue that legitimately has no solution. Speaking of Super Metroid, in one (and only one) room there is an enemy in an inaccessible area. Unlike every other enemy of its type, it follows Samus around and changes color based on her armor. Why does it do that, and why is it only in this one place? Shut up, that’s why. It’s just a fun little thing that happens without any implications. And it drives players nuts.
How Power-ups Work (and Don’t)
Power-ups are a major part of the Metroidvania experience. The general gameplay loop is “find a problem -> wander about -> find a solution -> use the solution”, usually nested several layers deep so the player has to track multiple problems throughout the game. If there’s no power-ups, if the player at the end of the game has the same abilities they had at the start, they’re not playing a Metroidvania. They’re playing Mario or Pac-Man. (I have a tongue-on-cheek theory that all video games are either Mario or Zelda, which ties into my serious theory of additive vs. subtractive video game randomizers, but I already used the Calvin’s Dad meme in this post.)
Luckily, D&D works this way by default. Characters do stuff, get experience points, gain new abilities, and do more stuff. The main issue is sustainability.
When Samus (Metroid) gets missiles, they’re limited, but she can get back ammunition by killing enemies. Alucard (Castlevania) has limited magic points, but they regenerate over time. Hornet (Silksong) has limited tool uses, but she recharges them at benches, which are dotted throughout the map. D&D doesn’t like any of that. An ability either works permanently or recharges on a short and/or long rest, which canonically takes time and intentionally can’t be performed rapid-fire. Worse, abilities tend to share resource pools. A wizard who can only cast two third-level spells per day has to consider whether flying past a hazard is worth losing a fireball, which is a real problem when your map requires flight for progression.
This falls in the same category as skill checks: if you want your players to do it, don’t make it onerous. Just like a mandatory puzzle clue or bit of knowledge shouldn’t be locked behind a skill check characters can easily fail, using cool new power-ups shouldn’t be blocked by the actions players want to perform. If you expect players to use a power-up regularly, make it at-will. Examples include Samus’s ice beam or hi-jump boots, items a player can toggle on or off and don’t cost anything. If you expect players to use a power-up occasionally, give it its own resource pool and a recharge mechanic appropriate to the expected frequency of use. Using spellcasting for progression seems logical but it’s effectively a tax on casters just for playing in the campaign. An item or boon with the power “once per short rest, the whole party can fly for one room” means the players won’t soar past every encounter, but also won’t have to turn back at the giant cliff face because they didn’t realize they had to ration their spell slots.
Mobility spells deserve special mention because the core loop of a Metroidvania exacerbates how tricky it is to balance them. Obviously, if a wizard can just fly and dimension door and passwall their way around, they’re going to skip many of the challenges and experiences you had planned for them. That’s not good for you and the effort you put into creating those challenges, but it’s also not good for the players if they use a spell to bypass a barrier and thus miss all sorts of cool things they would have found if they had treated the barrier as an incentive to look for power-ups elsewhere.
I recommend banning disruptive mobility spells entirely, contingent on buy-in from the players. If they’re playing in a Metroidvania campaign, you can reasonably expect that they want the feel of a Metroidvania. To some players, that means they explicitly want to break things open and skip whole sections of the game. For that I have little advice. It’s not how I like to play and it’s difficult for me to get excited about DMing a campaign where the players want to experience as little of it as possible. Rather, I tell the players during (or before) Session Zero that the campaign has certain expectations. “Look, there are spells that just make map design hard. Please don’t take them. If you see a problem and find a clever way to skip it that you know I didn’t expect, you will run the risk of missing some knowledge or lore or power-up and thus getting yourself into a bad situation later. I promise I will include ways for you to get past any problem the map puts in your way.” Most players I’ve had are amenable to restrictions that befit the genre. It also means that when they do get the “everybody can fly” power-up, it’s a major shift instead of “oh, just a bonus spell slot, okay”.
The Ideal Power-up
Metroidvanias—and games in general—have plenty of good, bad, and “meh” power-ups. It’s a lot like magic items in D&D where some loot lands and some flops. When the loot is tied to progress in the game, the downsides of disappointing items are even great. As I see it, good items meet at least two of these criteria:
- They solve a problem.
- They have combat application.
- They’re fun to use.
“Fun to use” is subjective, but I include “rewards cleverness” and “easy to handle” under that umbrella. Basically, any application of it that is handy, fast, or entertaining but not part of its required gameplay application. The more difficult it is to use an item as intended, the less enthusiastic players will be about it. Consider:
| Item | Problem-Solving | Combat Application | Fun to Use | |
|---|---|---|---|---|
| Bomb (The Legend of Zelda) Appears in most games |
Break destructible walls; open secret areas | High damage to multiple enemies | Speedrun tech / glitches | |
| Book of Magic (The Legend of Zelda) Never appears again |
None | Creates low-damage area effects (that hurt you) | Duplicates a low-level item; only usable with magic wand | |
| Varia Suit (Metroid) Appears, if not all, in most games |
Makes exploration safer, especially around lava | Halves damage taken | Visual change to character | |
| Long Beam (Metroid) Appears again only in the remake |
None | Hits enemies from farther away | Minor ease of exploration | |
| Leap Stone (Castlevania: Symphony of the Night) Appears in most games |
Allows double-jumping | …allows double-jumping | Look, double-jumping is just really good | |
| Mist Form (Castlevania: Symphony of the Night) Almost never appears again |
Pass through certain walls | Deal low damage (after upgrades) | Slow; allows minor sequence breaks |
These are somewhat extreme examples of high- and low-tier items, but I think it gets the point across. The point is that you don’t need every power-up to turn the game on its head or become a character’s go-to weapon and favorite popcorn seasoning. If you hand players something that gives them options and doesn’t penalize them too badly for mistakes or failures, they’ll find ways to use it.
Follow Your Heart (Container)
To reiterate, very little of this changes the core rules or assumptions of the game. You can adjust or redesign classes, races, backgrounds, weapons, all of it. But you don’t have to. Metroidvanias are about the way the player and the world interact with each other, not the round-by-round minutia of play or the lore and opportunities provided by the setting. You have all the freedom you want to make your own adventure or campaign, whether it’s in a video game world, your own setting, or—I acknowledge begrudgingly—Forgotten Realms.
In fact, if I had to pick a third thing that makes a Metroidvania a Metroidvania, it would be “a world the player wants to interact with”. That’s why people look fondly on Hollow Knight and not, say, Super Pitfall. (There are several reasons, but the fact that Super Pitfall punishes you for exploration is a big one.) Lore and setting encourage players to dig deeper, to learn about the who and why, to draw conclusions based on the little they know then act on those conclusions. It’s just that most games with a narrative work this way, including most D&D adventures, so I didn’t call it out specifically.
As always, the most important thing is to go by what you and your players want. If your party loves drawing their own maps, a pointcrawl may not be right for you. If your players love ridiculous sequence breaks, let them have all the mobility spells they can get (and maybe prepare an escape hatch for when they get in too deep). Both D&D and Metroidvanias are means to an end, and the end is a fun game.