Hi, I'm a game dev interested in all sorts of action games but primarily shmups and beat 'em ups right now.

Working on Armed Decobot, beat 'em up/shmup hybrid atm. Was the game designer on Gunvein & Mechanical Star Astra (on hold).

This is my blog, a low-stakes space where I can sort out messy thoughts without worrying too much about verifying anything. You shouldn't trust me about statistical claims or even specific examples, in fact don't trust me about anything, take it in and think for yourself 😎

Most posts are general but if I'm posting about something, it probably relates to my own gamedev in one way or another.


🕹️ My Games
boghog.itch.io/
🎙️ Game Design Vids & Streams
www.youtube.com/@boghogSTG
☠️ Small Updates + Dumb Takes
twitter.com/boghogooo

Warning : I'm a rabid, unreasonable metroidvania hater. If you're a fan of the genre prepare for venom.

I used to think I was into metroidvania games, but the more I played them, the more bitter I got & the more convinced I became that metroidvania is perhaps the worst genre in gaming. It's a genre that consistently fails to take advantage of its interesting design elements, and constantly devalues and butchers itself.

These games turn exploration into staring at a map screen without engaging with the environment. These games turn discovery into a shopping checklist. They reduce level design into repetitive hallways. If you want someone to despise exploration and beg for 2D beat 'em up level design, just get them to play nothing but metroidvania games for a couple of years.

But among that sea of trash, there's absolute gems like Rain World. Games that deliver on every promise of the metroidvania genre so well that to include them in that pile would be an insult to Rain World itself. So instead think of it like this - there's the stinky diseased dumping ground that leaks into other genres called "metroidvania", and then there's the shining pristine beacon of hope called [ World Mastery Games], my platonic ideal of a metroidvania. And Rain World sits at the top of that golden tower.

The idea of world mastery isn't really new and Rain World certainly isn't the only game to capture it decently well, even if nothing quite matches it. These sorts of games have existed for quite a while in different forms - STALKER being a really good example, especially with some mods like Call of Chernobyl, some mods that add more enemies/stalkers and Misery. Resident Evil 1/RE1make have aspects of this too thanks to their severely limited inventory, limited resources & the open ended mansion with a lot of different routes (& Crimsonheads soft-blocking routes in RE1make). Fan made randomizers in games like Zelda 1 and LTTP shift the games towards world mastery.

Metroidvania games themselves actually gain these elements during some self imposed challenge runs such as speedruns, but the potential is mostly wasted because developers treat the genre as disposable trash.

Metroidvanias are pain

World mastery games and metroidvania games share many of the same elements. They have big interconnected environments with interesting gating that makes the routes you take more involved than simply picking a direction and holding forward. They have a mix of gameplay styles - exploration, platforming, combat, perhaps stealth, perhaps RPG elements. They have sequence breaking, they have backtracking, they have ability-based gating of both the soft and sometimes hard varieties.

Metroidvania games use their worlds as content delivery mechanisms - their job is to connect one piece of prebaked content with another piece of prebaked content, and perhaps add a few more access points than you'd have in a linear game for sequence breaking. Different elements exist almost counter to each other - the games have backtracking as a defining feature, but never do anything with it and as a result have to constantly undermine it with shortcuts. They don't actually want you to go through the same area more than a time or two. "Exploration" itself is constantly undermined with a really detailed map and obvious ability gates, which are often color-coded. The genre is afraid of its own conventions.

Slugcat

PURITY

World mastery games aren't ashamed of the elements that make them up and embrace them, trying to flesh them out. They aren't trying to hide their backtracking, they tend to give meaning to it, either within the same run or across several runs. They will have life simulation systems, indepth movement mechanics & layered levels that let you take advantage of them, challenges that remain tricky even if you've seen them a bunch of times, interesting resource economies, randomized elements that force you to rethink your approach, constantly shifting goals, survival-esque time pressures that force you to branch out carefully or make reckless mistakes, and other such stuff. They will embrace exploration - they might abandon map screens entirely, or reduce the information said maps give you.

The games are built to reward knowledge of the world itself rather than any specific level, or even mechanics. The better you know the layout of the levels, the more efficiently you can do what you need to do like gather resources, avoid enemies, come up with plans on the spot, etc. The world design itself tends to be deeply integrated into combat. In STALKER this is mostly manifested as the amount of cover you get, some anomalies you can use to your advantage, terrain/buildings you can use to abuse the game's stealth system and ambush enemies. In Rain World, the stealth depends on terrain, and the viability/effectiveness of combat is directly dependent on which weapons/objects you're carrying or have next to you and what sorta terrain you're dealing with. In both games you will play better if you have a robust, highly detailed mental map of the world itself.

Needless to say, roguelike-inspired survival games don't quite fit the bill either - they have a certain degree of world mastery as well, but you absolutely need a stable world underneath all the randomized elements. Otherwise you're mastering systems & mechanics, not the particular, hand crafted layout of a world. Zelda 1/LTTP randomizers are a better representation of the style than RL survival games. The games need enough randomized elements to constantly force players to engage with different routes through the game & as much of the world as possible, but not enough to undermine the planning aspect.

Just wanted to do a messy rant. Will probably expand on this in a dedicated Rain World review, after I play that game enough. So far game's shaping up to be one of my fav ever, giving me exactly what I loved about STALKER in a more refined, interesting & condensed form. Go play it.


You must log in to comment.

in reply to @boghog's post:

i find it interesting that despite being very popular, metroidvania as a genre is so stangnating and flawed. more than half of these games are basically linear where you go forward, find the dead end, go to the only place that is not a dead end, get an ability and go back, which means that the whole world behind is non-existent to you. the other thing some metroidvanias do is give you absolute freedom right at the start by creating a set of non-interconnected fully unlocked stages that you can visit whenever you want. the problem is that such games don't force you to use your abilities in certain locations and making any decisions is pointless since you can clear all stages no matter what you've done before.

Because it's not a real genre, it's an adaptation to a world of Bigness. It's basically a marketing gimmick that lets indie developers sell their platformers to a mainstream audience who sees traditional platformers as "lesser", that's all.

World Mastery Enjoyer Gang represent. Been meaning to play Pathologic for like 12 years now LOL isn't it a bit more focused on the narrative stuff though than STALKER or Rain World?

God, I will never forgive what EA did to Mirror's Edge. The first game certainly wasn't perfect, but it had some neat ideas to build off of.

Taking a linear game that slowly ramped up in how you chain your abilities, to open world slop in Catalyst. They didn't even place things that give you XP in interesting spots requiring creative usage of your tools. Having to steal microchips (standing still for 2 seconds like you're inspecting a Counter Strike skin) in a game supposedly about going fast is insane to me.

Hm. I don't think you did your best on this article. I think you should have defined exactly what a World Mastery™ game is, and provided examples of why Rain World, and STALKER are world mastery games. (also, I think immersive sim is technically the genre here? Would Breath of the Wild and Tears of the Kingdom be world mastery games? or DXMD?)

Dark Souls 1 is considered by many (including me) to be a Metroidvania, but only with keys, rather than ability upgrades. And it's an undeniably solid game all around.

Hollow Knight is a metroidvania, and it emphasizes exploration by only filling in your map when you die, or sit at a checkpoint, and by having your location on the mapscreen only be shown optionally. This means while you're exploring, you can't see the rooms you've passed through, and have to rely on dead reckoning and memory to tell the layout of rooms. Hollow Knight has the issue where it's arguably a bit too easy and simple for most of the game, but the delicate flower sidequest ties it together by outright forbidding the larger shortcuts, and pressing you to find an efficient route across the map.

I can certainly agree with the criticisms that a lot of Metroidvania games let moment-to-moment level design fall by the wayside in the name of a wide explorable world. Just look at Super Metroid and Symphony of the Night (look at the long empty hallways!), both of which are a lot easier than the games that surround them (even if they make amazing speedgames). But there are also a lot of very solid Metroidvanias out there, and a lot of demonstrations that exploration can be complex, and the level design doesn't need to take a hit.

Overall, I think this article needs to define its terms better, and cite more concrete examples. I think you're assuming we know what's in your head too much, and gesturing at games that did it right more than explaining what the actual difference is.

You're right I didn't, I just really hate metroidvanias.

It's tricky cause I think this "genre" mostly exists as a kind of untapped into design possibility space that you can see when playing specific games in specific ways. There are very, very few games actually designed to work this way from the ground up, Rain World aside. It's actually kinda shocking that this space hasn't really been tapped into much by devs, given how fun randomizers are & how popular survival games have become.

Let's see though, so as far as metroidvanias go, they need :

  1. Interconnected maps with a bunch of different routes
  2. Discrete "chunks" within said maps with discrete entry/exit points, which separates them from open world games
  3. No distinction between "dungeon" and "overworld" & minimal or no "levels" - you might have a game where levels/dungeons have a metroidvania-esque design, but the goal of MV is to create connected worlds so this kinda segmentation works against it.
  4. Backtracking - you will have to go back in one way or another to get access to new stuff, usually via convenient shortcuts but not always.
  5. Gating - you have to meet certain pre-requisites before getting access to new areas, devs use this to create a stable progression/difficulty curve. Traditionally the games have ability-gating but personally I don't consider key-based gating to be meaningfully different. Gating can be hard or soft - it can be a knowledge or execution check, not just an item check

So Super Metroid, SOTN, Dark Souls & Resident Evil 1 are all metroidvania games in my mind. King's Field would fit the bill, a whole bunch of imsims would fit the bill too. Risen & Gothic would probably qualify too, but been a long time since I played.

World Mastery Games have all of that, but also :

  1. Randomized threats or goals that force you to take different routes & go to different places whether you want it or not. Either in a macro sense (go to area x instead of area y because the randomized quest/progression item is there this time!) or at least in the micro sense (lizard/crimsonhead camping near the entrance to this area, better go around!). Additionally, the player's starting area might be randomized too
  2. Mechanics that force backtracking unless the player actively works to mitigate it by Mastering The World™ - limited inventory spaces (RE1), time pressures (STALKER's emissions, Rain World's rain). Forward momentum is earned, not expected.
  3. Thorough intergration of player abilities and the surrounding level design (or maybe a resource economy). Players must lack something really critical and need to be looking for it in the world, and losing it. Maybe this can be survival mechanics like a hunger meter. Maybe their full moveset can't be available at all times. You need ammo to shoot in RE1/STALKER which is found in levels/bought with money gotten through stuff you find in levels. Can't attack enemies without spears/rocks/cherry bombs in Rain World, which can easily be lost. These can't be permanent either, it defeats the purpose

I don't wanna be too specific with implementation here because I think the abstract idea is what's important - get players to engage with the world as much as possible. Build everything around the idea that it'll be played not once, not twice, not ten times, but hundreds or even thousands of times. Integrate the world into the gameplay & learning curve as much as possible, do not let players get by on just their knowledge of player mechanics. Don't let them settle on too stable of a route, don't let them get comfortable camping in an area, force them to explore as much of the game as possible. Give them challenges that test their knowledge of the world. I'm absolutely sure that there are ways to achieve this that I can't even imagine but smarter devs will nail.

Make Metroidvanias But Good. And good in ways only these games can be, not a "oh it's fun action game that happens to have the metroidvania structure cuz normal action games don't sell as much anymore".

So TOTK, BOTW and Imsims would fit the bill when played in a particular way, if players create a mini game for themselves by, say, constantly coming up with new challenge runs that force them to go to different places, creating some sorta resource economies, that sorta stuff. It's one of many cases where players are far ahead of gamedevs - they try to twist their lackluster games into a more interesting form.

I mean, I think this could be summed up as, "Metroidvanias would be more interesting if there were threats that traveled across the map, and occasionally blocked off routes, and if you didn't have consistent access to the same tools, so you need to vary your solutions to various problems."

I can see what you mean about randomizers. Players have developed such a generalized mastery over the game, that you can shuffle all the rooms around, and players find ways to connect the pieces.

I think you're onto something, but statements like, "Forward momentum is earned, not expected," are value judgments that sound nice, but don't really mean or count for a lot. It's easy to take a hardball attitude. I think having some emotional distance helps, ie. "This is something limiting the genre, here's some things we could do to expand the genre." I think permanent upgrades are just fine for one. It's a choice, not a necessity.

Also, I think Super Metroid and SotN aren't honestly the best metroidvania games. I think a lot of other games surrounding them are better (Zero Mission, AM2R, Harmony of Dissonance Portrait of Ruin, Order of Ecclesia), even when they're more linear.

Yea forcing a generalized/holistic understanding of the games is basically the best non-specific way to put it. It makes sense anyway since that's a lot of stuff in games is designed - if you make a beat 'em up enemy you don't just give it a preset seed and put it in the same scenario over and over again. You wanna randomize its behavior and have the player fight it in very different contexts, so they learn how to deal with different "sides" of said enemy. Worlds aren't any different.

And yea you can def make really cool platformers, action platformers, action RPG's or what have you in the metroidvania format, but the style has so much potential that hardly anybody's tapping into or even talking about much, that it's been a persistent frustration of mine for the past decade. Thankfully between randomizers & Rain World, nature is healing

We haven't been seeing much innovation or refinement in action platformers, to be frank. It's hard for Metroidvanias to rise above a deficiency in game design knowledge.

I think we need to do more work illustrating the basics. I've drafted an article on dynamic level design, but I'm publishing it through some people I've been working with, and the draft needs refinement before it's out.

Kinda off-topic: Even though I strongly disagree about Zero Mission and AM2R being better than Super Metroid, I cannot help but respect you representing the best Castlevania games. Especially Harmony, no one appreciates that one!

Have you played Kirby & the Amazing Mirror, and if so, what do you think of it?

I think Super Metroid is the best Speedgame of the metroid series and has the best nonlinear structure by far, but the enemies and bosses were weak, for the most part. And I consider the core gameplay to be a bigger deal than whether or not the game has interesting speed tech and nonlinear exploration. I rate Hollow Knight only a 7/10 for the same reason.

I have not played Kirby and the Amazing Mirror, but I'll put it on my list.

Would love to hear what you think of it if/when the time comes. The most concise way I can think to describe it is it feels less like a retail Kirby release and more like someone made a very extensive romhack of a Kirby game to turn it into a metroidvania: So it still operates on all Kirby rules, and that plus the diabolical navigation means you absolutely MUST learn and adhere to the rules of how a Kirby game works to progress.

It's the kind of game a lot of metroidvania fans would whine about, but that I love because the act of exploration and progression is actually resistive. Kinda reminds me of Harmony of Dissonance in that way, which you like, hence my mentioning it.