Roguelikes: The Misnamed Genre (2011)

The author obviously did a lot of research, explained roguelikes' roots, and has an interesting take on something that a lot of the great roguelikes have in common. It was a good read.

Also, if you read to the end of the article, they aren't saying to use their definition of roguelike, since the name is already taken.

I've never cared much for the berlin interpretation (it focuses too much on the specific mechanics rather than the more abstract design principles than one would naturally associate with a genre), but this article's definition definitely bends far too strongly into its abstractions... to the point that almost every game is defined under his new term, and games that no one would put together... are put together.

IMO a roguelike's defining feature is the "fairness" of its simulation (most of the world's rules are equally applied to all its denizens.. though some may be more equal than others with custom spells/abilities/etc), in combination with the inevitable ADnD dungeon-crawling and RPG elements. The grid-based, turn-based, modality, procedural generation, etc of the berlin are too pointed to be relevant IMO.

Though procedural generation is a little interesting because I feel like it inevitably falls out of any game targeting a "fair" simulation -- how much good is a hand-rolled map design if you can't also hand-design the mechanics and interactions with the entities in the map? So it naturally lends itself to ceding control over the world generation itself.. and of course if your rules are fair and sane enough that the player can use them with their uncontrolled and arbitrary strategies and combinations.. then they're probably good enough to randomly smash together by the machine. So the monsters gain their flexibility, and everything just works -- so why not. The only issue is leaving the AI flexible enough to deal with the fact. And permadeath kind of just falls out of the procedural generation mechanics as well

Thus its not that roguelikes are an endless interaction with complexity.. its that the game-universe itself operates more closely to a proper simulation of the universe (not a detailed sim, mind you, but not a sim that simply revolves around the player(s) ala megaman), and complexity of the interactions are a natural outcome. And then differentiating between say a colony-sim (DF), a 4X and a roguelike is largely on the basis of the scale at which the player interacts with the simulation (colony, country, individual, respectively). [Note: you also have actual simulators, but they're clearly a different kind of interaction -- In DF the simulation is in the backseat, whereas in EURO TRUCK SIMULATOR 2 the simulation is the main point]

Of course you could probably differentiate even more within that.. but there are so few games that actually make a real effort at any kind of "fair" simulation that the game name alone is sufficient. Which is also why Noita feels like it belongs with the other roguelikes, despite violating most of the berlin interpreation -- it implements a "fair" simulation.

There's probably a better term for "fair".

Yeah, I'm with you on this one. "Roguelike" was, for more than twenty years, essentially defined as a game that is turn based and grid based in the vein of Rogue, centered around traversing procgen levels and usually with permadeath and dungeon crawling as key focuses. It wasn't until game journalism sites and their fans started going "um... roglik is when hard..." that use shifted, and the current use describes basically nothing.

I think the Berlin interpretation puts too much weight on periphery aspects of the genre, but this shit is just ridiculous. If you use the standards these people do to define Roguelike with anything more popular, it becomes plainly obvious how absurd it is. You could call the Warhammer 40K tabletop a Doomlike because it has chainsaws, demons, the ability to be customized, scifi guns, the firing of said scifi guns and space marines, despite not even being a first person shooter or a video game.

> Roguelike is a well-defined term, though there are multiple competing definitions. This article presents none of them.

Isn't multiple competing definitions the opposite of well-defined? How overlapping are those definitions?

Simplest definition possible:

Try Nethack and see if you "get" it. If you do, congratulations! You now know know what a roguelike is.

Topic is pretty dead horse at this stage, it's from 2011 which makes it slightly more reasonable. The raw tonnage of gaming discourse since then is astonishing.

Dark Souls - probably one of the most discussed-to-death games ever - hadn't even come out yet, for instance.

Not even a single Skyrim "I took an arrow to the knee" meme

There is something to be said about all video game genres being wrongly named.

I mean when you look at some romance written in first person, that takes place in Victorian era, you don't call it a First Person Victorian novel. You call it a romance novel.

Art is usually named after feelings it evokes and less on the actual techniques used.

Maybe puzzle/idle games are only games that actually have descriptive names. You play a puzzle game because you want to solve a difficult problem. You play a shooter because you want some carefree violence.

In lieu of that, Roguelikes definition misses the mark completely.

You can have a Roguelikes experience without any of the stuff defined in Rougelike Berlin definition. Noita is real-time; Elona/Caves of Qud allows save and load; Elona has little procedural generation outside wilderness and random events;

Haha, man, it's weird when a post you made ten years ago, that isn't even linking to its main site (see http://www.mandible.net/2011/04/25/roguelikes-the-misnamed-g...), hits the front page of something somewhere.

For what it's worth, I was absolutely aware of the Berlin interpretation and was giving it about as much attention as I think it deserves. In my opinion the Berlin interpretation is not a sane definition of a roguelike, it's a point-by-point description of what the common roguelike of the day was. It's a ridiculous gatekeepy way to design a genre; just imagine the definitions we would have got if "first-person shooter" had been defined in the wake of Doom ("a first-person shooter involves using a set of six to twelve weapons, bound to the number keys, to kill demons") or in the wake of Battlefield ("a first-person shooter involves large team battles with matchmaking and gradual unlocks of sidegrades, which can be bypassed by paying real money").

But the point of that post wasn't to attack the Berlin interpretation, rightfully or not, it was to trace the outlines of a concept that I think a lot of people aren't isolating. The existence of the Berlin interpretation - which was a lot more popular back then - was, and to some extent still is, blinding people to the interesting parts of the genre. The interesting part of an FPS is not the exact number of weapons available; the interesting part of a real-time strategy game is not the fact that units move smoothly, and not on a grid; the interesting part of a roguelike is not whether it's made with ASCII art. It's an implementation detail.

My attempt in that post, made kind of hamfistedly as anyone would say about something they wrote over a decade ago, was to get at the interesting parts of a roguelike and try to draw a loose outline around them.

If I were to write this today, I'd spend a lot of time talking about the XCOM 2 expansion pack, which I think made some fascinating steps towards being this kind of game. Yes, specifically the expansion pack.

In the base game, you can more-or-less plot out your build order before you start the game; you can know what will show up at every step, you can know when you'll get certain items, you can roughly predict when you'll get certain abilities (modulo soldiers dying, of course), there's no unpredictability.

In the expansion pack, though, they throw a bunch of random events at you that can seriously influence your build order. Maybe you get double-speed research, for a single research item, that you must either cash in immediately or abandon; this can mean you're encouraged to pick somewhat-random equipment. You can also be given the opportunity to buy weird unique items that are permanent upgrades to, again, somewhat-random equipment, so maybe in one game shotguns are really good and in another game it's sniper rifles that get the random bonus. Again, this can change your strategy significantly. And finally the big enemy monsters have random traits - they're weak to things, strong to other things, and all of this is re-rolled every game.

This isn't an enormous amount of variation. But it's significant, and this is the thing that I'm pointing at; situations where the player is left thinking "wait, hold on, how do I navigate this? how can I put these bonuses and penalties together into something really awesome for me?"

I think on Steam games that take clear inspiration from Rogue (so top down, turn based, maybe ASCII based) are tagged Traditional Roguelikes. I personally like the Roguelike/Roguelite split you outlined but I can understand why people get confused by the names. I agree that permadeath + procedural generation is such an amazing mechanic combo that it needs to have its own name to easily be described when you present it into new base mechanics or game form.

Roguelites unlike Roguelikes (the lite and like parts are self-explanatory), have some progression outside the permadeath runs. The Binding of Isaac, Dead Cells, Hades... are all roguelites, due to having this permanent progression (unlocking new items, skills, increased statistics...) that carry on to new runs. Roguelikes don't have a progression outside of individual runs, so each run is a fresh start.

People who are unfamiliar with the terms just end up using them interchangeably.

Same.

People yelling about how it needs to be a turn based dungeon crawling RPG to count as a roguelike are like those dudes still mad about how eSports aren’t even actual sports.

> For me, roguelike = procedural generation + permadeath

So any game with a random map lacking a progress-system is rogue-like? Like Civilization, Sim City, Star Craft? No limit on the genre or type of interaction?

> Some people call these games "roguelites"

Rogue-lites have a progress-system that grows outside the sessions. So every run has meaning beyond the run itself.

The thing is, in many cases "Roguelike" is not a helpful description. It's like calling Megaman an "R-Type-like" because it's a side scroller with a beam-up weapon. The things that make Megaman what it is, have nothing in common with R-Type even though they share a couple of game mechanics.

I got really confused when people called Spelunky a roguelike. If my friend called on the phone and asked me what Spelunky was like, I'd say "A Rick Dangerous style platformer with procedurally generated levels" - and he'd have a pretty good idea of what I meant. If I said "It's a roguelike" he'd probably think I was talking about Dungeon Crawl Stone Soup, or ADOM or something.

> For me, roguelike = procedural generation + permadeath, that's all.

By this definition, Minesweeper is a Roguelike. :-)

Procedural generation + permadeath includes Tetris and Solitaire.

I just try and avoid the whole word nowadays. The Berlin interpretation was a fairly decent guideline when there were just a bunch of turn-based tile-based dungeon crawlers with randomized levels, many of them with text-based graphics and/or developed under early free software models. Then Diablo came along, brought a real time aspect into it.

Personally, I think the procedural generation aspect is overrated when it comes to the genre definition. Say you fork Rogue and hardcode the level layouts, item and enemy placements and so on. It would probably be an inferior game, but still in the same genre of game, I think. At least the game would change less than if you made it real time or somehow removed the grid lock.

I tend to say something like "turn-based dungeon crawler". This way I can talk about games that are like Nethack or DCSS without having to clarify I don't mean platformers, twin stick shooters or -- god forbid -- deck building games of all things. I like action games just fine, often more than turn-based dungeon crawlers, but the itch they scratch is a different one.

Those two are just design principles and I completely agree that they are great together. However, those two don't form a genre. The classic roguelikes e.g. rogue, nethack, brogue, dcss, doomrl etc. very clearly are a genre. They're very similar yet all do their own thing and do interesting things in the confines of the genre. Modern roguelikes aren't a genre because there's so many variations that it's meaningless. Risk of Rain 2, Dead Cells, Hand of Fate, Binding of Isaac, Ziggurat. They're all roguelikes apparently. But they all play so differently. What's the point? It's like saying Contra, Super Mario Bros, Battletoads are all in the same genre because if you die a couple times you have to start from the beginning.

I think procedural generation, permadeath, and you can unlock tools or advantages that you can take between attempts.

Loop Hero, FTL, etc are examples

Hm, considering X-COM, the only missing important piece is 5. Non-modal. Hack-n-slash and grid/turn-based are not inherent to the X-COM genre, and are broken only in geoscape anyway.

There's also a bunch of low-value factors there: 2. Enemies are similar to players, 3. Tactical challenge, and perhaps 6. Numbers.

This is something that hasn't clicked for me before.

Your link discusses classic or traditional roguelikes.

OP's blog post is from 10 years so I can't blame them. But today "Roguelike" is usually something slightly different.

A roguelike these days is a game that has (at least some) procedurally generated content and permadeath without meta progression.

A roguelike purist would call these roguelites.

It's a shame "being so wrong as to be misinformation" is not a flaggable offense under HN guidelines.

I ran it in the past couple of years, maybe with DOSBox; I can't remember exactly.