Engine Builders: an analysis
In recent years, a new, very popular genre has emerged: games heavily focusing on upgrades and synergies. People creating games in this genre usually do not know how to call it; they usually go with “roguelike”, because games that inspired them are also called so. However, the hook of this genre is not any roguelike element. In particular, Rogue was not a part of this genre!
History of roguelikes
Let us quickly recall some historical background of the roguelike term. ‘Roguelike’ became a category on USENET in 1993. It was meant to group games that have a particular style of moving the character and performing basic actions (note the similarity to other genres such as platformer or first-person shooter — which are also defined by how you move and fight): the game is grid-based, you can press a single key to move your character to move to another square or fight the monster there, other commands like spellcasting are done similarly — a very simple system that leads to fast, tactical, and complex games. Since it was difficult to find a good name, the genre was named “roguelike” after the first popular game in the genre. Roguelikes were usually created as passion projects by communities of dedicated fans who were dissatisfied by the low complexity and difficulty of commercial games at that time. The mainstream gaming press mostly ignored them. The site RogueBasin currently lists over 700 roguelikes, in several subgenres. Roguelike maps are almost always procedurally generated, and hardcore roguelike fans respected permadeath:
reloading your save game was considered cheating.
Spelunky (2008) was a platformer inspired by the procedural generation and permadeath of roguelikes. It was very innovative, platformers did not feature such elements until then, and lead to a wave of other “procedural death labyrinths”, popularly called “roguelikes”, with some sources claiming “roguelike” means that now: a game with procedural generation and permadeath. Roguelike fans found it weird, because these games were clearly not roguelikes, but had trouble explaining it.
But then, the thing became even worse for people wanting to play games similar to Rogue. Games such as Risk of Rain (RoR), Slay the Spire (StS), Vampire Survivors (VS), and Balatro. These games were called roguelikes, but they were not “procedural death labyrinths” either. RoR has no procgen, StS was a deckbuilder (and deckbuilders had randomness and permanent consequence by default), VS and Balatro had neither permadeath nor procgen in any reasonable sense. In 2024, the majority of games marketed as “roguelikes” seem to be games like this. This is very confusing for people who have played actual roguelikes, leading them to conclude that anything is called a roguelike now, that videogame devs are calling their games roguelikes just to confuse people, and things like that.
Fans of these games are similarly confused by the reaction of original roguelikes’ fans. That is because the games mentioned in the previous paragraph actually do have something in common — however, that common thing is not actually a roguelike element!
So what the common element is?
The common element is the focus on upgrades and synergies. In the video game world, this seems to be largely an innovation of The Binding of Isaac. Edmund McMillen did not know how to call his game, so he called it an “action RPG shooter with heavy Rogue-like elements”. Presumably “RPG” is there to represent the focus on upgrades, most people do not seem to think of The Binding of Isaac as an RPG. When I tried The Binding of Isaac, I expected a game that is fun for similar reasons that make roguelikes and Spelunky fun. I did not get what I wanted as its roguelike elements are quite weak, and I did not find it interesting. But people coming to The Binding of Isaac with a fresh perspective enjoy its focus on upgrades, they crave more similar games, and assume roguelikeness is about that.
RoR, StS, VS, and Balatro took inspiration from this, and their devs and fans also did not know to call them, so they also got called roguelikes.
In the boardgame world, there is an existing term with similar meaning: “engine building”. In https://boardgamegeek.com/thread/2959433/what-makes-an-engine-builder the most upvoted definition is:
My definition of an Engine Builder is a game allowing the player to craft reproducible combos triggered by the game’s basic actions and gameplay
which revolves around expanding and optimizing these combos. The basic actions are the fuel for the engines and the resulting actions and resources are the engine’s output.
So basically a game about upgrades and combos. Other definitions mention starting weak and becoming exponentially stronger during the game. Deckbuilder is a subgenre of engine builder, here the engine has a form of a deck you draw cards from; in an engine builder, the engine can be of any form. It is a bit vague, more based on vibes.
So we will be using this name engine builder in this post. Does an engine builder need to be strategical, as in, should they involve making interesting long-term decisions? In my opinion, no, although probably some fans of engine-building boardgames would strongly disagree with that. Good engine builders tend to be strategical. However, restricting the genre only to games that are good according to some community leads to gatekeeping and confusion. One other possible name would be upgrade fest — contrary to engine builder or roguelike, it makes the general idea clear. It might sound a bit pejorative, but then, we have genres named walking simulator or bullet hell. I have also seen Tanya X. Short refer to such games as synergy builders, which is nice too.
Relation to other genres
RPGs and metroidvanias also have upgrades in most of their definitions, does this make them engine builders too? No — their upgrades are more about making the game varied in the long run, and to provide the feeling of going on an adventure and growing, they are not the main focus. This includes traditional roguelikes (as a subgenre of RPGs), which are usually not engine builders any more than other RPGs.
In 4X games (e.g. Civilization) and factory games (e.g. Factorio) you also get stronger, but this also feels very different.
How are engine builders related to run-basedness? Since upgrades are the focus, it is difficult to design long-form engine builders, they are more fun when the process repeats from 0. (Again, this contradicts traditional roguelikes, which are often be very long games.)
How are engine builders related to the idea of “when you lose the game, you have to restart”? Well, if you lose the game, it basically means that your engine was not strong enough, so it would not not make sense to reload anyway. So a way to provide a benchmark you are competing against, but not the focus. (Tabletop engine builders do not need this because
you are competing against other players.) That is very different from the strategical risk management of a roguelike such as Dungeon Crawl Stone Soup (DCSS), and the execution-based arcade permadeath of Spelunky.
Another genre focusing on upgrades is incremental games (Cookie Clicker, Universal Paperclips). These seem to be fun for similar reasons. Many incremental games feature “prestige systems” (restart the game after completing some goal, usually with some permanent benefit), so basically the same idea of run-basedness. One difference is that incremental games usually cannot be lost; however, you can still compete against time for a similar experience.
When people say “Space Invaders but a roguelike” or “Poker but a roguelike” or “Tetris but a roguelike” or “Minesweeper but a roguelike” or whatever, they seem to actually mean an engine builder. Strategy and arcade games have permanent consequence and randomness by default (in strategy games this follows from wanting the players to make correct decisions every time instead of memorizing them, and in arcade games, from the randomness of user inputs). If deckbuilder is a subgenre of engine builder, what about Slay the Spire? Well, Slay the Spire is not a pure deckbuilder, relics (taken from Isaac) are a big part of the experience, so calling it
engine/deck-builder makes sense. Even more the case in Balatro.
How are engine builders related to roguelikes? Not at all. Rogue was not an engine builder. None of classic roguelikes are engine builders. It is possible to be both a roguelike and an engine builder, like in Path of Achra. But roguelikes do not need to even have upgrades.
While The Binding of Isaac was heavily inspired by roguelikes, its roguelike elements are not what made it fun. Games should be marketed by what actually makes them fun. Procedural generation boils down to placing separate, predefined rooms randomly, contrary to typical roguelike levels where distant parts of the map would interact. While the unidentified potions of Rogue were iconic enough to be copied by many roguelikes, unidentified pills in The Binding of Isaac were not as iconic. Permadeath is not what made it fun either, what made it iconic is its upgrade approach.
Game developers calling engine builders roguelikes claim that “you know what we mean”, but typical players are not as knowledgeable in recent trends as game designers; looking for a definition of roguelike, they would likely find a traditional one, a definition of a procedural death labyrinth, or something like “a game where you restart where you die”, which does not sound fun. Engine builder fans do not get what they expect when trying procedural death labyrinths (Spelunky, Caveblazers, Vagante, Noita) or roguelikes — while the most popular roguelike in 2008 was DCSS with its extremely well executed permadeath, today, roguelike designs such as Moonring, Elin, or Caves of Qud which do not focus on permadeath are gaining popularity, and this trend is likely to continue. Just like roguelike or metaprogression fans do not get what they expect when playing an engine builder. Hopefully this genre will eventually get a proper, clear name, be it upgrade fest, engine builder or something else, just like we no longer have “doom clones”.
