Once a relatively obscure sub-genre that appealed to a hardcore fan base, roguelikes have been enjoying a recent renaissance. Let’s take a closer look at their central design elements, and see how different games are making use of the same underlying architecture.
So What is a Roguelike?
Until very recently, the term "roguelike" narrowly referred to games that were like Rogue, a 1980 Unix dungeon crawler that became the archetype of the genre it named. "Roguelike" was used to indicate that a game incorporated the primary mechanical features that Rogue had. The aesthetic of Rogue (that of an adventurer attempting to reach the bottom of a dungeon) was widely copied as well.
Since Rogue and its direct descendants were hack-and-slash RPGs, the unique design elements of the games tended to be conflated with aspects of the RPG genre they otherwise belonged to. This was formalized with something called the Berlin Interpretation, which outlined the defining aspects of a roguelike. These include some of the most important features:
- Procedurally generated environments
- Permanent death
- The randomization of item identities
The Berlin Interpretation also lists turn-based combat, grid-based environments, and killing monsters as key features, which are more closely aligned with the specific kind of RPG roguelikes originated as. As the structure of roguelike has expanded into genres beyond action RPGs, those RPG design elements have tended to be dropped or ignored.
The primary three features (procedural generation, permadeath, and randomization of items) have found use in a variety of different types of game, and are strongly associated with the term "roguelike". So, technically speaking, "roguelike" refers more to a shared set of mechanics and design principles than a genre in and of itself (though I'll continue to use 'genre' as a shorthand).
As a result, games from a huge variety of genres have made use of the roguelike design framework. It's a set of design elements that can be used equally well in action platforming (like Spelunky), exploration and base-building (like Dwarf Fortress), tower defense (like Dungeon of the Endless), top-down strategy (like FTL), stealth (like Invisible, Inc), first person shooters (like Eldritch), and traditional dungeon-delving RPGs (like Dungeons of Dredmor).
Procedurally Generated Levels
Procedural generation indicates that elements of the game, especially the level design, are not hardcoded in advance by the designer, but instead are generated randomly from a template each game. As a result, it means the player cannot memorize beneficial locations or places to avoid, and so must play each game blind.
Though the specifics of the map are randomized, the general form of the map stay the same each game. So, in a traditional dungeon crawler like Cardinal Quest, the map is always rooms and corridors, and similarly in the sci-fi FTL, a sector map is comprised of beacons linked by lanes. It is the specific arrangement of these links and nodes that is randomized each game.
Procedural generation has two main benefits to the overall design:
First, since the field of play is different every time, it increases replayability as the map is always unknown. This frees the player from having to memorize a level’s specific challenges in and of themselves (as one might do in a challenging platformer like a Megaman game or Super Meat Boy), so the player must attempt to adapt to the pattern underlying the challenges. While you can’t ever know for sure what’s behind the next door, you can develop a sense of what might be behind it.
Second, it ensures that every playthrough is an exploration of a new space. Since the map is never the same, the player will always be plunging into the unknown, whether it is their first game or their fiftieth.
Exactly how random the levels should be requires substantial testing and care from the designer. If the placement is too random, the difficulty between games may vary too drastically, with some being very easy and others being unwinnable. Depending on the complexity of the game, it can be an intensive process to ensure all possible random maps are fun and able to be completed. The best designed roguelikes make the player feel punished when they make mistakes, not when the game itself makes mistakes or generates a map that's unfairly difficult.
The difficulty in perfecting the maps increases the larger the levels are. A famously complex example is Dwarf Fortress, where a vast and detailed three dimensional world with biomes, minerals, creatures, civilizations and entire histories is generated each game. (Replicating just a fraction of these with graphics instead of ASCII was part of the inspiration for Minecraft.) Most games take a more limited approach by segmenting the map by level or area, so the generated areas are relatively smaller and thus easier to fully test.
"Permadeath" generally means that the player must restart from the beginning upon their death, and that they can not reload a save to reattempt a failed challenge. It’s this feature that often intimidates players unfamiliar with the genre. Kicking the player back to the beginning after every death can seem unduly punishing, and contrary to the design principle of interactivity.
But a game designed with permadeath in mind is built
differently than one that assumes the use of saves or lives. A typical
game will be built with portions of the game the player might see only
once (like easy initial levels), and balanced to have difficulty spikes
intended to require several attempts. Conversely, a roguelike is built
with repetition in mind, and so while death is still a failure condition, restarting is much less onerous than it might be in another game.
While permanent death can seem jarring at first, it does bring some design benefits. It essentially redefines the failure state of the game; rather than expecting the player to play through the game once, while re-attempting challenges they fail at, instead the player is expected to play through the entire game without dying once (a tall order!). So "death" isn't a temporary setback, but rather part of the cycle of play. Since the vast majority of attempts will result in death, it's less "You Lose", and more an invitation to "Try Again", while hopefully applying what you learned.
Most importantly, permanent death gives visceral meaning to each and every playthrough. Without the safety net of reloading or lives, the action of a roguelike can become heart pounding as the player's position becomes precarious. The stakes that permadeath create is why it's the most iconic design element of a roguelike. Every decision made is given critical meaning, because, unlike in most games, the player can't simply have another go if they make a mistake. This can certainly be stressful and frustrating, but it also makes success feel much more meaningful and hard-won; few games can match the feeling of accomplishment that roguelikes offer.
Though the cliché environment for a roguelike is heroes delving into a dungeon, the gameplay of surviving on limited resources against the unknown lends itself very well to a horror setting. Survival is a natural partner for the roguelike architecture of real consequences and a world you can't memorize, and dying is scarier when it carries more weight. Don't Starve is a good example that fosters an atmosphere of horror with its roguelike mechanics acting in concert with its ominous Edward Gorey inspired art direction. The Binding of Isaac's dark thematic elements are similarly bolstered by the underlying roguelike mechanical architecture.
Discovery and Limited Retention
This is a feature that has the most substantial variation in use. In a traditional roguelike, the only thing retained between games is the skill of the player. In NetHack, death means not just losing all acquired equipment, weapons, and levels, and losing the current random dungeon layout, but even also losing knowledge of any discoveries made during play. For example, a bubbly potion might heal the player, but in the next game it might contain acid instead. The Sega Genesis classic ToeJam & Earl uses power-ups in randomized presents in a similar manner.
Modern games don't tend to strictly implement this (especially since not all games have item attributes that could be sensibly randomized), but do tend to adhere to the spirit of the design. The intent is that the only thing the player keeps from game to game is their own increasing skill and knowledge of the system. Always needing to identify potions in NetHack or Angband is just an expression of that design. This ensures every playthrough is meaningful and unique, which procedural generation and permadeath also support.
There's been substantial experimentation with this aspect of roguelikes. A popular variation is to include unlockable game elements, which can help the player track their progression. In Spelunky there are multiple unlockable skins for the main character, but they confer no advantages besides different aesthetics. In FTL, though, there are 28 ships to pilot, each with different strengths and abilities, but only one is available to start with, and making incremental progress or achieving certain goals gradually unlocks the others. This would seem to violate the principle, as earlier runs have a demonstrable impact on latter ones (by making new things available), but it's really just a way of slowly unveiling the full feature set as the player becomes more engrossed. NetHack has 13 character classes available from the start, but it wouldn't dramatically change the game if some needed to be "unlocked" in some manner, since you can't play more than one per game anyway.
Some games may include more meaningful progression. In Risk of Rain, in addition to unlockable character classes, achieving certain goals unlocks the ability for items and powerups to appear in future playthroughs, and making them available can become necessary for continued progress. Systems like this may blur the line of not retaining any advantages, but their effect is generally small, and the main advantage the player gains is their own increased skill. This can make the challenges of a roguelike more palatable, because instead of one very difficult goal to achieve (beating the game in a single playthrough without dying), it gives the players some incremental challenges to beat as proof of their progress. Though this may dilute the premise slightly, it's a concession that gives players a tangible sense of progress they can visualize, since their increasing skill level is more subtle.
Some games may even ignore this entirely, however. Rogue Legacy features a progression system that allows you to purchase incremental skills and benefits (like higher HP) for each subsequent attempt. So while the game is otherwise very much a roguelike action platformer with procedurally generated levels and permanent death, the game requires repeated attempts to slowly build in-game advantages, in addition to building player skill. The upcoming title Temple of Yog takes a similar approach, wherein the player's success in the roguelike combat portion is rewarded with incremental advantages to the player's village and to subsequent runs in the dungeon. These games further stretch the definition of roguelike (Rogue Legacy identifies itself as a "Rogue-Lite" for this reason), but they still maintain the core familiar gameplay of repeated meaningful and unique attempts with gradual progression.
The defining aspects of roguelikes are instructive not for their exact features, but the style of play they give rise to. The roguelike architecture brings out the best in games which require repeated and iterative attempts at success. They raise the stakes for such play through permanent death, and ensure each attempt is a new experience by randomizing the layout and having no (or limited) carryover of advantages from previous games. It's why such a diverse range of games can make good roguelikes, from puzzlers to platformers. And since roguelike mechanics can be rich additions to games of any type or genre, why not try making one yourself?
List of Games Mentioned
This is by no means an exhaustive list of roguelikes, but these are the games mentioned as examples above:
Cardinal Quest, 2011
- Don't Starve, 2013
- Dungeon of the Endless, 2014
- Dungeons of Dredmor, 2011
- Dwarf Fortress, 2006
- Eldritch, 2013
- FTL: Faster Than Light, 2012
- Invisible, Inc., 2014
- Risk of Rain, 2013
Rogue, 1980 (playable due to preservation by the Internet Archive)
- Rogue Legacy, 2013
- Spelunky, 2009
- Temple of Yog, Upcoming
The Binding of Isaac, 2011
Subscribe below and we’ll send you a weekly email summary of all new Game Development tutorials. Never miss out on learning about the next big thing.Update me weekly
Envato Tuts+ tutorials are translated into other languages by our community members—you can be involved too!Translate this post