For A Game To Be Considered A Roguelike

Within the roguelike genre, there is a lot of arguments over what concepts truly describe a roguelike or make up a roguelike. A lot of these arguments like to think of roguelikes as adhering to a rigid set of features for the game to meet the criteria of being a true roguelike. Other games which simply use only parts of the standards are at best just roguelike-likes or roguelites - something entirely different and not a part of the genre; or so is said.

For the most part after reading a lot of these arguments, I have to say I disagree with a lot of them. Besides some of the staple points being completely out of date with the time, they really don't feel like they get to the heart of what roguelikes are all about; while in some situations those 'roguelike-likes' seem to touch the inspiration of what Rogue is - a game that even the developer can play and not know what is going to be coming their way and doing so in a fun way.

The definitions given also aren't very developer friendly either, stifling the ability of developers to produce games for others to enjoy because they want to create a game that is a roguelike. This is rather poor because developers should be encouraged to incorporate innovative ideas, and the definition can’t be so constricting to regulate games in a genre to being Rogue-clones.

As such this write up is meant to add my own opinion of what a Roguelike essentially is defined as and to provide a definition that gives working room to future roguelike game developers to make new games to play. The approach I take for my definition is to be as inclusive as possible of existing games within the genre and to try to take into account a little bit of future proofing as the roguelike genre evolves.

Potentially my definition may even included some games under the roguelike genre banner that some people may disagree with belonging there (Diablo, Desktop Dungeons) but I’ll note that I consider the genre to be rather broad and that it consists of many, many sub-genres. Ultimately, the goal here is to provide a clear definition with concrete lines which help set roguelikes apart from other games - not to draw concrete lines in defining certain roguelikes against each other.

About Defining Roguelikes -

Before continuing into the definition itself, I figure a quick note regarding a couple of important questions are in order - What Purpose Does Such a Definition For Roguelikes Serve and Who Am I To Try And Form Such A Definition?

To the first question, the purpose for defining what roguelikes are is to determine what sets them apart from other games. Additionally, while not everyone has to technically agree with the definition, a definition is only as good as the number of people who agree with it. Hopefully this write up will help in forming a definition that people can not only agree with, but if they don’t can work into a structured definition that eventually will lead to the definition people can agree with.

To the second question, I am PC Gamer. I have played a great deal of computer games over my life, including a great many roguelikes. I also know a bit about their history, and the relation of influence that roguelikes have had on some other notable games. Unlike others who may have put forward their own  'interpretations' or 'definitions', I haven’t actually developed a roguelike either - a key component that has really gone missing from such definitions, until now.

Now, before continuing into my definition, I will also note that you do not have to agree with my definition, or at least, not agree with every specific part of it. While this definition is formed from my own opinion of what defines a roguelike, I fully admit that there may be some aspects which people may feel do not belong in defining what a roguelike is.

The main purpose for posting this write up is to help advance the discussion on what the most clear definition on what a roguelike really is may come to be; not necessarily that this specific definition ‘is’ the most clear definition of what roguelikes are. So feel free to agree with the definition entirely or to take it and remove or add to the definition as you feel is necessary to make a new one; perhaps a definition everyone can mostly agree on can be formed.

What I Define As Core Defining Factors -

What You Play As:

You play as a distinct entity (character) that has a physical existence in the game's world; the entity doesn't need to be living or even sentient and an entity could consist of a party of people (multiple characters) if the game is centred on that party and not any specific person in the party. You could be an inanimate object, an artificial intelligence, or some little puppy wandering the street. Regardless of what character you are though, you need to be in full control of all thought processes when it comes to action making for that character - at least in single player.

Note that by thought processes I refer to non-reaction type behavior of the entity (referred to hereafter as character). By this I mean that if you are allowed to move around, the game can't at anytime take away that behavior from you - outside of cutscenes, transitions between one level to another where the game features more than one level, or where in the game you have had your moving appendages cut off to prevent you from moving.

The game is also allowed to have reactionary behavior, by where the game will always make your character, or attempt to make your character react to something - for example, if you move into an unlocked closed door, the game may always make your character automatically open the door; or if your character moves into a staircase to change the level, your character will automatically utilize the staircase and change the level.

It may also be possible for more than one player to exist in the game world and play with each other. However, there are few instances as to multiplayer roguelikes, so at this time, the only true requirement for multiplayer roguelikes is that other players have to be their own distinct entity that is also distinct from every other player, or all players must play as the same distinct entity when playing the game (such as where the entity is a party of characters, and multiple players play different characters within that party).

Regardless of what you play as, you don’t head a major faction within the game; as opposed to taking the reigns of major groups residing in 4X Strategy Games, directly or indirectly as may be the case. You don't control a military power either like in a grand scale Real Time Strategy Games where you control a sizeable part of or the entirety of a massive army where you have direct control over your armies movements.

You also aren't an invisible force, manager, or god that can manipulates the world regardless of whether you are in the world or not; though for clarification, you could be a god with an avatar that is limited in what actions can be exerted on the world, as the actions must all originate through and are localized to the god's avatar. Ultimately the game mechanics and strategy boil down to what actions your character in game can do in a localized environment.

In a roguelike game you simply play as your character, taking on the role of whatever it is your character is doing. The game is usually won by achieving the goals that the character is pursuing, though this assumes that those goals are achievable - obviously if the goal is simply to get as deep as possible in an endless dungeon with an adventurer for example, then there is no winning the game, it simply ends when the game ends or you simply stop playing.

Note as well that in regards to goals, if the goal of the game is simply ‘just get as far as possible’, it may not necessarily belong to the roguelike genre. To be specific, if the game is just about travelling, racing, or moving forward non-stop then it is likely more apt to classify the game as part of the Endless Runner genre.

Level & World Design:

Progression in the game is randomized with non-static levels via random map generation/world generation and enemy placement. This doesn't mean however that the game can't include static elements (in-game vaults), and it is possible to have a roguelike consisting primarily of many static elements with a minimal amount of random elements connected them or altering them - the key thing is that the game is random enough to change when and where the player experiences the game's levels so that they can't necessarily predict how their run will unfold from the start of the game to the end of the game.

It is also allowable that a Roguelike always have static levels which are set to always appear in the same place, in every playthrough, but these elements should be minimal in occurrence and not take away from the randomized nature of the game when included. The player should have minimal idea of where the static levels are going to be found as well, though the player can know when they will be found - ie, the Boss Level will always occur at Dungeon Level N and the Town Level appears on the same place in the World Map at Location X.

In regards to the levels the player transverses, the makeup of the map that the player plays on can have defined static parameters that random world generation can’t exceed. However, there needs to be enough room within those defined parameters that the player shouldn’t necessarily have a definitive guess where they need to go; or least guess what pattern of rooms, corridors, or obstacles lies between them and known static elements of the game.

Notably in most roguelikes, levels are broken down into tiled segments. While not every roguelike needs to use tiled segments, where they do one can expect them to exist on a grid. In general, every tile in the grid corresponds to a different part of the level; with every tile containing some element of the world - that being as simple as a floor, a wall, a door, a misc object, or containing the player or the enemies he faces. Generally, each tile that is enterable by an enemy or the character can’t be entered by another enemy or character.

Depending on the degree that tiles are used though, it may be possible for tiles to be designed so as such that it consists of many subtiles instead. Subtiles may be on a 1 to 1 ratio with the player character’s avatar or the avatar of enemies he faces; though the avatar of enemies or the player character may also be big enough to exist in more than one subtile. This rule may also extend to misc items that exist in the game world.

As a note, an individual avatar may be big enough that it it simultaneously exists in more than one subtile; or even cover more then a full tile. Big avatars such as these may also partially fill different tiles where the game engine allows it. Alternatively, an avatar may be smaller than a subtile, and thus only take up a portion of a subtile. Multiple enemies that are also small enough may also exists in one single subtile.


The representation of elements within the game world is depicted through distinct images or markers of some type; an element referring to some physical aspect which makes up something within the world. These distinct images may only represent one static environmental feature; a specific enemy, an inanimate item, etc. This representation need not be graphical, and there is usually an emphasis to depict the game world as simply as possible, such as with ASCII Art; though there is no hard rule that depiction can’t be more complicated.

As a rule, the distinct image used must follow the rules of the physical aspect it follows - which means that if representing an enemy, players can notably expect all other enemies represented by that distinct image will act in exactly the same way - assuming they see all of the actions that the enemy in question is capable of performing. Note though that this representation for an enemy can be pretty broad; where classes or weapons with different properties are involved, it is possible for an enemy to have a wide berth of actions available to it.

It is possible for a physical aspect to adopt new rules where its function allows it to change when certain conditions are met, or if the rule of the physical aspect is to 'lie' to the player initially. For example, where high grass might be represented in the game, if a player walks on top of the high grass it may meet the condition of turning into a new physical aspect - trampled high grass, which comes with a new distinct image and ruleset for what it now represents. Another popular physical aspect could be that of a door - which could have the properties of being open or closed, or even destroyed or shattered.

Also worth noting is that roguelikes can have transformative enemies or mimic type enemies. Transformative enemies could be the likes of a cocoon which can transform into a butterfly or an enemy with the ability to sacrifice itself to summon a demon. Mimic type enemies may 'adopt' another distinct image to match that of another physical aspect in the game world - until such a time when they decide to ambush the player and turn into the distinct image of a mimic.

As the character progresses through the game, the character may progressively change. This can be through the character gaining experience, items and equipment, mutations, traits, and whatever else the game might do to advance the character. The game may assign how the character may progressively change automatically, but usually the player is given the option of deciding how their character changes as it progresses through the game.

It is not the case however that when the character progresses through the game that it has to change. It is possible that any changes that occur may in fact not be the better for the character - such as the character picking up cursed items that hinder the character or diseases which hamper the character from performing. It is also possible that the character may get weaker through dwindling resources. Usually some negative changes are preventable, but it is also possible that the character is guaranteed to change for the worse as part of the gameplay.

The game environment that the character transverses is also subject to change. This may be through new appearances of enemies or different classes of enemies, through different tactics used by the enemies from predecessors, through the game's natural environment becoming more dangerous or unpredictable, or any number of changes that can be thought of be the developers of the game. Much like the character as well though, the game environment need not have to change and may generally remain the same.

Pacing & Thinking:

When not under constraints due to a time limit or another player, it should be noted that the player is free to set the pace at which they play the game at. Players are free to play as fast or slow as they wish, and this may also even include pausing to consider one's options. However, the game can include constraints such as time limits or making even mundane activities have an associated cost to them.

For example, in a turn based game, it is allowable that just entering into the player's inventory takes a turn, or equipping an item to use, and so forth. Players may also have a time limit before the game automatically ends the player's turn as well - such as in orientation to multiple players. The game can even be run in real time, but the key thing is that whatever the constraint is, there must be an element of decision making allowable for the player to decide their actions.

Throughout a roguelike game, there is a heavy emphasis on decision making rather than reaction; though this doesn't mean there can't be elements that rely on player reaction to function. Roguelikes are thinking games, and the gameplay must have elements that stimulate decision making and give time to the player to pause and think before making decisions; though this doesn't mean the player has to be given the option to make a decision at every moment or for every single activity within the game itself. In relation to reactionary actions and decision making though, the decision making has to be of sizeable importance to playing the game.

Derived from decision making, there is also an emphasis to consequences that the character suffers as a result of the decisions the player makes; whereas if the player does well, their character gets rewarded and if they screw up, their character gets punished. Commonly, the most frequent iterations of this is permanent failure and lack of saves, where your character only has one life and should the player’s character die they are prevented from going back to somewhere in their run that occurred after their run began before they died - though this doesn’t mean that a player can’t have the option to save their game and end a play session and then come back later to pick up the run from where they ended their play session last.

However -


So long as the game accomplishes some semblance of consequence to the player's actions through other means, it is possible for a roguelike to forgo permanent failure or not allow game saving. There is no ‘hardcore rule’ within the roguelike genre or one that requires games within the genre to be heavily punishing towards the player. If the developer of a roguelike game wishes too, they can forgo the need to have permanent failure or no saving points in their game.


Ultimately, regardless of any other factor referred to here in my definition, the most important factor that determines that a game is a roguelike is in regards to replayability. This refers to restarting the game from the beginning and playing the game again as well as replaying specific parts of the game.

Reasons to replay the game can be varied. A player may replay the game to try a different starting setup where the game allows it, or to replay their character differently if it isn't limited in how it can be played. The player may replay the game because they are looking to accomplish a specific goal of some sort, whether that be related to the actual game or not, both within the aspects of the game itself or perhaps in relation to scores where points are tallied in some shape or form. Where permanent failure or losing the game is involved, the player may replay the game because they have yet to beat the game.

Regardless of why the player might replay the game, replaying the game should be for the most part 'a unique experience' as compared to previous runs the player may have experienced while playing the game. This should even be to the point that if the developer of the game was to play their game, they themselves wouldn't necessarily know what to expect when playing the game. Specifically, each time the player starts playing a new game, there should be something new or interesting, capable of catching the player off guard, or presenting a new/different challenge for the player to overcome.

When replaying the game, it is allowable for some element of persistence to exist between previous and subsequent game sessions. However, this persistence must be strictly controlled to prevent early game sessions from just being foundation blocks for future sessions to succeed. If an early game session is unwinnable and is only useful for unlocking content, future resources, or a means of winning in a later game then the player isn’t so much restarting the game but rather continuing the previous session.

It should also be noted that something as vast as a game world which you already played in during a previous game may also be allowed in roguelike games. This may include a vague semblance of a previous game session’s world but with rearrangement involving subsequent session’s worlds or a world where every miniscule detail from a previous session is saved and a new character can begin. So long as the subsequent game session feels distinct from its predecessor, the game may still be a roguelike.

Another example of persistence between current and future game sessions that may be present in roguelikes is unlockable content. Unlockable content may refer to features in the game that become available permanently from the point on, being available for the current session and/or future sessions to benefit from. This may include new starting options for subsequent games that the player may start or for unlockable features that may exist in future games that the player can use or interact with through their characters. Unlockable content can run the gauntlet in terms of scope and can range from just being a simple item recipe to additional content in the overall game that you can play.

Regarding The Berlin Interpretation -

I likely couldn't post this without relating at least to the 'first' attempt at making a definition for roguelikes, which was the Berlin Interpretation. For those who don't know or weren't aware, the Berlin Interpretation was the first 'attempt' at creating a definition for what is a roguelike game. Unfortunately however, it failed to give a definition that covered the vast depth of what roguelikes are - heavily in part due to the fact that it was biased as to what is a roguelike game on the part of those involved.

My opinion of the Berlin Interpretation is that it is a 'good' definition to define a 'Traditional' or 'Classic' Roguelike game; which is to say it essentially covers a sub-genre that is nestled within the actual roguelike genre as I have described it. The Berlin Interpretation however can’t cover the the entire roguelike genre itself as it is too narrowly defined and centred on games which the authors of the interpretation represented ‘the canon’ of roguelikes.

Specifically when looking at the Berlin Interpretation, a key failing with it is that it doesn't try to define the 'essence' of what a Roguelike is; and worse it takes an exclusionary nature which people have used to cast off other roguelikes as being 'less than' roguelikes and not a part of the roguelike genre - ie, roguelike-likes or roguelites.

Another key failing is that it doesn't try to really be a definition, as definitions are exclusionary in nature. The Berlin Interpretation really fails to get across what really excludes a game from being a roguelike, and even gives a wishy washy disclaimer. As such, one may be free to not even consider the Berlin Interpretation to really be a definition in the first place for defining one game as a roguelike while coming back to use it to define a game as not being a roguelike on another occasion. Personally, I prefer not referring to the Berlin Interpretation as a definition at all - especially to describe the vast set of games that make up the roguelike genre.

All in all I don’t think much of the Berlin Interpretation. It comes off as far too elitist to me when read and it looked like more of a list of ingredients for making games with - which is honestly the wrong approach for forming a definition for what is or isn’t a roguelike. Roguelikes aren’t about what physical aspects make them up, but rather what the game itself orients the player to do or not do. Aspects that roguelikes have are about accomplishing a goal to how the game will be played and not about just having a physical aspect for the sake of having it.

Also of note is probably one of the worst sins of the Berlin Interpretation: the group of people who came together consists of only roguelike developers - and no roguelike players. Discussion likely would have benefited a lot more with a little player input as how a player interprets a game. Players playing the game will usually have quite the different opinion as opposed to how a developer does - and the same can be said for how they define the games they play as well.

On Games Questioned As Belonging To The Genre -

Whether or not a game is a good roguelike or not in view of someone playing it is not a reason for it not being included within the roguelike genre. Roguelikes as a whole cater to the taste of many individuals, and their is a vast amount of tastes that roguelikes can cater too - some of which may interest one person while completely turning off another.

Just because one game that fits the criteria I gave above doesn't cater to the interests of a person or group isn't a reason to not call the game a roguelike. "So long as a game meets the factors used above in my definition for a game to be a roguelike, it can be considered part of the roguelike genre.”

A key example that people like to bounce around for example is in the case of Diablo, which was developed by Blizzard Entertainment. Many dispute that Diablo is a roguelike, however it fits into the definition I gave above - which means that it is part of the roguelike genre under my definition.

As a note though, one is free to argue that diablolikes however aren't roguelikes. To give some examples, Dungeon Siege II and Titan Quest do not include randomized levels, so they aren't roguelikes. However, I leave it to another article to define what diablolikes are.

Final Thoughts

As some general advice to roguelike developers, if the run of your game takes 5h+ from start to finish or really can’t be beaten within one sitting, you should include other means of playing your game outside of permanent failure and not including save points. Roguelikes needn’t be masochistic and the norm should be for any game developer to avoid making games that hard.

At the same time, roguelike developers who opt not to include such things as permanent failure and not including save points should look into including game mechanisms which can still facilitate penalties for bad decisions on the part of the player. It is also possible to include a way for the player to lose the game as well, even if such features as permanent failure and not including save points exist - though ultimately one should take into consideration whether players losing the game is desirable.


Written by Davion Fuxa, a frequent roguelike gamer and Let's Player on YouTube

DFuxa Plays -