Psychochild's Blog

A developer's musings on game development and writing.

14 July, 2016

What makes a good roguelike?
Filed under: — Psychochild @ 3:31 PM

Although I didn’t list it as one of my favorite genres, I do like roguelikes. They are pretty popular these days, too, with a ton of variations on the theme.

One of my friends said he was on a mission to own all the roguelikes and “roguelites” available on Steam. My good friend Dave Toulouse was looking at his recent game, March of the Living in the context of other roguelikes. We talked a bit about how fast you fail can make a difference in a good roguelike.

So, I figured I’d talk a bit about what elements make a good roguelike.

The classics

The name “roguelike” comes from the classic game Rogue, a game where a dungeon was created originally using ASCII characters to represent different elements. Levels were procedurally generated, with simple rooms connected by hallways. The game featured randomized items, allow for repeated play since items had to be discovered each time.

From this game other games were developed that were inspired by (or simply cloned) the game; the most famous of these early games was probably NetHack, a game that is the gold standard for deep and involved gameplay. Given the plethora of games inspired by the original Rogue, the term “roguelikes” was born.

In the modern era, we see games that took on a lot of elements of these early roguelikes, but applied them to different types of games. For example, the game Rogue Legacy takes the idea of procedurally generated areas, procedurally generated characters, and permanent character death and applies it with platformer type gameplay. FTL took a lot of procedural generation and permanent character/crew death and put it in space with interesting situations posed to the player. Some people prefer to call these “roguelites”, as they deviate pretty far from the original formula laid down by Rogue and the ASCII-based games that followed. I’m not interested in splitting that hair, so I’m just going to lump them all together; keep your holy wars to yourself.

Anyway, let’s take a look at the elements that make a good roguelike/roguelite.

Procedural generation

One thing that the early game Rogue did that surprised most people was generate the levels procedurally. Instead of having a static map, the map was generated using an algorithm that created rooms and connected them with hallways. This created a map that had to be explored each time.

But, it can be more than just the map. Rogue Legacy has the characters you select be procedurally generated, based on the attributes of your previous character. Traits can be passed down to your descendants. This adds an interesting twist to the usual process of letting you create a custom character.

The idea here is that you can get a lot more replayability out of the game if you don’t know precisely what is going to show up each time you play. A procedurally generated map and randomized items make it so that you will have a different experience each time you play.


Related to procedural generation, there’s a strong sense of exploration in roguelikes. Not only are you exploring the map, but you’re also usually exploring the game mechanics.

For example, in the early roguelikes you generally had to explore the different items and what they did. For example, knowing the approximate ratio of good potions to bad potions makes you think about when to try a potion and when to wait. In the game Pixel Dungeon, it’s useful to know that healing potions tend to be a bit more common at lower levels; therefore, the first potion you get 2 of is more likely than not the healing potion for this run. I’ll talk a bit more about this meta-knowledge later.

You also get to explore how game mechanics and in-game items work together. In NetHack, investigating how blessings work helps a lot in the game later. (Mild spoiler) Knowing that you can dip an item into holy water to uncurse an item or to make a blessed item is super useful in succeeding in the game.

Crazy combinations

Related to exploration, a a major part of the gameplay is to figure out useful combinations in the game. The earliest version of this is probably the “Ascension Kit” from NetHack. This was a set of gear that had various abilities required to fight the last boss, along with any “intrinsics” your character had gained. Having the right ascension kit made the endgame easier… or maybe even just possible.

The king of crazy combinations is probably the game The Binding of Isaac. Usually the involve some sort of degenerate combo, that allows one item to take advantage of some element of another item, or eliminate some penalty. My favorite combos involve Brimstone, which is a fun ability to charge up and kill a bunch of things at once. These crazy combinations help add to the exploration aspects of the game, while the procedural generation parts make it so that you don’t always get the exact same crazy combination every time.

Resource management

Most roguelikes have a scarce resource you need to manage. It might be health and healing items. In a lot of the original roguelikes, you had to eat food to keep fed otherwise you might starve to death.

In some cases, this resource management is a stand-in for time. The starvation mechanic in early roguelikes was basically a timer; you couldn’t stay on a level forever finding every secret door, eventually you had to move on or you would starve. In Crypt of the NecroDancer, playing as the main character Cadence gives a time limitation in the form of the song. You can only move on the beats, and the song only has so many beats; once the song is over, you are forced to the next level automatically.

Most games have items that are found and used in the game, and these must be managed carefully. Often inventory is limited, so even the number of items you carry has to be carefully managed. Sometimes you can even lose items and gear, so you have to be careful to have backups somewhere in case you lose your items.


Some might list “perma-death” as an element of roguelikes, but I’ll make this more general and call it “failure”. A good roguelike lets you fail. Usually this means that a character only has one chance, and if you die then that character has failed. A lot of games expect you to fail many times before you accomplish the final goal.

Not every game focuses so heavily on failure. Dungeons of Dredmore was one game I was able to beat on what was essentially one of my first few characters. Obviously I didn’t crank up the difficulty all the way, so I could play it again with more challenge and more risk of failiure. :) But, the death of a character is still there, lurking.

Failure in roguelikes has to be carefully managed by the designer. First, failure has to feel like something the player has some control over. Randomly dying is frustrating. But, if a player pushes the limit and doesn’t manage a resource properly, then it becomes their own fault. For example, spending too much time in one area in a game with a starvation mechanic makes the death your own fault. Getting into a dangerous fight when you have few hit points makes the death your fault. A player blindly rushing into an area or using an item at the wrong time makes the death their fault. Of course, players might still try to blame terrible design. :)

The other thing is to borrow a page from startup culture and “fail fast”. It can be frustrating to hang on the ragged edge for a while, only to be killed in what feels like an impossible situation. Likewise, it sucks to die and then feel like you have to work a long time to get back to feeling cool. For example, in The Binding of Isaac, a bad setup would probably kill your starting character quickly. However, if you did find a cool power, it’s likely that you’d find some crazy combination soon after and get back that feeling of power you had before you made your previous fatal mistake. Making the player work too long to get back to a place where they feel cool can be bad.


The game should have some way to reward you even after the character dies. In the older games, this was in the form of meta-knowledge. A unspoiled person coming into the game has no idea what holy water is for, but the experienced player who has experimented will know how to use holy water to their benefit. The very experienced player will know the best ways to use holy water, saving it for the optimal application. This goes back to the exploration part, where exploring the game mechanics can give you a big advantage once you get the resources in a future run.

More recent games have made this reward more explicit. The Binding of Isaac lets you unlock new abilities depending on how you played previously. Defeating end bosses with certain characters made some items become available, which made even more crazy combinations. Rogue Legacy has the manor, which lets you build up benefits by spending the gold you pick up in the castle; death will take your gold when you go back into the castle, so you might as well spend it!

This is what makes failure tolerable. Knowing how to play the game better, or having some sort of perk you can rely on in the future helps losing a character not feel quite so bad when you die and have to start over; after all, you did get something from that round even if you died messily.

Different games do different things

Early roguelikes had a lot of other features that were pretty standard, and which are perhaps useful to identify as unique to them and not necessarily found in later “roguelites”: turn-based combat, overhead view with grid-based movement, enough complexity to have multiple solutions, etc. In fact, there’s a formal “Berlin Interpretation” that defines what a roguelike is according to a very narrow, specific definition that attempts to restrict the definition to only certain games. Again, I’m not all that interested in splitting a hair or maintaining genre purity. If you prefer to make a distinction between “roguelikes” and “roguelites”, knock yourself out; I’m conserving my resources for bigger fights. ;)

One thing about the more recent roguelikes or “roguelikes” is that they put different emphasis on these areas differently. Whereas The Binding of Isaac is all about the crazy combos, Spelunky is a lot more tame in the items you can find and ways you can combine them. But, Spelunky‘s platforming-focused procedural generation of levels is more complex than the relatively simple requirements for the layout of The Binding of Isaac‘s mostly rectangular rooms.

I think you could also have interesting games where you start even dropping some of these elements. Would a game with a complex procedurally generated world that put more emphasis on exploration and less emphasis on resource management still feel the same? Or would that feel like a completely different type of game unlike any other roguelike? I don’t know, but it sounds interesting. :)

What do you think? What are your favorite roguelikes? What do you think the future holds?


  1. For me, it is “being Binding of Isaac”.

    Comment by Murf — 14 July, 2016 @ 4:15 PM

  2. Tales of Maj’Eyal is hands down the best successor to NetHack I have found. Massive replayability due to many distinctive classes, multiple races and randomness of both exploration areas and items. It also has extensive background with an unfolding story and deep history of stuff that has happened and rumours of why the world is how it is now.

    Comment by Bobturkey — 14 July, 2016 @ 8:23 PM

  3. I love Spelunky and Pixel Dungeon (since you mentioned both). One side note about the latter is that there are also explicit meta rewards in the form of class specializations you unlock when you have reached a certain point with the class.

    Doesn’t really matter, I like a good roguelike.

    Comment by unwesen — 14 July, 2016 @ 8:52 PM

  4. Add another vote for Tales of Maj’Eyal. The sense of scope of the world, the excellent work done with the UI and skills and leveling system that makes it feel intuitive to a modern hotbar MMO player, the steady pace of class unlocks and achievements and accumulated meta-knowledge that makes each playthrough varied, and all done by one guy.

    I also have a nostalgic fondness for Angband and its variants. The simplicity of its gameplay loop (enter dungeon, fight things, level up, return to town, rinse and repeat, go deeper) made it easy to grasp and progress the furthest with, at a time when I was still very new to roguelikes and unable to come to grips with Nethack’s style (what? Not explore every last corner and kill everything on a level? No way… => starved to death / died to something or other)

    I’m a mild purist in that I have a special place in my heart for true ASCII roguelikes. There’s a level of abstraction in colored letters and symbols and text that isn’t there when replaced by graphical tiles, though obviously tiles make it more accessible to a general audience. I like how complex everything else can get (eg. up to Dwarf Fortress levels of procedural and narrative generation.)

    Still, the roguelike spin-offs are intriguing in another way, to see how much of a innovative riff a dev can get mashing genres together. I liked playing Realm of the Mad God once upon a time, before whatever they did with microtransactions. Flash game meets bullet hell shmup meets roguelike permadeath and unlocked classes = fun speeded up action game where reaching max level is 20 min or less of one’s time.

    Unreal World is probably the most simulationist wilderness simulation game I’ve ever played (while still remaining “game” enough to grasp and play/learn.)

    FTL was interesting in that it really stretched the boundaries of what one might call a roguelike (to me), but it was a compact and “good” feature-complete game in its own right.

    Comment by Jeromai — 14 July, 2016 @ 9:51 PM

  5. I used to be pretty obsessive about NetHack back in the early-to-mid 90s. I even tinkered with the source code and would play with changes. I still have a spiral bound copy of the essential NetHack guide to the game that was compiled over the years at or whatever that Usenet group was called, back before the green card lawyers ruined Usenet.

    I think the combo of simplicity and depth grabbed me, but the one feature I really liked is that you could scratch something on the floor or die and in a later game come across these things as evidence of past attempts. That bit seems to have been lost over time. And, of course, it was at a time when few were dreaming of graphical games of such breadth and depth. It was very much a game in its time.

    Over the years I have played many variations on NetHack and Rogue. There were a few attempts back in the 90s that involved slapping a simple graphical interface on the whole thing, but that never had the same feel a me running around as an ASCII character. The influence is everywhere though. Even Pokemon Mystery Dungeon is essentially a variation on NetHack with Pokemon, short dungeons, and no perma death.

    Comment by Wilhelm Arcturus — 15 July, 2016 @ 6:23 AM

  6. . . . Castle of the Winds. I liked it, loved it, and my brother and I both stomped it into the ground pretty thoroughly without excessive save scumming. I say excessive, because we did sometimes reach levels which just were not beatable when we set foot in them.

    I did play Nethack, didn’t . . . really like it, the learning curve appears atrociously high and there’s so many useful functions which are never explained . . . and so many functions anyway that you really have like a textbook which is called a “Strategy Guide” to slog through. I respect the heck out of this game, and what it can do, but I can’t play a game like that – on top of the permadeath, meaning the only way to learn from mistakes is to start over and hope the first two floors don’t kill you.

    Comment by Kereminde — 7 August, 2016 @ 6:56 PM

Leave a comment

I value your comment and think the discussions are the best part of this blog. However, there's this scourge called comment spam, so I choose to moderate comments rather than giving filthy spammers any advantage.

If this is your first comment, it will be held for moderation and therefore will not show up immediately. I will approve your comment when I can, usually within a day. Comments should eventually be approved if not spam. If your comment doesn't show up and it wasn't spam, send me an email as the spam catchers might have caught it by accident.

Line and paragraph breaks automatic, HTML allowed: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>

Email Subscription

Get posts by email:

Recent Comments


Search the Blog


August 2020
« Aug    



Standard Disclaimer

I speak only for myself, not for any company.

My Book


Around the Internet

Game and Online Developers

Game News Sites

Game Ranters and Discussion

Help for Businesses

Other Fun Stuff

Quiet (aka Dead) Sites

Posts Copyright Brian Green, aka Psychochild. Comments belong to their authors.

Support me and my work on