Gamasutra: The Art & Business of Making Gamesspacer
'Roguelikes': Getting to the heart of the it-genre
View All     RSS
October 23, 2014
arrowPress Releases
October 23, 2014
PR Newswire
View All

If you enjoy reading this site, you might also want to check out these UBM Tech sites:

'Roguelikes': Getting to the heart of the it-genre

May 21, 2014 Article Start Previous Page 2 of 2

Anyone who has made a game has first-hand experience with gameplay systems that never took off, and games that never delivered the promise of their initial prototypes. Roguelikes are also insurance against facing that scenario, argues Cook.

"There's nothing worse than finding yourself six months into production only to discover that the mechanics that seemed convincing enough in test tube of preproduction are in fact shallow and boring," he says. "By prototyping with procedural levels, you are forced to make your core mechanics robust in the face of really bizarre scenarios. This robustness tends to yield playspaces that deal with all sorts of abuse in terms of future expansions or balancing issues while still remaining fun."

He zooms out: "On a higher conceptual level, roguelikes pose a strong aesthetic stance. They say: What if our game was treated like code? With simple modular objects; with systems of interaction; with a certain physics to the world."

This allows for better collaborations -- a particular must for small teams -- argues Short. "I think roguelike design is extremely programmer-friendly… which means indie-friendly, since most indie companies are at least 50 percent programmers. Roguelikes are puzzles to be solved, and programmers (and technical designers) dig puzzles."

This leads to games that are well defined by their own gaminess. "They aren't shy about being games with game verbs and abstract game-like rules," says Cook. "Sure, that fluffy writing and art stuff can improve a roguelike, but the core fun is the logic of the systems. Systems beyond spreadsheets," adds Short.

All of the above cascades into a game that is primed for constant evolution -- which is increasingly crucial as game developers strive to keep players interested in a game past an initial purchase, and to build communities. That's another secret to their appeal for developers. 

"They're perfect for constant evolution. Every additional item isn't a new episode – it's a new world of possibility," says Short. "Now that it's common knowledge that you can sell your initial game, and keep working on it, potentially forever… it's an indie dream come true."

"Here's a style of game that thrives on never being finished."

"With a roguelike, you can add in 2,000 new objects and the whole system adapts with the press of a button. It is still hard work, of course, but here's a style of game that thrives on never being finished," says Cook.

"There's a lifetime of potential expansions you could pour into a roguelike. How wonderful would it be to have a community playing a game 30 years from now that they still consider fresh and exciting?"

It goes further than that, however, says Short. "The value of proc-gen isn't just in replayability; players comparing stories and strategies of how they survived a roguelike is part word-of-mouth marketing and part player generated content! Normally, you don't get that kind of community-building virality except in multiplayer."

That has already driven success for Don't Starve. "Social media and the whole Let's Play scene are providing a platform for players to present their play as performance, which works really well with procedural and emergent gameplay. It's also free advertising, which is hard to resist," Forbes says. 

Flexibility, adaptability, and the future

Roguelike elements are fast becoming part of the fabric of gaming -- even creeping into triple-A titles. "I think the fact that a game like Demon's Souls has mass appeal means that players are warming to some of the ideas," says Forbes.

Even players with much more casual tastes are also prepped to appreciate the fruits of the genre, Short argues: "Games like Candy Crush demand mastery and improvisation. It won't be long before a roguelike takes the mainstream by storm."

"It won't be long before a roguelike takes the mainstream by storm."

The genre is highly adaptable, as Doucet points out, which also bodes well for its success. "I can start one of those up, have a unique and interesting experience in 15 minutes, get my butt kicked, and try again. This is very important when you're an adult and have less time. Within those parameters, it's the perfect genre for me."

Cook suggests that the "roguelike" is a framework to build a game on. He sees Edmund McMillen's The Binding of Isaac as a great example of how melding roguelike concepts to ideas taken from other games can still lead to a coherent and compelling whole.

"What I find exciting is how people are now breeding these wild chimeras out of a half dozen different genres and gluing it all together with roguelike architectures. The architectural element, how the various individual design patterns fit together into a robust whole, is something that is worth more attention," Cook says.

Its changeability opens up thematic possibilities, too. "I want to use the genre as a way to explore real-life conditions. In real life, you can't hit the reset button and use knowledge of the future to subvert the challenges in front of you," says Doucet. In his spare time, he's been working on Tourette's Quest, a game that explores his own personal challenges. Thanks to the genre's strengths, he says, "I can make a game that's fundamentally about risk management and learning to embrace the physical limitations of a disability."

Despite the potential of the genre, Cook also worries that it could become played out. It's a danger of any genre that comes into fashion. "There's an opportunity to make better games by smaller teams. The risk is that they just copy tired patterns and run the concept into the ground."

Looking ahead...

This article has consciously sidestepped the issue of orthodoxy that has been such a piece of the discussion around the roguelike genre. Forbes' take on his own game exemplifies this conversation well: "Don't Starve uses permadeath, procedural world generation, and discoverable rules-based systems. It's quite a bit less directed than a classical roguelike, and in fact I would consider it more roguelike-inspired than an actual example of the genre."

"Probably, the genre should be destroyed, and actually I'd argue it maybe already has been."

Both Doucet and Short have covered this topic in depth, if you're interested in learning more. Burgun, in fact, doesn't see the value of considering it a genre rather than a set of mechanics to experiment with: "Probably, the genre should be destroyed, and actually I'd argue it maybe already has been." The goal of this article is concentrating on what core roguelike mechanics make possible for game developers.

It's clear that, genre or not, the design elements that make these games compelling are both identifiable and useable in a wide variety of contexts; they are now part of the lexicon of game design.

Our interviewees paint a bright future for games which take inspiration from these mechanics, and it isn't purely because they're compelling to play: there are also production and promotional reasons that make them extremely attractive. That adds up to a recipe for an enduring legacy.

Article Start Previous Page 2 of 2

Related Jobs

Nexon America, Inc.
Nexon America, Inc. — El Segundo, California, United States

Localization Coordinator
Petroglyph Games
Petroglyph Games — Las Vegas, Nevada, United States

DeNA Studios Canada
DeNA Studios Canada — Vancouver, British Columbia, Canada

Analytical Game Designer
Xsolla — Sherman Oaks, California, United States

Senior Business Development Manager


sean lindskog
profile image
You guys nailed it. Thanks Christian and interviewees.

Don't Starve and FTL, both games with very roguelike qualities, are two of my recent favorites.

The thing that really stands out to me is this. A roguelike is make or break on the quality of your game systems. 'Cause that's pretty much all you got. You can't rescue a shitty roguelike with shiny production value or Hollywood-esque "movie game" scripted content. These games kneel down and worship at the altar of the Game Systems Gods. These game systems better be great because there's nothing else to hide behind.

Patrick Casey
profile image
Agreed. Both Christian and the interviewees nailed it.

And YOU nailed it with this comment and your one below. Well done!

Bart Stewart
profile image
The question I keep coming back to is this: Why does deliberately combining exploration-friendly PCG with exploration-punishing permadeath ever seem like a good idea to any developer?

I don't say they "shouldn't" ever be blended; trying different combinations is interesting. But the reason why Rogue/Hack and their ilk killed off the player character randomly wasn't because of any modern notions of challenge or excitement or social competition -- it was because programmers ran out of RAM! When you had so many interesting things you wanted players to be able to see, there wasn't room for or interest in gentle failure mechanics. You "died," you started a new game.

Why do game developers today think it's a good idea to copy that mechanic when the constraint that originally required it no longer exists?

For various reasons I also strongly support the development of "procedurally generated, system-driven games." That is a valuable design model for all the reasons given in this article. It's why I've been watching with great interest Josh Parnell's Limit Theory ( ame) since its successful Kickstarter.

It's the retention of that other structural choice now called permadeath that I'm not convinced will be broadly appealing... or necessary. It was tolerated 35+ years ago, and obviously there are some gamers who like being abused with a "game over" after a lot of time invested in progress. But I'm not persuaded that's going to be a popular component in more mainstream games.

I'm familiar with the "mastery" argument. I'm doubtful that it will resonate with most people, who don't like to lose.

The really ironic thing is that this article shows up just as I am finishing porting a 1985 version of the classic STARTREK game from PL/I to JavaScript. It's filled with completely random game-terminating moments. So this question of permadeath as a "feature" is one that's very fresh for me.

Why in the world should this instant-loss-of-progress thing be perpetuated when the other component -- procedurally-generated complex systems to explore -- is so much more interesting and welcoming?

sean lindskog
profile image
Permadeath is all about the visceral feeling of true risk.

If I don't win this battle, it's game over. In my brain, risk/reward are balanced emotions. Exploring to that new, deep, dangerous part of the game is tremendously more satisfying to me if it is *risky*. It is much less satisfying if it is just the eventual result of putting enough hours into the game (and reloading as many times as needed).

Along the same vein, the feeling of accomplishment is much greater if there is a real chance of failure along the way.

Now, certain game mechanics make permadeath work better or worse. For my tastes, I tend to dislike anything I perceive as an "unfair death" in a permadeath game. Alternatively, if immediately after dying, my brain goes, "here's what I should have done...", then I'm likely to dive back in.

Also, procedurally generated games tend to work well with pernadeath, since you get a fresh new game world to match your newly honed skills against. Rather than just repeating the exact same content.

Paul Schwanz
profile image
I think your last point gets the correlation right. Permadeath benefits from procedurally generated content in order to stay fresh each time around. But I'm with Bart in that procedurally generated games don't necessarily benefit from permadeath.

I think that permadeath is attractive primarily to achievers and not to explorers, while procedurally generated content is attractive primarily to explorers. But permadeath can grow old quickly even for achievers if they have to do the same thing over and over again with each restart (though Dark Souls is an interesting counterpoint). This is where roguelikes solve an issue for the achiever who is interested in permadeath.

Roguelikes may also attract some explorers because of the possibilites presented through procedural generation and complex system interactions, but these gamers will find permadeath an annoyance to endure. I think there are valid game designs out there that will embrace the PCG mechanics of roguelikes while eschewing permadeath that have the potential to hit the sweet spot for many explorers, though achievers will probably be lost in the process.

Ujn Hunter
profile image
If there is no risk of dying, it's boring. Who wants to just click through a never ending procedural dungeon?

David Paris
profile image
Having written some of the really early Rogue-likes, permadeath didn't have anything to do with RAM. It was more that it gave you the opportunity to put the player through another iteration of the full random and customization process from the beginning, giving another pass at hitting some of the large amounts of content that would not all show up in a single-play through, and reset their progress so they'd have to keep playing. Beating the entire game, although possible, should be a monumental and lucky event, and rare enough that every discrete doing so was a unique diamond, worthy of retelling and different from the experience of the next guy (or same guy) who would do the same.

However, that sort of progress reset is really tailored towards a fairly hardcore gamer, and not very friendly to less laser-focused players. Over time we have come to realize that although these super-hardcore types do exist, and that most people enjoy a challenge, the average player got more enjoyment and utility out of the game if it was a bit less punishing. Since we're usually looking to reach a broader audience, this translated into more forgiving reset mechanics, lack of permadeath, etc...

Lorenzo Gatti
profile image
You seem to confuse skill-intensive roleplaying games with easy and grinding-intensive adventures.
Expecting to win any individual game by accumulating enough "progress" and "exploration" is an extravagantly arrogant notion; the real progress and exploration, and arguably the real fun, is learning strategy and pulling off cool stunts over the course of many games.
The character deaths you see as "instant loss of progress" are, for normal players, the appropriate consequence of a mistake or an excessive risk, and to some extent an appropriate form of roleplaying (adventures in dangerous places should be actually dangerous).
When a character dies, progress in the wrong direction isn't really progress, and loss of a failed game or an unsuccessful experiment isn't really a loss.

Andrew Pellerano
profile image
There is the concept of N*M design where I create N things along one axis, M things along another, and then the resulting combinatorial matrix allows for N*M possibilities. Traditional games take more of an N+M approach which requires lots of content and large teams.

It's only by coincidence that roguelikes rely heavily on N*M. There's opportunity to find other uses for N*M in other genres as well

Karl E
profile image
Seems like a low ambition... shouldn't N^M possibilities be the goal to strive for? Every new orthogonal variable is a new axis in the space of possibilities.

sean lindskog
profile image
My games are cos(N ^ log(M)) * rand(42)

Andrew Pellerano
profile image
I think what you're describing is just N*M*L*K*J... For example a player stat system may let you choose from 3 classes and then stat points across 6 properties that range from 1 to 10. That's 3*10*10*10*10*10 character possibilities, not 3^6.

I'm having a hard time wrapping my head around a concrete example of N^M, do you have one in mind?

Karl E
profile image
The social aspects of PCG cannot be overemphasized. These games can be lonely and players are itching to share any events they believe they are the only person in the world to have experienced. Also, players should be able to experiment with the PCG mechanisms themselves and share the results. After all, the best PCG algorithms can be used by anyone who can enter seed values and wait to see what comes out.

Matt Mirrorfish
profile image
I second the comments on PGC and risk. The feeling that you have this seed/world once and will lose it if you make a mistake is very compelling to me.

Also for someone like me who is predominantly mechanics/systems oriented and not interested in long stories the fact that I can have a 15-45 minute experience that feels complete and exciting is very valuable.

I think perma death is also an interesting motivator for learning, which is the primary axis of progress through these worlds. I really appreciate an opportunity to practice and learn a set of adaptive skills against a difficult but fair, non-repetitive teacher. Playing the system feels much more like learning the way your friend plays go then memorizing a series of jumps or whatever. Great piece!

Fabian Fischer
profile image
It's funny how we call so many things "roguelikes" these days.

Thankfully we are slowly but surely rediscovering what strategy games actually are about: Decisions and consequences. And to support that we simply need(!) a way to lose (“permadeath”). Also, we need to make sure that the system doesn’t degrade into a puzzle that you just learn the solution to. The latter is the reason why we need(!) some form of randomization to make sure that the player will always actually make decisions instead of just inputting the solution.

By far most video games of the last decades were (extremely elaborate) puzzles with little to no dynamic gameplay, most of them trying to be like movies on top of that (thereby completely disrespecting the uniqueness of the artform of creating interactive systems). Today we’re finally getting away from that idea again.

The only reason we associate these “features” (actually they’re really just necessities) with Rogue is that roguelikes have been this bastion of actual meaningful decision-making in the gaming world for a long time. Most of today’s “roguelikes” are not at all like Rogue, except for these two fundamental necessities of strategy games. I know, people have thus invented terms like “roguelike-like” or “rogue-lite”, but in any case it really doesn’t have all that much to do with Rogue anymore.

Anyways, because we’re rediscovering these things, it’s an amazing time to be a gamer! ;)

Charlie Saunders
profile image
This article is really insightful - I wonder how long the big boom in roguelike games will last? Hopefully a while! I think the blend of instant death and random generation is a winning formula. It's a reasonably simple genre to develop, too, as small additions can add whole new levels of game play experiences.