Roguelike Elements in Games: What the Indie Scene Gets Wrong

While the indie scene is often thought of as more creative than the AAA industry, it also has its fads and trends that dominate the market. The most well-known of these was the popularity of pixel platformers, although even before the indie boom, there were the occasional independently-made games that became popular. Nowadays, many indie developers are interested in borrowing mechanics from roguelikes, especially the ones that defined the genre: permanent death and procedural generation.

As with everything popular, there are critics and detractors. Many people who aren’t too fond of the “roguelike-lite” games are those who played a lot of traditional roguelikes. While some criticism might come from a place of elitism, there’s also a lot of valid arguments that might explain why most modern games with roguelike elements probably won’t have the staying power of NetHack or Angband.

This article examines problems with several roguelike-lite games. As those games can have rather dedicated fanbases, I feel the need to say that its intention is not to convince anyone that his/her favorite game sucks, but rather to compare their design choices with those of traditional roguelikes and see what can they learn from them.

The Purpose of Roguelike Games

The idea behind the original Rogue was to create a replayable adventure game–its creators wanted to counteract the fact that once the games they’ve played (mostly early interactive fiction titles like Colossal Cave Adventure) were finished, there was no reason to play them again, as the actions required to win stayed the same.

Rogue achieved its goal by placing the player in procedurally-generated dungeons. This meant that while playing it for the second time, one would not have the advantage of knowing the maps, or the placement of traps, monsters, or items. Aside from creating different levels for each game session, Rogue also randomized the appearance of items. A red potion could heal you in one game, and paralyze you in another. The only way of finding out was either trial and error, or using a Scroll of Identify–which was a guessing game in itself, as the scrolls were randomized as well.

The lack of a typical save system meant that the player couldn’t just save scum to find out which items were which. Additionally, the character’s constant need to eat meant that you couldn’t just grind levels until you could kill everything with your bare hands, as the only way to get more food was to find it on the lower levels of the dungeon. Thus, Rogue eliminated common ways of exploiting the game, and forced those who wanted to win to become proficient.

While later roguelikes emulated Rogue in everything from turn-based combat to ASCII graphics, its procedural generation and permanent death became the most important things about the genre. Roguelikes were all about challenge and consequence, as learning the game’s intricacies was more important than memorization or grinding.

The games became complex–but the complexity actually made them more fair, as it reduced the effects of luck, allowing the players to prepare themselves even for the worst situations. While Rogue would sometimes become unwinnable when the game didn’t generate enough food rations for the player to survive, NetHack gave skilled players many ways of avoiding death. Unavoidable deaths still happened occasionally, but they usually required extraordinarily bad luck.

Diet Rogue

While the huge influx of roguelike-lite games is a fairly recent phenomenon, titles inspired by the genre have been made for a long time. The late 1980s gave us a series of semi-random text-based, realtime-with-pause puzzle games known as Kroz, as well as Castle of the Winds, a two-part tile-based dungeon crawler, which is kind of like Moria or Angband, but allows unlimited saving and reloading. In the 1990s, randomization and item identification puzzles were used to great effect in Toejam & Earl, while a few years later Moria and Angband influences shaped Diablo, a gamer that revolutionized the action-RPG genre.

The popularity of roguelike elements in modern indie games can be traced back to Derek Yu’s Spelunky, which inserted procedural generation and permanent death into an arcade-style platformer. In later years, the genre was popularized by The Binding of Isaac and FTL: Faster Than Light, the former of which combined the aforementioned elements with Zelda-style action-adventure, the latter with space exploration. As those titles became popular, ‘roguelike elements’ turned into an indie fad.

The Appeal of Death and Randomness

The popularity of roguelikes and roguelike-lite games is intrinsically linked to the fact that, after many years of dumbing down, the difficult games have returned. It’s a great thing too, as most players want to be challenged, and developers lately aren’t afraid to oblige them. There’s been a lot of talk about ‘games as experiences’ versus ‘games as challenges’ –some people reasoning that, as games have gradually developed victory conditions, cutscenes, and storylines, those things became more important than achieving a high score.

As both stories and their presentation became more elaborate, completing the game and seeing how it ends became not a reward for the dedicated gamers, but an essential part of the experience, like watching a movie to the end. The resurgence of challenging games seems to prove that idea at least partially wrong. Many players are interested in mastering the gameplay just as much as experiencing the audiovisual and narrative aspect of the gaming experience.

Undeniably, there is something intriguing about a game where failure might result in having to start over. While the influx of retro-style platformers has made the high level of difficulty acceptable again, the difficulty of most of these titles was very specific. In games like Super Meat Boy, a very high level of precision is required, but only for a few seconds. The player must perform perfectly, or almost perfectly, for the duration of a level or until he/she reaches a checkpoint. However, levels are short and checkpoints are plentiful.

Thus, all the consequences are short-term. A bad move will stop you from progressing, but will not erase the progress you’ve already made, nor will it have consequences that aren’t obvious or immediate. The roguelike approach to difficulty is the opposite. Playing it safe, conserving resources, and planning become crucial, as lost progress can be counted not in seconds, but sometimes in hours.

Procedural generation is an appealing concept as well. Clever use of algorithns may lead to a lot of content being put in a seemingly small game. The idea that a game is completely different with each playthrough sounds great as well, as players can get sick of strict linearity and hand-holding.

Unfortunately, it’s not that simple.

Designing Levels On Infinite Typewriters

The thing about randomness and permanent death is that they’re not universally good things. Like many things in video games (cutscenes, RPG elements, non-linearity, open world), they can be used in better or worse ways, and there are some game concepts, styles, and genres in which they work better than in others.

Roguelike-lite games can often be too random, not random enough, or random in all the wrong places. While Sunless Sea shuffles some places around the map, it’s not random enough and it always plays like the same game. The exact coordinates of places you need to visit are randomized (and it affects the difficulty and profitability of certain quests), but you still have to visit those same places and do those same things. This is especially annoying in games with permadeath, as the player will quickly grow bored of the repetitive early game actions.

On the other hand, games like FTL, Zafehouse Diaries, and Darkest Dungeon can often be too random–similar actions can give wildly different results that depend more on the random number generator than on the player’s choice, making them too dependent on luck. Note that the randomness or lack thereof is in completely different areas of the game. Generally speaking, aiming for more randomness in level creation with less randomness in the outcome of player actions is a good rule of thumb.

Another thing that needs to be considered is whether roguelike elements work well with the game being developed. Traditional roguelikes belonged to a specific genre: role-playing games focused on combat and dungeon crawling. They were also fairly short, and their difficulty was based on player’s limited knowledge and resources. Obviously, an extremely long and narrative-driven game wouldn’t benefit from permanent death (unless it’s treated as an optional challenge like in Pillars of Eternity). You wouldn’t want to start from square one if you got killed by an end boss in one of the Final Fantasy games.

Procedural generation, on the other hand, would be an unwelcome change to games in which difficulty comes from enemy placement and environmental danger. If something like Dark Souls put random enemies in random places, most of the levels would become either laughably easy or nearly impossible. Outside of gameplay, it’s also important to note that in almost all cases, storyline, music, and visuals work much better when hand crafted, even in largely procedural titles.

Roguelike-lite developers sometimes disregard an important part of roguelike design: mechanics that prevent grinding. As the Skinner box approach to RPG elements is, unfortunately, pretty common, it sometimes creeps into the very genre that sought to avoid it. Games like Sunless Sea force the player to grind for money, grind for stats, and even grind to make subsequent playthroughs easier. They might do it to give the player sense of progress, or to make them spend more time on the game session, but whatever the reason is, it actually makes the games less enjoyable. Not only is grinding itself pretty boring, it also defeats the idea behind the roguelike difficulty. Ultimately, roguelikes are about the risk of losing progress and about having to master the game. If that risk is minimized, and mastery can be substituted with the ability to do the same thing for hours without getting bored, this kind of difficulty becomes pointless.

Who Needs A Life When You Have ASCII?

While putting roguelike elements in indie games has become a fad, it would be unfair to assume that those flawed games are flawed because the developers don’t know or don’t care about the consequences of their design choices. Even though I spent this whole article criticizing some aspects of the genre, there’s one thing I need to admit: games like NetHack, Angband, ADOM, Dungeon Crawl and Dwarf Fortress are so elaborate, interesting, and well-designed due to how many work hours were spent building, refining, and fine-tuning them. NetHack is still in development, and was first released in 1987 while borrowing a lot of code from an earlier game, Hack, which started out as a clone of Rogue. The gameplay ideas were built upon since 1980 while disregarding graphics, story, and even an intuitive interface. Creating a great roguelike is a monumental task, and the end result will be enjoyed only by small groups of dedicated fans.

Modern roguelike-lite games often don’t offer much to fans of traditional roguelikes, and they’re sometimes not even that fun to play. Ultimately, though, they’ve done their job by introducing new players to a largely obscure genre, through their attractive presentation, and gameplay that doesn’t require encyclopedic knowledge of a fictional world. It might be the casualization of one of the most hardcore genres out there, but the end result is still more hardcore than the norm. While most of these games will probably not have dedicated fanbases in ten, twenty, or thirty years, for the gamers of today they might be an entry point to things they wouldn’t even think of playing. I probably won’t be picking up too many upcoming games that describe themselves as having ‘roguelike elements’ but hey, at least more people will have an idea of what constitutes as a roguelike.

Further reading: