Random Video Games

What this list is for

You have four hundred games and you play the same one. Everybody does. The library is not a library, it is a monument to intentions.

Press Generate and something else decides. A few of the titles in here:

• The Legend of Zelda: Breath of the Wild
• Red Dead Redemption 2
• Among Us
• Minecraft
• Cyberpunk 2077
• Total War: Three Kingdoms
• Solar Ash
• UnderMine

Now try to sort those into genres in your head. You will reach, almost immediately, for words like roguelike and soulslike and metroidvania. Stop there for a second, because something strange is happening in that sentence.

Drawing a game

1. Press Generate and play whatever it gives you for one hour before you judge it.
2. Ask for three if you want to choose rather than obey.
3. Type into Contains to search titles.
4. Copy to keep the shortlist.

Where these titles come from

Games gathered and checked by hand, spanning enormous blockbusters and things made by four people. The tool picks from that pool in your browser.

The only art form whose genres are named after products

Think about how other media name their categories. Film has the western, the thriller, the musical. Music has jazz, punk, ambient. Books have crime, romance, science fiction. Each of those words describes a quality of the work: a setting, a mood, a subject.

Now do games.

Roguelike means like Rogue, a dungeon game written in 1980 by university students in Santa Cruz, displayed in letters of the alphabet, in which death is permanent and the dungeon is different every time. Metroidvania is two products glued together, Metroid and Castlevania, and means a map you unlock by finding abilities. Soulslike means like Dark Souls. Doom clone was what we called first person shooters for years, until the industry decided that naming a genre after a competitor's product was commercially awkward.

No other medium does this. Nobody calls a detective novel a Holmeslike. There is no film genre called Casablancavania.

The reason is that games are defined by their verbs rather than their subjects. Two games can share a setting, a mood and a story and be nothing alike, because in one you are shooting and in the other you are negotiating. What actually needs naming is a bundle of mechanics, and mechanics are extremely hard to describe in ordinary English. So people point at the first game that did it and say: that, but different.

It is lazy, and it is precise, and it works.

The weekend some people tried to write a definition down

Naming a genre after a product creates an obvious problem. As soon as somebody makes a game that is a bit like Rogue but not much, the arguing starts, and it never stops. Fifteen years of it.

So in September 2008, a small number of developers and players of these games met in Berlin at something they called the International Roguelike Development Conference, which was less grand than it sounds, and they did something almost no fan community has ever done. They sat down and tried to write the definition of their genre.

The result is known as the Berlin Interpretation, and it is one of the strangest and most likeable documents in games.

They began not with principles but with five games they all agreed were the real thing, and they treated those five as the canon. Then they went looking for what those five had in common, and they sorted the results into things that matter a lot and things that matter a little.

What they actually produced

The high value factors are the ones you would guess. The world is generated fresh every time, so no two runs are the same. Death is permanent, and you start again from nothing. The game is turn based, so you may take as long as you like on any move. It runs on a grid. Movement and combat and everything else happen in the same mode, without the game changing gear. There is enough complexity that any problem has several solutions. You are managing limited resources. You are exploring, and working out what unidentified things do, from scratch, each time.

The low value factors are the ones a modern reader finds funny. The game should have a single player character. The monsters should obey the same rules you do. It should be displayed in ASCII characters. It should take place in a dungeon.

And then the preamble does the thing that makes the whole document remarkable. It says, in effect: this list tells you how roguelike a game is, not whether it is one. Missing points does not disqualify a game. Having points does not qualify it. The purpose is to help the community understand what it is studying, and it is not there to constrain anybody.

A group of people sat down to end an argument about definitions, and produced a definition that refuses to settle the argument. On purpose.

Genre as family resemblance

There is a real idea underneath that, and philosophers got there first.

Wittgenstein pointed out that you cannot define the word game. Try it. Some games have winners, some do not. Some involve skill, some are pure luck. Some are competitive, some are solitary. Every feature you propose as essential is missing from something everybody agrees is a game.

What holds the category together, he said, is not a shared essence but a network of overlapping similarities, the way members of a family resemble each other without any single feature running through all of them. Your grandmother's nose, your uncle's laugh, your cousin's build. No one thing, but you would know them in a crowd.

That is exactly what the people in Berlin built, whether or not they meant to. A list of resemblances, radiating from a handful of prototypes, with no requirement that any game have all of them.

Which means the endless internet argument about whether some game is really a roguelike is not a factual dispute at all. It is a disagreement about how far from the family a resemblance can travel and still count. Nobody is going to win it, and the argument is not a failure of the community. It is the category doing its job.

Questions people ask

What is the difference between a roguelike and a roguelite?

By the strict reading, a roguelite keeps the permanent death and the generated levels but drops the turns, the grid, or the rule that you keep nothing between runs. In practice the word roguelite was invented so that people could stop arguing and get on with playing, which is the most useful thing any genre label has ever done.

Is it worth playing an old game I have never heard of?

Often more than a new one. Recommendation systems bury anything that is not currently selling, so the games you have not heard of are not the bad ones. They are the ones that stopped being marketed.

How long before I give up on a game?

An hour, and be honest about whether the game is bad or simply unfamiliar. Most games teach you how to play them in the first hour, and the discomfort of not yet knowing is easy to mistake for the game being poor.

Do genre labels actually matter?

They matter to the people buying, because a label is a promise about what you will be doing with your hands. They matter far less to the people making, who mostly just want to build the thing and let somebody else file it.

References

1. The Berlin Interpretation, full text, as posted after the International Roguelike Development Conference 2008. https://www.roguebasin.com/index.php/Berlin_Interpretation
2. Cartlidge, D. Genre, Prototype Theory and the Berlin Interpretation of Roguelikes. Game Studies. https://gamestudies.org/2403/articles/cartlidge
3. Roguelike, on Rogue's 1980 origin, the 1993 coinage of the term, and the canon used in Berlin. https://en.wikipedia.org/wiki/Roguelike