So I went on an off-the-cuff, typo-riddled tweet-a-thon about the Berlin Interpretation, the generally-accepted list of “factors” that constitute a trueborn roguelike. The collaborators behind the Berlin Interpretation said that they never meant their list to be proscriptive—that they were writing a field guide to spotting roguelikes in the wild, not a Save The Cat for making new ones—but even so, the Berlin Factors are oft cited when someone or other wants to claim that Spelunky flatly doesn’t belong in the same conversation as Shiren the Wanderer.

As I said on Twitter, the central question of the Berlin Interpretation is, “What makes a game feel like Rogue?” But the frankly more interesting question is, “What makes Rogue feel different from other games?” In that light, we need to reconsider or at least re-prioritize the Berlin Factors, wherein turn-based play is a “high value” roguelike quality, despite lots of turn-based games having nothing important in common with roguelikes—and wherein monsters that “are similar to players” constitute a “low value” quality, despite that being relatively unusual in games, and thus a key factor in, for example, Spelunky’s roguelike lineage.

Derek Yu says as much in his Spelunky write-up for Boss Fight Books, recalling, “When I was working on Spelunky, I focused on just three attributes of roguelikes that to me held the essence of the genre,” those being “randomized level generation,” “permanent death (also known as “permadeath”), whereby the player has one life and cannot reload their game to take back mistakes,” and finally, “a ruleset for physical interactions that is shared by the player, non-player characters (NPCs), and items.” He then goes on to explain:

The third attribute is my version of “monsters share rules and behaviors with players” and it extends not only to monsters but to items as well. In many games, monsters and items are coded separately and handled separately. You might be able to pick up an item and kick a monster, but you couldn’t kick an item and pick up a monster. In a roguelike, you should be able to pick up and kick items and monsters, with results that are based on consistent rules.

As I’ve said before, in practice what we call roguelikes (or roguelites, or roguelike-likes, if we’re trying to keep the purists happy) are at best selective about their use of the Berlin Factors. They’re more about adapting Rogue’s key formal and tonal selling points, the things that attracted players and designers to it in the fist place—consistent rules, persistent mystery, abundant danger—than about filling out a lengthy and exacting genre questionnaire.

More than that, in holding onto the Berlin list while discarding its preconceptions, we can better understand why some games that are not Berlin-blessed in any degree nonetheless possess a roguelike feel and a roguelike appeal. Here I’m thinking of the Souls games, their ever-expanding indie progeny, and discovery-rich, exposition-free adventures like Hyper Light Drifter. These games have no procedural levels, no permadeath, and yet they’re recognizably roguelike-adjecent.

If we think of the Berlin Interpretation as a litmus test, then this is a brain-straining paradox, a reckless disregard for what words even mean, a challenge to fight the this game doesn’t count fight yet again. But the Berlin Interpretation, I’d argue, is most useful when we use it in sort of the opposite way: Here’s how to recognize some Rogue-inflected design sensibility wherever it appears.

So I said all this on Twitter (in fewer words, obviously), and the first reply I got was from Droqen, the developer behind Starseed Pilgrim, Probability 0, and a plethora of other weird and lovely projects, who said “now i’m thinking how fun it might be to make a No Berlin roguelike,” in other words “a roguelike that excludes every Berlin Factor: not a jab at the berlin interpretation, but a fun exercise in game design & an exploration of strange roguelike space.”

Droqen wrote a post on itch.io, listing out each Berlin Factor and briefly considering the implications of doing the opposite. Some of them are fairly straightforward. “NO Permadeath: When you die, you keep… something. I’m going to interpret this as you either keep your character or your progress,” and “NO Random environment generation: Thank goodness this specifies random environment generation. Other random elements are fine, but the map (and generally the placement of things in it) is predetermined & designed.” Multiple player-characters instead of a single one. No grid-based map or maps. No dungeons, at least not in the sense of “stonework, rooms, corridors.” Easy enough.

And no resource management, which might mean no HP, and ah yes, no numbers, which almost certainly eliminates HP. But that might be OK, since we’re also avoiding the hack-and-slash, combat-centric Factor.

Other Factors are ambiguous. “NO Complexity: Eesh. Let’s go with LOW complexity.” Complexity is of course a relative term, so the key question here would be, do we mean low complexity as compared to other games, or simply low complexity as compared to other roguelikes? Spelunky has much higher complexity than Super Mario Bros. but much lower complexity than Dwarf Fortress, so there’s some wiggle room here. The game wouldn’t have to be simple in order to be vastly simpler than a roguelike enthusiast is probably expecting.

But here’s where it gets complicated: “NO Exploration and discovery: What??? How…”

How indeed.

“No discovery,” taken literally, is probably the one and only deal-breaker in the list. Learning how systems work, and seeing the particulars of a run unfold as those systems ping off of one another in surprising ways—that might be the hard kernel of what makes roguelikes interesting. We can mitigate or remove procedural generation and permadeath, but lose discovery and you lose most of what makes roguelikes—maybe even what makes games—compelling.

That being said, in the same way that we could eliminate randomized environments without altogether eliminating randomization, we could probably eliminate exploration without eliminating discovery. This dovetails well with the “no random environment generation” rule, as well as the rule about there being multiple player characters and with the idea of avoiding “dungeons” in the traditional sense.

Imagine a large but unchanging map. You don’t need to explore the map in order to reveal it. The whole thing is revealed to you from the start, and you can scroll around and look at the layout, including areas that are currently distant from your player-characters, whenever you like. However: Your characters’ positions are randomized, as are some qualities of the items, NPCs, and “monsters.” (More on why monsters aren’t necessarily monsters in a moment).

Droqen specifies that the positions of these elements would be more or less fixed, so let’s say they could change stats or other qualities without changing position, thus emphasizing the discovery without exploration theme.

Initially, you’d have to get to know the map. No avoiding that. But once you’ve acclimated yourself to the layout, you’re never exactly exploring. You’re just discovering new things within a familiar space. (Think about Space Alert, where all manner of hectic and hilarious things can happen to you while you float through space, but the floorplan of your spaceship is consistent). The map would follow no obvious grid of uniformly-sized tiles—Sunless Sea’s sprawling waters and undulating islets come to mind—and would avoid discreet “rooms,” since we could hand-design a coherent space rather than relying on computer-arranged chunks of logical and navigable dungeon stuff.

Then there’s this: “Monsters are NOT similar to players: Rules that apply to the player don’t necessarily apply to monsters. This is actually one of the more interesting rules to break severely. How different from the player can I make monsters?” For that matter, what constitutes a monster in this hack-and-slashless game of ours? What differentiates a monster from an NPC or an item? Are the “monsters” potentially unexpected events rather than wandering punchable baddies? That would make them fundamentally different from the player-characters, after all.

All of that to say, I’m fascinated by this hypothetical No Berlin roguelike, both as a formal exercise in sideways minimalism—how many expected features can we remove while retaining the essence?—and also as a refreshing thing to play. It could shake up our understanding of roguelikes, and by extension, of what we can expect to discover when we play games.

Because it’s all about discovery, probably. Maybe.