A Review of Dwarf Fortress Reviews

In my series "The Warp and the Weft", I'm taking a look at how games generate stories that we want to tell to others. The series works as a prelude and background work to creating my own colony sim and dynamic story generator. In this article, I'm going to examine some of those "stories we tell to others" - namely, Steam reviews of Dwarf Fortress.

Pictured: Urist McDwarf, the fortress's beloved mayor, is throwing a tantrum because he doesn't have enough marble chairs. Five innocents were injured.

It's worth thinking about the purpose of a Steam review up front. Steam reviews are written by a player, mainly in order to let other players know whether or not they should buy the game. They vary in form, from straightforward ratings to descriptions of the experience of playing the game to ASCII pictures of cats. Reviews can be rated by other Steam users as helpful or not helpful. As internet users are wont to do, a vote for a review being 'helpful' could also mean that the user thought the review was funny or touching, or simply that they like ASCII cats.

A large proportion of the reviews written for Dwarf Fortress take the form of stories from Dwarf Fortress games. I read 20 of the top-rated review stories in order to see what makes those stories tick. Why did the player want to share this story? What underlying game mechanics and systems were needed to make the story possible? What common forms, themes, and patterns start to emerge?

HIDDEN DEPTHS

These story reviews all have an undercurrent running beneath them: look what this game can do. Marvel at the depth of the systems, look at the gameplay and circumstances that emerge. They advertise Dwarf Fortress as a game with a huge amount of systemic complexity. The reviews demonstrate, through concrete examples, that Dwarf Fortress heaps systems on top of systems, and that entertaining results can occur when they interact with each other.

On another level, these stories show that there's a point to the systems. The systems don't interact in a vacuum, but rather they have legible consequences to them. That legibility is important. If a tree falls in a forest and no one knows about it, does it matter if it makes a sound? A player can look at the output of all of these systems and develop their own narratives.

In some way, the breadth of the systems mentioned is a sign of the depths of game. If every review mentioned the (extensive) health / body part system of the game, then you'd expect that to be a major focus of the game. Of the 20 top-rated stories, 6 of them involved the body part system directly or indireclty. You see that it's part of the game, but not the whole thing.

REMEMBER THE UNIVERSE

There's a certain interesting property of a simulation-focused game, particularly one that swings so hard in that direction as Dwarf Fortress. If you're playing something like Splinter Cell or another stealth game, then all of the systems in that game are in service of that initial stealth idea. The game promises to make you feel like a bad-ass special ops agent slipping into an enemy stronghold. A system in such a game can be judged purely on whether or not it contributes to that experience.

For a game like Dwarf Fortress, though, the simulation is the point of it all. A system is additive to the game's idea if it allows the world to be simulated in a more detailed (and interesting) way. A newly added system doesn't have to have a short-term goal, such as enabling stealth gameplay. It can be tied to the larger goal of simulating a universe.

This comes out in the story reviews I read. There were certainly some systems that showed up more often than others, but no system emerged as being dominant or even necessary. A story was as likely to include crafting or livestock as it was to include combat and severed limbs. Taken together, the reviews paint the picture of a game with seemingly limitless depth and with an infinity of possible stories.

A masterwork image of an adventurer in a tower.

LESSONS LEARNED

Finally! Bullet points.

- A number of stories involved the player making a mistake, or unintentionally causing the downfall of a dwarf or a fortress. There's a slight flavor of "pride goes before the fall" here.
- Other stories had a mystery - some event or circumstance that seemed random or inexplicable, where a deeper investigation led to an epiphany or punchline.
- Lots of stories had a surprise for the storyteller. This could be a previously-unknown system revealing itself, or a known system proving itself to be deeper or richer than expected.
- Fewer stories had a traditional narrative structure (rising action / climax / falling action) than expected. I would wager that the stories we tell about games hew to this path, too. The game is not played in a vacuum, and the story about the game includes the player (even if they go unmentioned). Many stories involve how the game suprised or delighted us as players. There could be moments of tragedy, but the moments of comedy are much better remembered. I'd wager that there is a gap between what we expect of "regular" stories as opposed to these game stories.
- Stories generally were not based on in-game relationships much, if at all. A story is more likely to reference a dwarf's profession, martial skill, or age rather than a dwarf's relationship. Part of this might be because relationships in DF are less legible. They don't have much of an effect on gameplay, and they aren't likely to be noticed by the player.
- The "events" that occur in a story are not necessarily the "events" that occur in the game. The different beats of the review stories may or may not line up with something that would show up in the game's announcement log. A small detail, but worth remembering. The player chunks the game's events into larger or more abstract ones. This can work in the game dev's favor, as you can focus on designing the smallest elements and events. It can work against you, too, in that those larger chunks of events will then have less design (or intention) behind them.
- Notably, within all of the story reviews I read, I did not see anyone 'hallucinating' systems or ascribing events to non-existent systems. The creator of Rimworld, Tynan Sylvester, tried to design Rimworld (a similar colony sim / story generator) to take advantage of apohpenia, or the tendency of humans to find patterns where none exist. He used an in-game example of a surgeon amputating a patient's leg to save their life, and the surgeon and patient are married. The player might imagine extra drama based on the couple's marriage, but that drama is not encoded in the simulation. No reviewer imagined systems that weren't there, and they all had a solid grasp of the in-game systems (and where the boundaries of those systems may be).

This was a useful project, and not just because it was entertaining to read these stories. These stories are subtly different than a traditional narrative, both in structure and formation. Players are commonly trying to demonstrate the complexity and interplay of Dwarf Fortress's systems, and they enjoy the humor found in unexpected events. If a game wants the mantle of "story generator", then its systems need legibility, consequence, and breadth. Dwarf Fortress offers much, and much to learn from.

Subscribe on iTunes

POWERED BY SQUARESPACE