Action-Adventure Storytelling and the Setpoem

The secrets of Angeline Era's world design, revealed!

Hello everyone! I hope the beginning of summer is going well. The weather in Tokyo is warming up, and it's hard to not just want to go out and play, rather than sitting in front of a computer and working… anyways, here’s the 6th Dev Timeline post for Angeline Era! (Older ones here)

Here's a photo from the other day: I was walking around Tokyo, till finding Zenpukuji River (GPS), a uncharacteristically green (for central Tokyo) area. I was somewhat moved by the area because it reminded me of the way some bike paths, neighborhoods and forest preserves would intertwine back home in Illinois.

Anyways, Angeline Era development. So what's happened recently?

• Demo preparation! - finalizing levels, music, art, the level-entry bushwhacking game, leveling up, etc.

• Demo testing + feedback incorporation, with release on May 9th as part of LudoNarraCon. Check it out on Steam up until release next year!

• Story/Cutscene Work

• Level revisions, overworld design (making the demo overworld was great practice). This is in a more ‘finalizing’ phase than a ‘making new levels’ phase right now. Gotta draw the line somewhere.

• Designing bosses!! On our 9th right now…

I'm of the opinion with game dev that the more vacations and breaks, the better… Let’s actually talk Angeline Era!!

This year, I particularly enjoyed biking to the top of Matengai Cliff over on Nishinoshima Island during the start of their sakura season. Biking up a mountain near the ocean reminded me of (I am serious) Quest 64's world layouts and how often the ocean is within sight, as a vertical reference point. And there's something game-like in that the entire island is essentially a tourist island - the roads are well-kept for people to bike/hike on. But more on that for another time!

Dev Timeline Reflections

Okay, so now I will move from what we just did, to reflecting on what we did last year. In my last post, I left off in mid-2023. Here’s what we did over the rest of 2023:

• Did playtests, discovered Setpoems (the meat of today’s essay)

• Made 4th/5th bosses

• Prepared a demo for funding pitches (which we're still trying to get… ha ha.)

• Made more progress on the musical direction (in particular, writing Jade Wonderland)

• Sephonie console port, merch, Anodyne 1's PC remaster, started work on the Anodyne 1 Fan Book

• Prototyped the level-entry Bushwhacking game!

• Added Artifacts (subweapons), gun upgrades and weapons

• Created fun organizational spreadsheets for tracking progress

• Story, Art Style, tools for the art (lighting, vertex color,etc)

And vacations! Haha! Including one that served as a lot of inspiration for the (currently shelved Sanpo project - although in some ways that's also become the Danchi Game I'm working on with my partner).

General Thoughts about AngEra Dev in 2023

However, I think the most work on my end was spent on doing level design. For example, some of the levels you can play in today’s demo were initially drafted in this period. As we shifted into full production, level design and story took the lead over audiovisuals. Maybe because for us, the music and art often act to convey additional layers of info built upon the story/design.

Perhaps the art and gameplay could have used some more time before our reveal and funding pitches, but in the end I think it was better to have made the deadline of 2023 Summer Game Showcases. It's arguable that without taking the step of publicly promoting it, we wouldn't have generated the reflections needed to then push the visuals and levels to the next level (as they currently stand!) The strongest tool for a game designer, really, is feedback - and the ability to process and discard it as necessary.

Various things come to mind as worth discussing, but let’s focus on the most interseting one… how Angeline Era developed the concept of the “Setpoem” and utilized it in service of “Action-Adventure Storytelling.”

The Core of Angeline Era: Theoretical Foundations for the Setpoem

A year into the project in early 2023, the game found its footing in controls and the poise-based bumpslash system, but not yet in level design. Generally speaking, I find level design is harder to develop a good direction for, vs. bosses, enemies or mechanics. It’s like the difficulty of writing a good sentence vs. a good chapter.

A theoretical foundation for setpoems came from a trip I took in March 2023 - in particular, the Site of Reversible Destiny (a sculpture park with uneven footing) which I go into in a post from that time. In short, that trip got me thinking about the tactile nature of 3D environments, and how simple variations in the steepness of a slope can strongly color how we feel moving through a space.

A 3D level designer’s dream! And an ankle hazard. The Site of Reversible Destiny..

The practical foundation for setpoems came from observing playtests in April 2023. The combat fundamentals were great, but the level design didn’t flow right.

Eventually I thought about how, like moving through real life, we feel different sensations. And likewise, a game's action can create unique physical sensations. Some call this 'friction' in the physical sense. An action game might have 'primary sensations' like the Bumpslash, but, when you Bumpslash in other contexts like with enemies, or things to dodge - this creates 'secondary emergent sensations', which are more complex to describe. In real life, this might be catching a football at a weird angle when there's 10 seconds left on the clock! Or perhaps it's multitasking in the kitchen. Or in our game, it's performing a bumpslash only to have to deftly hop out of the way of an oncoming charging enemy. The primary thing you do is necessary to have the sensation, but it’s very much all the context for that thing, that create all the definition and color of the sensation.

By Spring 2023 I had felt that this was the potential in Angeline Era's combat and levels! …but the levels weren't really living up to the task in conveying the feelings. They felt like a mushy mix of action or exploration… some combat rooms wouldn't have eye gates (which gate progress through the level). The pieces were there, they just weren’t being articulated well.

Discovering the Setpoem

A few things helped us figure out the next direction for levels: In particular, I had an overreaction to playtesting:

I think most gates should just stay open after clearing, even if you die.

This is clearly a horrible idea, but that's why you work in teams! Marina replied with useful comments:

It also depends a ton on whether it's a linear level or not. A more dungeon-like level can have fewer respawning gates naturally. However a linear level without many respawning gates feels kind of pointless, like why even make the player run through those rooms again if you're not challenging them to do better than before so they have more HP for the later hard part that they died on? Imagine like… Bayonetta's series of action encounters, if you could walk past the encounters… it wouldn't make any sense.

Part of the issue at the time was that you still had to kill every single enemy in a room to open the gate and progress. That's wild! It feels so slow. In the current game, on any given level you usually just end up killing 40-50% of the enemies to make it through, since gates unlock based on killing certain enemies.

Sorry I’m too lazy to remove the borders. The bottom left is an “Eye Gate”, you must kill certain enemies to open them.

Marina brings up another idea we had where certain enemies open the gates (funnily enough which was used in Anodyne 1) - and that combined with me thinking about all the strange little action-adventures like the Hydlide trilogy, as well as the humor of something like Dark Souls, or the ideas behind the The Site of Reversible Destiny. These things didn't just feel interesting and memorable - something to their level design was also memorable. Which led to some thoughts:

I've been thinking about this Robert Yang tweet on how level designers all love Elden Ring but learn nothing from it -

https://twitter.com/radiatoryang/status/1647020756352761858

I've been feeling this sense something is missing with (Angeline Era's) level design, even in recent levels, but especially in Oak Forest. The move towards 'top-down' designing has helped a bit but I think the missing step was "Action Storytelling" (made this term up)…

Oak Forest is the test level we playtested, and by Top-down designing I mean using a theme to drive the level (deep woods, lots of spiders, etc). I continued:

…Which is when the layout of enemies in a room isn't just fun from a pure action standpoint but is arranged in some slightly memorable/absurd way. So the level feels sort of like a 'story' in that the stuff you have to do with Action might be very hard but it's memorable simply bc there was some 'storytelling thought' put into the room Robert's tweet gets at this, but I think Action Storytelling only works when the game has a very good/pure action core that doesn't outbalance the enemies… and enemies feel very distinct even on a simple level. So Souls, AngEra, etc

Today, I'd revise my thoughts a bit. I'd call it Action-Adventure Storytelling (AAS?), and it works best when the game has a clear - and constrained - design language: E.g. "These things are harmful, these things block your way, your moves have certain affordances, and you must observe or recognize things about the environment or yourself to survive - rather than always mashing A and forward." I would say clear/constrained over ‘good'/pure’ nowadays.

When an action-adventure game has clear design language, you can start to create scenarios that can immediately be "read" by a player - and they will instantly perceive designer intent! For example, here are some extreme examples I came up with:

There would be easy things like, put a goliath next to 4 regular ones, and if the goliath dies the other ones stop moving (they get sad). Or there's a single bugdog and killing it makes 4 Goliaths run out at you. or just like more thoughtful layouts rather than the random jumbles

As Robert Yang puts it in the tweets:

if you use LD as a language, then you can tell jokes, hug the player, punch them in the dick, etc… and these aren't "reward nodes", these are expressed in the actual experience, the reward is the level

Here’s a silly example from the demo. What’s going on here? Hm…

I then expressed that AAS is kind of a vague term, and maybe what I'm describing is smaller in size than a big Action Setpiece. Marina responded with:

Hm a punchy term for it… an Action Setpoem? Actually i think what would be really neat and sorta bring this all together is if the conditions for clearing the room are generally not stated Each time you face a new room there’s a sort of air of mystery. If you play the same room multiple times you can eventually sort of narrow down/triangulate what the conditions are Its never like "hard" because 90% of the time killing all enemies will clear the room, but you can learn by rote or make poetic guesses and thus speed up how fast you can go through levels So like in your example, you only have to kill the goliath bugdog. ther’es no UI that tells you that and its not always that way on screens with goliath bugdogs, but on that screen its how it goes. you could miss that entirely but its there. And then you develop this relationship with like every screen of expectation, memory, etc

We shortened Action Setpoem to "Setpoem". I then went on to make a revision of the playtest level, thinking about Setpoems, and it seemed to be good!

Setpoems vs Storytelling

I will add that, a year and a bit on, I do actually think Action-Adventure Storytelling (AAS) is a distinct idea from Setpoems. To me, Setpoems are when a level designer uses the game's design language to construct a moment that is memorable for any sort of reason. It's a granular, small thing. In Angeline Era, this is like stumbling on the Turnip Cave, the Honeybun shop, watching the bugdogs hop over stumps, the immense amount of pots in that One Level, a room filled with dangerous chargers only to have them immediately killed by a spiky log, and so on…

Nothing to see here, surely…

It might require some one-off code, but not much, as most of it builds on the game's core language. Compare Uncharted 4's expensive car chase, with merely creating a room of Bugdogs in Angeline Era that rotate to stare at you. The Uncharted example is bigger - a "Setpiece" - while the Angeline Example is smaller - just a "Setpoem".

Where Action Storytelling comes in is when I then employ multiple setpoems to construct a level's overall 'story', in conjunction with art/visuals/writing. For example, the demo's intro level, River's Rest, contains a single piece of framing text describing the game:

A quiet river branches ever smaller, tiny ponds at its tips. And the Fae, weary from a day's travel, splash about a momentary peace.

The rooms of River's Rest are usually simple. Visually, there are many basic bugdogs that are standing around in pools. It's a wet forest of sorts. Some are hiding behind trees, there's even hidden 'bathhouse' rooms.

Overall it's a very leisurely level with mild threats. On its own, you might get the sense that some of these Fae could care less about fighting. Many mind their own business. Together, these setpoems combine with the framing text and visuals to roughly tell a 'story' about the whole area. Then, that loose narrative unit fits in with the game's universe and explicit story!

This, I think, is one of the biggest tricks behind effective action-adventure level design, and I think it's how memorable action-adventure games achieve that status (Dark Souls, etc, included.)

Writing a good setting, item lore, etc, is important - but those can't function without an engine - and that engine is the game's combat, enemy, enemy encounter, and level design.

This "engine" can be used as a foundation to create all sorts of setpoems, which then, allow the designer to create levels and spaces that embody a wide variety of emotions.

Of course, in Angeline Era's case the levels still have to function as action levels! Eventually we realized that the game ought to delineate different types of levels - Scale Levels and Non-Scale Levels. While both would operate with Setpoem logic, the flow of a Scale level is more linear and action-focused, while the flow of a Non-Scale level is looser and more exploratory. We decided to call Non-Scale Levels "Encounter Levels".

Encounter Levels

They are called "Encounter Levels" because they're about seeing what things you encounter! We also decided there are often little to no 'progression rewards' like missable new equipment. Although often you can search around and find some money or small secrets. Also, you almost never need to 'have a certain item' to see everything in an Encounter Level. The result is that players are more confident to fully engage with a space upon reaching it, rather than always worrying "am I completing this 100% or missing an important item?"

Encounter Levels are necessary in Angeline Era as there are many Setpoem ideas that break would the flow of a action-y level. For example, what if I want to have a mysterious giant tree hollow? A wide-open field? Enemies doing funny things? Over 2023 I got better at avoiding mixing the two types of level design as much, and part of my revisions this year are untangling these old less focused drafts. Turns out, you can combine multiple setpoem ideas to not just make an action level - you can make a compelling, almost actionless experience.

If you know… you know.

You could view it as the way that novelists write chapters or sections in different styles. In deep, descriptive passages, readers know to take their time and take in the sensory aspects. At other climatic moments, the writer may shift into writing in a more fast-paced way, where it's less about taking in details and more about enjoying the emotional ride.

Having the paradigms of "Scale Level" or "Encounter Level" in the back of my head allow me to better "write" the levels in Angeline Era. It also lets me better decide how to utilize individual Setpoem ideas. For example, the idea of a "bloodthirsty sniper stalking you through the woods…" that became best expressed through the demo's challenge level, "Open Season", a level with many combat rooms themed around that idea, with normal enemy (the sniper) modified a bit to be vicious. But the idea of "strange genetically modified turnips" was a better fit for, e.g., Ivyleigh Farms' strange vegetable rooms.

From working on Angeline Era, I've really been able to learn and think about action-adventure, or Exploration and Action, and how they go hand-in-hand to build out a world. But not just that: I think I've been able to more clearly think about the way that the fundamentals of a game (movement, interactions) can synergize with the narrative setting and audiovisuals, to create form interesting levels and spaces. This has been hugely helpful when playing older games as well and thinking about what makes them work!

In fact, this thinking has already been helpful in brainstorming and working on other projects, even outside of an action context. What will the future hold?

Well, if you've read this far… why not play the demo for yourself, or subscribe to this Substack… or our Youtube?