Team 7: Francis Santiago, Nicholas Bui, Thanh Tieu, Seamus Allen, Oumnia Chellah
Artist’s statement
Our intention with After The Fire was to create a cozy game for casual gamers who enjoy puzzles.
As humans in a world recovering from a pandemic, we saw the need for a cozy game to offer players comfort and retreat, especially during difficult times. In our experience, many popular games demand a lot of energy from players. Therefore, we envisioned a cozy and comfortable game that can provide enough challenge to make players feel fulfilled.
After The Fire explores soothing themes such as nature, connection, and growth to achieve this goal. In After The Fire, our main character uses their powers to rebuild their forest. As players unlock more seeds and heal more territory, their own magic and knowledge of their forest’s history grows. We wanted our players to feel the connection between their own growth and their environment. While our forest and little forest spirit might be fictional, the underlying message – that nurturing our environment can foster our own healing and growth – holds true.
Therefore, the game balances exploration and puzzle-solving to enrich the player’s connection to their environment. While the exploration helps the player uncover what caused devastation to their environment, the puzzles add challenge and fun to the player’s experience. The puzzles illustrate the impact of the player’s efforts in restoring the forest. By combining these elements, we hope that our game can appeal to a broad audience: from players drawn to compelling stories to puzzle enthusiasts and everyone who appreciates a game that marries an engaging narrative with thoughtful challenges.
Spatial Architecture, Narrative Architecture, and Puzzle Design Interactions
After the Fire’s narrative architecture is primarily communicated through interactions in the spatial architecture. Interactive items and puzzles are scattered throughout the game’s physical space, allowing the player to discover more of the embedded narrative through their interactions with these objects. The core parts of the story that guide the goals and motivations of the character are presented after unskippable puzzles, or once a player enters an area to ensure that the player does not miss the essential parts of the narrative, even if they choose to skip the optional interactions. The spatial architecture also communicates pieces of the forest destroyed by the fire, allowing for environmental storytelling while communicating atmosphere and tone. Finally, the regrowth of the forest as the character progresses acts as a visual progress bar of how the player progresses in an area while expanding the story of restoring the forest.
The spatial architecture also guides players from puzzle to puzzle, allowing the player to follow a linear progression of scaling difficulty with relative ease. These tile puzzles are placed directly in the spatial architecture rather than their own isolated levels ( a la Candy Crush), effectively connecting their outcomes to changes in the environment while justifying their presence through the story. The puzzles are explained within the narrative architecture as the main method of restoring the Book and the forest, and the seed economy is also justified through the story. The puzzles are also much more narratively and spatially satisfying than the optional interactions that do not create magic, rewarding players for their effort to solve them.
Getting Started
After the Fire began as an in-class discussion on the joys of having a small forest spirit as the protagonist of our game, it evolved slowly as our group wired towards a shared diffraction for the game. We started with Moodboard and Spotify playlists, where our group seemed to agree that our game should consist of a peaceful, nature-bound experience, as evidenced by the soft shades of green and tones of satisfaction in our mood boards. Our Spotify playlists contained similar themes of loneliness and peace – a sense of calm after a tragedy. We ultimately decided our game would elicit feelings of coziness and calm with an undertone of grief as our protagonist dealt with the destruction of his home.
Choosing which direction our gameplay would move in was a more challenging task. To achieve some common ground, each team member proposed a few game ideas with mechanics in mind, resulting in a master document of game directions and proposed gameplay mechanics. We passed this document around as a heat map, where our team members could select the game they were most excited about making. Finally, After the Fire was born, a tile-placement puzzle game where a young forest spirit whose home has burned down attempts to restore the ecosystem with his nature-based powers. We decided to create a slice that would compromise the tutorial level and the game’s first level.
This approach used all three kinds of space: narrative space, spatial architecture, and puzzle design. As a team, we divided these into two branches, combining narrative and spatial architecture as one series of playtests and level design as a separate one. Then, we combined the best versions into a final product, which we iterated upon. This post describes our iterations within each version of space and our final combined product.
Narrative and Spatial Architecture
Beginning the story
We knew our story would revolve around a forest spirit rebuilding a forest ravaged by a wildfire. We were unsure if we wanted to treat the fire as a mystery that the player sets out to solve as they heal the forest, but we were torn between an embedded narrative and an emergent narrative where the player embodies the spirit. We wanted discovery and narrative to be the core aesthetics of the game, so we initially started by creating a sense of mystery by not explaining why or how the fire happened.
Iteration 1: Interactable Elements
To evaluate our narrative, we created cards with an environmental object illustrated on one side and a related story fragment on the other. We arranged these cards in front of classmates for a playtest. The players picked up the cards, flipped them over to read the story fragments, and assembled the narrative piece by piece. For one of the cards, we gave the player an option to choose if their character voted the fire spirit out of the forest or not to potentially test an emergent narrative where a player can decide how the protagonist acted in the past.
For our narrative playtest, our questions were:
- How do interactable elements contribute to the narrative?
- What parts of the story were players interested in?
Successes
Our players nodded as they read each card. Even though the story was broken up in a non-linear fashion, they were able to piece together the narrative that the fire spirit was the main antagonist or, at the very least, caused the fire to burn. Embedding the narrative through environmental objects made the players more invested and eager to discover the rest of the story and the forest. Moreover, players liked that they had a sense of control when exploring the story.
Shortcomings and Room for improvement
Due to the lack of architectural design, there was no order in which the players could learn about the story. While this worked in most cases, there was confusion when the player started with the option to vote the fire spirit out/in, as they did not yet know who this fire spirit was. Furthermore, creating an antagonist changed the story’s tone in a way we didn’t like. We decided to keep the embedded narrative aspect but base the mystery around how the fire happened rather than alluding to an antagonist fire spirit. Because the player didn’t interact with the world beyond selecting and reading these interactions, there wasn’t much player autonomy. This also limits the discoverability because of the lack of further spatial storytelling. Moving forward, we wanted to have more structure for discovering the narrative through level design and environmental objects.
Iteration 2: Introducing Spatial Architecture
In the next iteration of this prototype, we chose to keep the interactive objects we presented players with in the first playtest. We supplemented them with a low-fidelity map of our two levels that they could navigate freely using a game piece as their avatar.
With these changes, we aimed to convey more of the story through spatial architecture rather than having the player simply consume the story through clicks. Attached is an image of our initial architecture interaction maps and physical game space that the player moved through.
We adjusted the story slightly, moving away from the emergent narrative that allowed players to choose how their character had acted in the past, opting for an embedded narrative that supplemented the exploration of the world. Since our target audience is casual gamers who enjoy puzzle games, we figured a game with a fixed story would be more welcoming than an emergent role-playing game, but the story’s mystery would motivate them just enough to keep them adventuring through levels. As players moved between puzzles, we hoped they would search for answers to the environment’s fiery tragedy. In this case, we aimed to have the space provide constraints and obstacles for players to overcome since the ruins of a magical forest should feel challenging to navigate. We also aimed to fulfill the psychogenic need for information and the aesthetic of discovery with a spatial architecture that encouraged exploration, which, in this case, meant separating the interactive objects in a wide field by a river and only providing information when the player asked for it.
Our main questions with this iteration were:
- Is the protagonist’s motivation clear when the information is spread across a spatial environment?
- Do people navigate fairly well through our spatial architecture?
Successes
We found that players were excited to interact with everything in the environment and even asked questions about objects not labeled as interactive. Players would tap random parts of the paper, like the river’s bend, asking, “Is there something over here?” On one occasion, one of our playtesters expressed disappointment at not receiving an answer to the mystery within the first level. This iteration catered well to “explorer” player profiles and people with a psychogenic need for information, especially since this puzzle-less prototype mostly involved adventuring through the space we had created. With this, we were confident our game effectively communicated the aesthetics of discovery and narrative. Interestingly, we also found that most players approached objects based on proximity, choosing to interact with closer objects first, a fact that would be useful in future iterations of the map.
Room For Improvement
Acquiring more seeds
In this playtest, we simulated puzzle completion by asking the players to collect tokens and hand them in as loose representations of seeds since it wasn’t what we were testing. However, this resulted in extreme dissatisfaction and confusion as one of our playtesters was increasingly confused about how to earn more seeds as the game progressed. This revealed an underlying disconnect between our puzzle space and narrative since we simply handed out puzzle pieces without providing a narrative justification of their origins. This led to discontent among playtesters, who reported they lacked autonomy when collecting their main resource for story progression. With help from Krishnan, we identified that while our main narrative arc was steady, the loops to acquire more information and resources felt random and undefined, a shortcoming that violated the consistency heuristic of spatial design. To address this in the next iteration, we would standardize how the player acquires more seeds to use in their puzzles so that each player could rapidly identify how to acquire more puzzle pieces if they come up short at an obstacle.
Story Shortcomings
As players moved through the level, we found that they expected the silent protagonist to have as much information as they did, which, in this case, was none. One of our playtesters asked, “Wait, am I supposed to know I can do magic?” We had incorrectly assumed that players would know the character’s backstory as well as we did with the limited information we were providing them, an artifact of the emergent narrative we attempted in the first iteration. The character’s motivation became clear as we explained it. Still, we wanted the player to experience the story with the protagonist, which meant providing them with the same amount of context. To address this, we decided to give our forest spirit a voice and created an introduction scene to provide context for the spirit’s predicament. By communicating the main character’s thoughts through dialogue boxes, players can see how our protagonist feels about the setting, enhancing their emotional connection to the story and facilitating the progression through the main story arc. This also would allow us to communicate the narrative and character motivation to the player more often, allowing the narrative space to dominate even as they explore the map.
Iteration 3: Onboarding and Characterization
For the next iteration, we went one step further, creating a fully analog version of the game with interactive objects, simplified versions of the puzzles, and index cards with writing and drawings as dialogue boxes. We implemented a character monologue and standardized loop process to acquire more seeds through interactions and an introduction scene to familiarize players with the world and context. We also used this as an opportunity to run players through a tutorial level with little input from the moderators, relying on introducing players to magic through plants with action over reading, being careful to blend simple versions of the main interaction loops into the first level, opting to make them simple enough to communicate our main interaction loop, puzzle solving and acquiring more seeds, in just two trivial puzzles. This way, we could implicitly use adaptive messaging through the environment so expert players could breeze through the tutorial without being held back by tutorial text. New players could take their time learning the simple mechanics.
With this, we set out to playtest to answer the following questions:
- Do the dialogue boxes make the world feel more dynamic and guide the player with little moderation?
- Does onboarding in its current state prepare the player for puzzles in the future?
Successes
The onboarding was a huge success, at least in the analog version of the game. Our playtester, who initially tried to breeze through the tutorial puzzles, encountered resistance almost immediately, later reporting, “It was pretty easy when I stopped actually to think about it.” This signaled that our onboarding text and level were resistant to brute force to at least some degree while still allowing for a minimal amount of player elaboration to learn our main mechanics. We also found that players were even more willing to interact with objects when it meant they would receive more information for doing so, reinforcing our finding that players crave more discovery and exploration. Finally, players made the connection that solving puzzles and interacting with objects gave them more deeds to complete puzzles, with our playtester recognizing that they “must have missed something” every time they encountered a puzzle they did not have enough seeds for.
Room for Improvement
Getting Lost
Our main obstacle during the playtest was the players who were lost when they could not solve a puzzle. While they recognized that they had missed something somewhere else in the level, they were usually unable to deduce which puzzles they had to solve first, given that they were given access to all of them simultaneously. This represented a major flaw in our spatial architecture. Even though our puzzles could only be completed linearly, they were not presented to the player as such. In our attempt to give players the freedom to fail and encourage exploration while they searched for answers to the puzzles they could not solve, our spatial architecture was doing a poor job of providing constraint or guidance, leading to frustration. While they had purpose and autonomy, they felt they lacked mastery simply because of the lack of implicit or explicit direction. At the start of the first level, our playtester said, “I don’t even know where I’m supposed to go.” They ultimately chose to follow the river to their left since it was their only continuous structure in view. To resolve this, we created a path on the ground of our first level, hoping players would follow it to navigate the space better. We created a branching path that allowed players to choose which to follow, but they would ultimately be guided to every interactive object should they stay on the path. Their autonomy would be preserved since they wouldn’t be railroaded in any particular direction, and they could explore more effectively or find their way back to an earlier puzzle easily if they missed something. This would benefit our target audience of casual players, who may not be familiar with top-down exploration, while allowing experienced players to explore off the beaten path.
Fidelity Issues
We found that as our playtesters experienced more of the first level, they lost much of their emotional connection to the space as they explored. We suspect that while the narrative text communicated the goals and the character’s experiences well, the spatial architecture was not. The paper prototype map was doing a poor job of communicating a forest ravaged by fire and a home that was destroyed. This meant the architecture could not provide good constraints or a compelling atmosphere as drawing on paper. We would seek to improve this by utilizing a digital, higher-fidelity prototype with sounds and animation for our next iteration.
Iteration 4: Adding a Path and Color
Our fourth iteration introduced the aforementioned path and color to our spatial architecture, allowing players to explore a digital version of the game where they cannot see the whole map at once. The goal of this prototype was to have the spatial architecture provide guidance and constraints while contributing to the atmosphere, and the narrative was presented through dialogue boxes upon completing puzzles. These two types of space would create the aesthetic of discovery and narrative while the puzzles presented a challenge that progressively increased in difficulty. We set out to answer the following questions:
- Will adding a path guide players through the level more effectively than having completely free reign?
- Does the inability to see the entire map at once impact the ability to explore?
Successes
We were correct in assuming that people would follow the path laid out for them, and in most cases, following the path allowed us to overcome the obstacle of having a narrow field of view on the camera. In the tutorial level, one player commented how simple it was to navigate: “There only looked like there was one way to go, so I wasn’t going to get lost.” Because of the path, the spatial architecture guided our players through the level, demonstrating we had finally provided effective constraints and guidance in exploration. Because of this, players were imbued with a newfound sense of mastery that improved their motivation to keep exploring the interesting locations along the path. The path also acted as a great corrective force, as evidenced by one player commenting, “I should have gone the other way, I bet.” when they accidentally arrived at the final puzzle ahead of time and couldn’t solve it.
Room for Improvement
An Unclear Path
While the path did lead players to every important location on our map, such as the puzzles, it came to our attention that almost every player put too much faith in the map to lead them to the correct destination immediately, stopping at the first location the path led: the final puzzle. We expected players to acknowledge the impossibility of the puzzle and move on but were surprised to find that 4 out of 5 playtesters stayed at the puzzle and attempted it multiple times despite not having enough seeds even to fill the squares. This was mostly due to a flaw in the spatial architecture, which led to a difficult location instead of the puzzle they were supposed to solve first. We had incorrectly given the players too much autonomy under the guise of guidance, leading them to believe they must already possess the mastery to accomplish the puzzle. We had unintentionally misled the player through our spatial architecture and quickly realized that we needed to have the constraints of the map mimic the constraints of the puzzle progression. We assumed that players would enjoy encountering an unsolvable puzzle that they would later be able to crush. Still, that assumption simply led to frustration. It decreased motivation as people took the loss as a sign that they lacked mastery, giving them the illusion of a nonsensical difficulty curve. To fix this, we would completely redesign the map so that players would encounter the puzzles in the order they were able to solve them, a choice we should have made earlier, but that would largely benefit casual players who needed a simpler spatial puzzle in the first level of the game.
An Empty World
In the transition from analog to digital, we were struck by the reality of travel time. Our players were navigating through a lot of empty space to arrive at objectives, interacting with almost nothing on the journey. As they moved from loop to loop, they barely interacted meaningfully with the environment or the story arc. The aesthetic of exploration was capped at being able to explore empty space. As our lovely TA Amy put it, “It feels like I’m walking through a lot of nothing before I get somewhere interesting.” To fix this, we would include interesting interactive items between important locations that give the player greater narrative depth to characterize the forest spirit better, give some history on the setting, and give the player something to find while exploring. This interaction would be fully optional and occasionally off the puzzle path, allowing players replaying the game to breeze past these narrative boosters while rewarding players for taking the time to experience the world they’re playing in. This fulfills the psychogenic need for information, rewards “explorer” players, enriches the narrative space, leverages the environmental storytelling, and gives the protagonist more personality to increase the emotional connection to him. There is roughly one optional intractable between puzzles that discusses the main story, leaving players who explore them in an “arc sandwich,” where they get bits of the main story between puzzle loops. The path occasionally winds and branches but only has one entrance and exit, allowing players to still feel like they have autonomy and aren’t forced to walk in a straight line.
Level Architecture
Our choice of puzzle type for our game grew naturally out of our vision for the story. Once we knew we wanted our story to be about rebuilding a burned forest, it made sense for the gameplay to feature planting heavily.
Planting in the real world is a highly spatial activity: you need to consider what environment different plants prefer, how much space each plant will take up, and how different plants can be planted side-by-side to populate all canopy levels with life.
We realized we could create a spatial puzzle out of that dynamic. Inspired by games like Dorfromantik, Castles of Mad King Ludwig, Cacassone, and Cascadia, we knew that a simple set of competing spatial constraints could result in enormous amounts of emergent complexity.
Iteration 1: The Fundamentals
Our first puzzle prototype created these constraints featuring multiple different kinds of plants, each with a unique scoring rule. For example, a tree might want to be next to as many other small bushes as possible, and one of those small bushes might be competing for spaces adjacent to water. The goal was to establish a set of competing opportunities large enough for the player to carefully consider how they could simultaneously take advantage of as many as possible.While we had initially envisioned a large area players could place their plants into, potentially cultivating an entire forest’s worth of bushes and trees, we ultimately decided to push the boundaries in our first draft to test this assumption by using the “rule of two.” The rule of two is based on the idea that if you adjust a value to what you think the right number is, you get very little information. By picking from two extremes with a gap in between, we could adjust the puzzle’s size within that range to achieve the intended effect. If we needed a much larger or smaller puzzle, we eliminated the range of unsuitable values.
Applying this technique to our own puzzles, we wanted to see how small a grid could make for a satisfying puzzle and make you feel like you were restoring a forest. The goal was to provide our target audience with a small amount of challenge without overwhelming them with a high bar for mastery. So, we halved our planned puzzle size to just six tiles.
Our goals when playtesting this iteration could be summarized by the following questions:
- Would players feel intimidated by a puzzle that mostly involved using math as a solution?
- What would be the best size for a puzzle that could be achieved in small periods of time?
Successes
As we found in our first test, six tiles turned out to be just about the right size, and the puzzles we included in our final game ranged from one to seven tiles in size. The small space made every placement impactful and rewarding. On top of that, the game was more approachable — new players had a reasonable number of options to consider for where to place their first plant. Our target audience involves casual players seeking relatively simple puzzles, and this size seemed to provide players with challenges and fulfill the psychogenic need for achievement in a small amount of time.
Room for Improvement
Calculating and Tracking Score
While our first playtest overall went quite well, it wasn’t an unmitigated success. The initial system contained several plants that affected those around them. While this does provide a lot of design space to explore, it created an enormous cognitive load. When our playtesters went to figure out how many points a specific plant was currently getting, they needed to scan across their entire board for anything else that potentially modified it. One player sat there and waited while we calculated the score in our head for them, an awkward experience that prolonged the game unnecessarily.
While we knew this would be mitigated somewhat by a digital program that could calculate scores for the player, we still felt that this kind of tracking complexity didn’t serve the core spatial puzzle we were aiming for. As such, we implemented a principle going forward: Plants would only affect their own score. Rather than a Starleaf Tree increasing the value of all your Moonglow by two, it would be worth two points for every Moonglow on the board. This is much easier on the player because they only have to remember the rules for how plants are scored one at a time when they’re considering that plant directly. This allowed the player to track their progress through the puzzle satisfactorily and encouraged players to place before doing calculations in their head, reducing their cognitive load. Plus, once a digital interface was available to show them how many points each plant was generating, it would be much easier to assess where those points were coming from if they could only come from that plant’s scoring rules.
Iteration 2: Experimenting with Cognitive Load
Whereas our first test tested the general concept of tile placement puzzles as implemented through a side of plants with competing scoring rules, our second test tested how the players acquired those plants and how those puzzles could be integrated with the rest of the story.
The first playtest used a draft system, where players would receive a choice of seeds each time they placed a plant. This created some fun design space. We could design plants that are good in multiples, and the player would need to gamble on whether they would be offered enough copies to make it worth it.
However, concerns emerged that this method was too separated from the game’s overall story. By contrast, if players needed to interact with their environment to get seeds, the puzzles and the story could flow into each other seamlessly. We could integrate the experience with narrative justification while rewarding the player for completion.
Still, we foresaw a potential problem with this approach. As we know from lecture, more options are often worse. Players can quickly get paralyzed by overload. Even once they make a decision, they’re less satisfied with it. If players got all their seeds before planting (often 6-8), that could be too many choices, especially when compounded by the variety of spots they could place plants in. Meanwhile, humans can only hold 3-5 items in working memory.
With this playtest, we wanted to push against this notion of cognitive load by giving our players too many options for their first plant. We set out to answer the following question using a small amount of spatial architecture, a location where they could collect more seeds whenever they wanted.
- Would providing the player with more seeds than working memory can handle cause cognitive overload?
Successes
Thankfully, our fears did not come to fruition. We presented the player with a grid they could place in and a set of locations they could interact with. Our testers reliably switched between these options, placing seeds as they acquired them, keeping the options available at once fairly constrained.
However, to ensure our game worked even if players did not follow this pattern, we also gave them all the plants at the same time. This did cause a notable increase in the amount of time before the player made their first placement, but it wasn’t egregious. Players still got started within one minute and finished the puzzle within five, an excellent outlook for our target audience.
We suspect there are a few reasons cognitive load wound up being manageable:
- We allowed players to pick up and rearrange plants at any time. This made them feel comfortable guessing and experimenting because their choices weren’t locked in, leveraging the aesthetic of discovery through experimentation.
- While players had many plants, they had just four unique types to work with in this playtest. Multiples of the same plant could be grouped together cognitively, reducing the load.
- Some plants had fairly straightforward rules, like a tree that wanted to be placed in a location adjacent to as many other plants as possible (place it centrally) and reeds that wanted to be next to water. Players found these a good place to start and build up from there. Once they used that plant a few times, recognition took precedence over recall.
Learning from this, we decided to keep what worked going forward. We allowed players to rearrange as much as they liked, limited our final Level One to just four plant types unlocked one at a time, and kept two of our plant types easy to place first. We also decided to commit to players receiving seeds from intractable story elements based on our success here, as the associated rewards made them much more excited to experience the narrative architecture while exploring the spatial architecture.
Iteration 3: Moving to Digital
As we switched from paper to digital, we wanted to test how the transparency of the game state changed and how players responded to our visuals and controls.
We set out to answer the following question:
- Would introducing a digital interface and automatic score calculation improve performance by reducing cognitive load?
We hypothesized that having the computer calculate the player’s score for them would substantially reduce the player’s cognitive load and make the overall experience more pleasant.
Room for Improvement
We couldn’t have been more wrong. On the paper version, the effects of each plant were printed directly on each card, providing easy reference. Plus, the player needing to calculate their score made it very clear how much each plant contributed to the total and what moves could improve that. As the player calculated their own score, they elaborated on the effect on the plant, committing it beyond their sensory processing.
When the score was calculated for them, players skipped this step and were often unclear about how much each of their plants contributed to their success. Once they reached this state, players sometimes calculated the score themselves (a task now made more frustrating by the lack of an easy reference), or, more commonly, just moved plants around based on guesswork until they got a solution.
The problem here was classic HCI. We’d violated two of Nielsen’s usability heuristics simultaneously: the visibility of system status and recognition over recall. Players couldn’t see how the calculation was done and needed to remember how every plant worked.
Based on this experience, we began implementing several changes to our interface, but they took some time to complete. In the meantime, we provided the player with a paper sheet detailing each plant’s rules, which created an immediate improvement. With the sheet available, our players consistently solved a seven-tile, three-plant-type puzzle in approximately two minutes. They reported high satisfaction with its difficulty level, demonstrating that we were providing bite-size challenges to an audience unfamiliar with this type of puzzle.
Our final product implements the following to solve the issues we found here:
- When near a puzzle, our UI shows a sidebar with the rules for every plant prominently displayed. This allows people to easily reference plants to reduce their cognitive load.
- Players can hover over any plant to see how much it is contributing to the score, a tool that allows for recognition over recall.
- Each plant features a puzzle meant to highlight how it works before applying it in a more complex setting. This allows players to develop a feel for each plant before needing to apply that knowledge. For this, we used the onboarding technique of letting players use new mechanics before challenging them with them.
While these fixes allowed us to achieve recognition over recall, the visibility of the system status is, in our opinion, still one of the largest weaknesses of the final product. If we had more time available, we would have created visual indicators of how the computer arrived at each plant’s score (such as highlighting the tiles of plants contributing to its score when the player hovers over the plant).
Iteration 4: A Series of Puzzles
In our final playtest of the puzzles, we unified the puzzles and gave the player a single level consisting of them in our spatial architecture. Here, we sought to explore how players interacted with puzzles when allowed to move between them. We set out to answer the questions:
- Can players successfully complete the puzzles when presented with them as they are placed in our spatial architecture?
- Do players use losses in more difficult puzzles as a signal to try another puzzle instead?
Successes
Players struggled with each puzzle somewhat, but usually not more than a few minutes, and reported that the difficulty increased at a rate that “kept me in a satisfying flow state once I figured out what I was doing.” The progression of difficulty when the puzzles are completed in the correct order increases at a simple pac, and players are guided by plants already placed in some of the puzzles to use as hunts for how to complete them.
Room for Improvement
However, it is the nature of puzzles that you don’t know the solution until you’ve found it, and as such players would get stuck and frustrated working on a puzzle they didn’t yet have the plants to solve. With one playtester, once they gave up on one puzzle, they would leave their plants behind on an unsolved puzzle instead of taking them with them, leaving them with insufficient resources to solve even the easy puzzles. This lack of guidance created a sense that the player lacked mastery, leading to a visible loss of motivation.
While we could have revisited the method by which players got seeds to solve this problem (and even considered giving them unlimited seeds as an easy fix), the system of interacting with the story for seed rewards was working so well that we decided to solve this problem by changing how we used space rather than changing the puzzle system.To that end, we made the path through the level functionally linear, taking players past puzzles in the right order. While previously we had aimed to show players some puzzles early on that they could not yet solve to get them excited to come back and solve those big ones once they had enough resources, we now moved those large puzzles to the very end, decreasing autonomy but improving the player’s sense of accomplishment and progression. To handle the case where the player ran to the end of the level anyway, we blocked off the final puzzles with a dialogue prompt and a large log as an environmental hint that other steps had to be taken before the puzzle was solvable.
We also came away with many minor takeaways, all of which we’ve addressed for the final project, including:
- Teaching players about the hover for mana option
- Make the mana display on individual plants correspond to the overall mana count
- Teach the players directly about Rich Soil
- Inform players that they have failed when they fail to solve a puzzle
- Make the distinctions between different puzzles clear
- Make all puzzles contiguous
- Allow players to bring up a menu to show the controls
- Don’t use the number keys 1-4 to represent slots in a vertical inventory hotbar
These changes make us incredibly proud of how our puzzles use space. They create an interlocking optimization puzzle network that frequently produces satisfying insights and beautiful patterns. Players are rewarded with audio, visuals, and an embedded narrative to make the completion of a puzzle fulfilling.
The Final Playtest
For our final playtest, we combined each of the three kinds of space into a mostly finalized product that implemented:
- Narrative Architecture: An intro scene, interactive puzzles and objects, and character monologue to communicate story
- Spatial Architecture: Fully realized maps, animations upon completing puzzles, a linear path accented by interactive objects both along it and off it
- Puzzle Design: Linear completion structure, hints on failed maps, easily accessible reference sheets, and clear rewards
With a final product in place, we set out to do a final playtest that answered the question:
- Can our game be experienced without guidance from a moderator?
Successes
As evidenced by the video of our final playtest, our playtester enjoyed the animations and puzzle-solving capabilities the most. They could progress through the game without becoming stuck (save for a minor bug that was rapidly corrected). The issues from past playtest involving poor spatial architecture and puzzle complications were largely resolved, as evidenced by the playtest recording.
Room for Improvement
In this playtest, there were several instances where the playtester was not interested in the narrative architecture of the game. However, this is likely not because the story is not compelling. Because this is only one playtest and this playtester may not be a part of our target audience that enjoys slowly experiencing a narrative, their opinion on the reading elements of the story would need to be studied further before making changes to the game. Largely, playtesters have requested more information and narrative throughout our iterations, so the goal of eliciting narrative and exploratory fun through narrative and spatial architecture stands.
For future iterations of the game, we would adjust some of the minor bugs that occasionally slow down the game, but we are happy with where the game is at the moment. With further technical capabilities and more time, we could supplement the narrative architecture with cut scenes or flashbacks to break up some heavier text-based portions, such as including a cut scene between scene changes to give more context on the book.
System Design
Accessibility
As a game that relies on sorting and arrangement of different types of plant life, we were careful to create assets that relied on multiple visual differences rather than simply using color to make our game accessible for people who have different types of color vision deficiencies.
1.Each of our plants, which serve as our main puzzle pieces, has a distinct shape that sets them apart by more than just color.
2. Our textboxes include changes in brightness and the inclusion of a book icon in the top left corner to differentiate when text is in the book as opposed to spoken by the character.
3. Despite burned ground having a similar hue to the fertile soil, we differentiate the interactable soil with high levels of contrast to make the interactive ground visually accessible.
4. Animations that rely on color still produce non-color-based effects like changing tree shape and a rising river, along with audio cues.
All art assets, aside from the main menu, were made using Aseprite by our team. The main menu screen was made with the assistance of ChatGPT, Aseprite and Photoshop.
Sound attributions list:
- S: rushing river in the woods by CastleofSamples | License: Attribution 3.0
- S: plants-rustle.wav by AlessandroSposetti | License: Attribution 4.0
- S: Bush Branches Rustling 1 11 by Sheyvan | License: Creative Commons 0
- S: Big Drawer by andersmmg | License: Attribution 4.0
- S: bubbling fountain.wav by cognito perceptu | License: Creative Commons 0
- S: extinguishing fire by soundslikewillem | License: Attribution NonCommercial 4.0
- S: FireBurning_v2.wav by pcaeldries | License: Attribution 4.0
- S: Kampina forest spring008 190322_1321.wav by klankbeeld | License: Attribution 4.0
- S: Magic Stars Retro Sparkle by smokinghotdog | License: Creative Commons 0
- S: Royal sparkle whoosh centre.wav by alexkandrell | License: Attribution 3.0
