About the Project

"Treasure Chase" is a custom level made for Starfield in Creation Kit. It is a narrative-focused quest that features exploratory elements and dialogue trees.

Made in Bethesda's Creation Kit editor.

Design Deep Dive

I had two design goals for this level: make a quest with exploratory elements and dialogue trees. I wanted to explore the player talking to NPCs in any order and unlocking new dialogue trees as they explore the space, find key items, and bring up new subjects learned from talking with other NPCs. I achieved this through first crafting the narrative and looking at how gameplay elements could be used to support this narrative. Combat encounters were used to show not tell the constant battles the town NPCs face against relentless bounty hunters, and the key items the player finds in each NPC's house provides further narrative context into their life and relationship with the other NPCs. To emphasize the exploration aspect of the level, I placed chests with various supplies throughout the town the level takes place in in addition to the key items in each NPC's house. The level quest features two endings, a choice I made to give the player more agency.

Level Post-Mortem

What Went Well

I was able to design a narrative that was clear yet still had depth. I did a good job of pacing the narrative beats and gameplay elements, and had non-required dialogue trees to reward players who wanted to dig deeper and learn more with more narrative context. I did not bog down the quest with multiple gameplay mechanics, and the gameplay elements that were added enhanced and/or supported the narrative. Ultimately, I view this level as having a successful narrative-focused quest that provided a satisfying story with a good mix of purposeful gameplay interspersed throughout!

What Went Wrong

The level layout feels somewhat plain. The level play-space consists of five main buildings that are all on the same ground level. There is no change in elevation, and the layout of the buildings is not true to a real life town. Granted, the focus of the level is the narrative quest so the layout is not of utmost importance and the layout does fulfill its purpose of clearly defining each NPC's house. Still, given more time, I would have loved to create a space that is more interesting to explore and is more akin to real-life "cowboy" towns as was my aesthetic point of reference.

I did not lock down the narrative before starting development. I developed the level over the course of 8 weeks, which is not a lot of time. I gave myself one week for planning before I moved into building the basic layout in editor. Unfortunately, as I was creating the quest, I was reworking parts of the narrative based on peer and playtest feedback. The changes I made to the narrative made the story stronger, but it meant that I had to go back and change parts of my quest, such as pacing and flow, to better suit it. This ate up development time that I could have spent on aesthetics and polish.

Lessons for Next Time

Moving forward, I will be planning and iterating on the narrative early, and definitely before I get anything built in editor or engine. Only once the narrative is at a good spot will I lock it and proceed with building the level geometry itself. Narrative is the same as design in the way that both need to be solidified before giving other disciplines (such as engineers) the go-ahead. Changing narrative, like changing design decisions, during development is costly.

Narrative-focused levels can still have interesting layouts. Even if a level layout fulfils its basic purpose, that doesn't mean it can't be expanded upon. One solution is to add verticality to add more dimension to a layout. The layout of a building or space can be altered to make it stand out to the player and help them develop a mental map of the play space. This not only helps the player navigate the level better but it adds visual interest.

With this project done, time to take these lessons and apply them to my next project!

About the Project

"Run Me Up" is a custom level made for Half-Life 2 in Hammer focusing on vertical ascension.

Made in Hammer editor.

Level Layout Map

I used Adobe Illustrator to make the level layout map for my level.

When translating anything from map to actually in engine, there are always challenges. This is because when making a map, there are no limits because it's translating from imagination to pen and paper. This means it's easy to design too big or too small for the actual level. The only way to really see what a level will look like is to get it in engine as fast as possible, which is what I always strive to do. The 2D map is helpful in knowing where I want puzzles or combat to be placed throughout and the overall flow, but building it in engine is the best way to get a sense of scale, difficulty, and whether the level actually is good or not. From there, it's all playtesting and iterating to get my level to be as good as it possible can.

Level Post-Mortem

What Went Well

I was able to design a linear level that flowed well as the player progressed vertically through. I achieved this by giving the player only one path forward so they wouldn’t get lost. I used both indoor and outdoor sections so that the player wouldn’t be in one area for too long and give it a nice sense of progression. This included indoor and outdoor stairs, and indoor and outdoor sawblade platforming. This met my design goal of always climbing up while including outdoor parkouring.

What Went Wrong

I realized that I tend to design very large spaces. When replacing blocks with real assets like furniture, the rooms that once felt like a good size felt way too big. I had to shrink rooms by blocking off sections and making new walls to make them feel appropriately sized, which ate up time during the Aesthetics milestone so I had less time to make my level look pretty.

Another issue I ran into was that the combat felt one note. The level focuses on sawblade platforming so I wanted to include sawblades everywhere to nail in the theme of the level. The combat sections were designed to provide half and full cover that allow for circular flow or gameplay. The half cover would prevent the zombies from getting to the player but not stop the headcrabs from jumping at them. However, even with the introduction of the poison zombie in the second and third combat encounters, nothing felt particularly special about the combat. Food for thought.

Lessons for Next Time

I realized aesthetics take a long time to make everything look good. Even if I made a smaller level, there's a lot that goes into making a level look pretty other than just swapping out textures. Small things like the set-dressing of a space, making it look lived in or at the very least full of enough stuff to not be empty, and filling in the outdoor areas to make the level look like it's a part of the world and not an isolated space. Next time, I would allow myself enough time to set-dress my level because aesthetics do matter!

Another lesson is I need to be aware of how big everything is before starting to build in-engine (aka metrics). It saves me headaches later because I know how big the player and props are when building the level so things will be appropriately sized from the get-go.

One final thing is I would lean into different types of cover (half, full) for the player and explore different types of combat, not just what I’m familiar with (circular flow). This allows for the player to explore different types of combat gameplay while playing through my level.

Now that that's all said and done, time to take these lessons and apply them to my next project!

About the Project

A collection of 3D first person levels made in Unreal Engine 5 to explore designing levels under different constraints.

Made in Unreal Engine 5.