Pioneering an innovative design framework for a community of rocket scientists.

Kerbal Space Program is a realistic physics and engineering simulation game where players design and fly their own spacecraft while managing a spaceflight programme.

Though the game’s modding community has existed for 12 years, modders rarely displayed design intent resulting in an fractured, inconsistent ecosystem that provided little value to players.

Key achievements

Engaged with technically sophisticated subject matter, internalising key technical concepts in the realm of spaceflight engineering.

Invented a quantitative, data-driven, player & user-centric design framework of developing and balancing new mod parts that worked across systems and were actually fun to play!

Innovated on internal workflow practises of developers by developing new authoring tools, speeding up asset creation by 30-50%.

Developed and launched my own mod applying this framework, adding 11 new rocket engines and attaining 14,000+ downloads while showcasing the end-user gameplay benefits of this novel approach.

Mentored 2 community members in the framework who have now developed and launched their own mods.

1,009

Downloads in the first 10 days

14,000+

Total downloads as of March 2025

“You approach stuff with a seriousness and introspection that I've never really seen in anyone else before.”

— Fellow modder commentary

“Beautiful engines. I look forward to using them in my career save!”

— Player feedback

“Shoji, literally of the people I know, you are one of the ones most deserving to be in the industry...I just see it with how you've handled every project you've taken on.”

— Fellow modder commentary

“You've actually nailed the Near Future [mod suite] art style so fantastically and added your own flourish. So good.”

— Fellow modder commentary

“I'm addicted to these engines, you have no idea.”

— Player feedback

“I've seen your other art and loved it. This, too, is fantastic. Nicely done—you've gotten me to consider playing KSP again.”

— Player feedback
The Context

Kerbal Space Program (KSP) is a strategy sandbox game where players are challenged to design and fly their own spacecraft in a realistic physics simulation.

The highly technical nature of this game has granted it a very passionate modding community, many of whom are often experts in aerospace engineering or video game industry professionals.

The Problem

Mod authors often worked in silos with little coordination, so new mods lacked clear gameplay design and intent.

New mods were often made based on guesswork and intuition. Over the course of 13 years this led to a modding landscape that, while populated with visually compelling creations, still often struggled to improve player experiences...

Poor game balance

Modded parts were frequently over- or under-powered compared to other parts, even within the same mod.

Lack of meaningful choices

Mods unintentionally reduced player choice, as some parts were so powerful (or so weak) that they became no-brainers.

Inconsistent design between mods

Players often mixed and matched mods from different authors, who often approached gameplay in very different ways.

Not strategic, just tedious

Players spend lots of in-game time trying to understand what these modded parts even did in the first place.

Unclear gameplay intent

Mod authors developed based on guesswork and intuition, with few guiding principles or means to evaluate success.

Limited and unfocused impact

Without direction, mod developers waste time on ineffective designs, while reducing the quality of user-generated content.

The Challenge

Develop an engine parts mod for KSP that put compelling strategic gameplay first.

There were three goals that I'd need to hit with my mod to rise to this challenge:

Create a system for evaluating balance

Not only would this provide a means to ensure the mod expanded on the strategic gameplay choices instead of limiting them, but would also provide a means to identify unaddressed gaps in the modding space. It'd also better ensure cohesion with many existing, popular mods, which in turn would make it more natural-feeling for players to install.

Embody functional visuals

The visual design of parts in the game need to communicate their function and purpose to the player, helping them navigate the complexities of the game's building mechanics and allow them to focus on strategising and vehicle design. Oh, and these parts would have to look real good alongside their peers, too!

Provide a replicable process

If successful, I'd want to make sure that myself and others can leverage this same process, helping to raise the standard of quality for our creations in the community.

To ease adoption among the very technical-minded playerbase, I created stylish infosheets for each part in the mod explaining their stats and features.

The infosheets incorporated a visual branding style for the fictional in-universe manufacturer and illustrative 3D renders of each part. These are made available in the published forum thread for my mod.

Impact 1

My key innovation: a new systemic method for evaluating mod balance where none existed before.

This quantitative graphing method provided a repeatable, quick, and consistent means of evaluating the balance of my in-game engines, ensuring they supported compelling gameplay across the mod ecosystem.

This framework has been adopted as the standard method of identifying engine balance issues or new part opportunities within the modding community, and to help characterise engine archetypes, breaking down complex stats into easily memorable concepts.

Impact 2

I've set a new standard for process documentation and instructional materials in the community.

Prior to making CEX, our community didn't have a good, up-to-date guide on how to mod new engines into the game.

As I recorded my process throughout making the mod, I was able to leverage my documentation into an end-to-end tutorial that has served as a reference guide and tutorial for new modders.

This has since been transformed into a GitHub wiki that serves as our single, standardised knowledge base, and I continue to assist in its maintenance and growth.

Impact 3

championed the use of procedural generation that expedited the mod authoring workflow.

I developed a number of procedural tools in Blender that auto-generate textures and 3D geometry key to our community's "house style".

Combined, the new tools turned steps that took 2-3 days into a process that could be done in 30 minutes, representing a time savings of about 20% when authoring a part.

Tools I developed include:

  • Procedurally generated edge highlighting, auto-generating edge details for texturing 3D models
  • Procedurally generated pipes with highly configurable, non-destructive customisations
  • Procedurally generated struts, also highly configurable non-destructive geometry
Impact 4

I gave back by mentoring two players, both of whom have now released their own successful mods.

Thanks for checking out my stories.
Let's do excellent work together.