This was a level I started in my University's Experimental Level Design Module, the goals of that module were to use a new editor, so I challenged myself with Radiant for COD 4 using written scripts in GSC and older bsp workflows. Also wanting to explore stealth combat with the more linear nature of the Call of Duty series feeling this was as good a place as any, giving myself the added challenge of placing all this gameplay in the real world location of Kenilworth Castle.
The project itself took a total of 17 working days (8 hours a day). With design tweaks, code refactoring, rerecorded audio and additional meshing work coming after the module was over.
What I did:
First and foremost this project was my first proper experience learning written script for games as opposed to the visual blueprinting seen in engines like Unreal. With that in mind I felt it was worthwhile having a short video discussing some of the ways I've gone around writing some of these scripted sequences in a timestamped video below:
(The stab encounter is the real meat of the video so feel free to skip ahead to that section.)

The stealth mission opens with some tension, suprising the player with a dog enemy early, showing the threat they can pose before the player faces them directly.
This moment also reinforces the level's core theme, of Price learning to put trust in the player.
The player's desired path in the gardens being blocked with the use of a smoking guard. Using simple AI placement and geometry to show the solution, that of the hole in the fence.


The enemies run past the stairs, telegraphing the next cinematic beat, adding an element of reactivity to what is otherwise a scripted sequence.
Once again Price doesn't take these targets alone, but with the help of the player.
A sniper occupies Price's future vantage point, this providing another narrative beat whilst also acting as a navigational pointer as to where Price will provide cover from for the rest of the mission.

After a lot of iteration and testing I ended up with various fail states for the different sequences, allowing player agency with how they tackle these cinematic beats, often having an intended experience but not completeley failing the mission if they didn't follow the desired path. Below you can see some examples for the stairway encounter of how I tackled these fail states.
Price having assumed the position of the sniper it's up to the player now to drop the stealth and push his way through to the documents. That same sniper spot providing the player a clear view of the encounter space before the player enters it.
Placing player and enemy cover in a way that gives the player to push forward and control space in the area, clearing a path towards the documents.
The encounter starts with the dog enemy rushing towards the player as a nod to the intro, before the rest of the guards are alerted by this and spawn in seconds later.




This was an extremely fun project to work on. Looking through developer code to figure out what was a largely undocumented engine gave me a lot of perspective on how industry writes mission script, especially compared to modern tools. Implementing the stealth mechanics from scratch also gave me a much clearer idea on how games create tension and suspense through those scripted sequence.
At this point I'm looking to challenge myself in other ways, however if I were to continue working on this project here are some things I would be looking to change:
