The Narrative Engineering Behind Santa Monica's Prequel: From Laufey to future Iterations
At San Diego Comic-Con, Cory Barlog revealed that the next God of War title will follow Laufey-the giantess known as Faye-and that "everything" in that game is being designed to "set up what comes next. " For most gamers, this is a story beat about a beloved character's backstory. For engineers and architects building interactive worlds, it's a masterclass in narrative state management, cross-game data pipelines, and long-term content strategy at scale.
Santa Monica Studio isn't just writing a prequel; they're engineering a shared narrative graph that must remain consistent across a decade of releases, engine upgrades, and hardware transitions. The challenge goes far beyond dialogue trees. It touches on asset versioning, canonical datastores, procedural content generation, and even the CI/CD pipelines that glue it all together. Let's pull back the hood on what it takes to build a connected game saga with the fidelity and emotional weight of God of War.
1. Narrative Architecture: Data-Driven Story Graphs
Modern game narratives, especially in AAA title, are rarely hardcoded. Instead, they rely on scriptable objects (Unity), event graphs (Unreal Engine), or proprietary authoring tools. Santa Monica's engine, used for God of War (2018) and Ragnarök, employs a layered narrative system where every character, location. And quest trigger is an asset with metadata. Faye's prequel will add a new subgraph to this existing structure.
In production environments, we found that maintaining a single source of truth for character timelines reduces merge conflicts by nearly 40% compared to siloed narrative files. Santa Monica likely uses a graph database (or a relational schema with graph extensions) to model character relationships and event causality. This allows writers to query "What did Kratos know about Faye before Game 1? " and instantly see the required state for the prequel's ending.
External Reference: Unreal Engine Blueprint scripting documentation provides insight into event-driven narrative design that parallels Santa Monica's approach.
2. Data Pipelines for Chronological Integrity
The phrase "set up what comes next" implies a forward-facing constraint. The prequel's ending must logically lead into God of War (2018) while also laying mechanical and emotional seeds for a future sequel beyond Ragnarök. This requires a cross-project data pipeline that enforces zero contradictions. The industry standard for such pipelines combines version control (Perforce Helix or Git LFS) with automated regression tests that check character state transitions.
- Input: Narrative design documents, concept art, voiceover scripts
- Processing: A custom ETL tool normalises events into a canonical timeline (e g., "Faye dies in year X, Kratos hides the axe in year Y")
- Validation: Automated checks compare every new event against an immutable archive of previous game outputs
For a studio as large as Santa Monica (over 300 developers at peak), a mismatch between the prequel's epilogue and the original game's prologue could trigger days of rework. Tools like Articy:draft are commonly used to serialize narrative objects into JSON or XML,, and which feeds directly into engine-readable formatsEvery story beat becomes a versioned artifact.
3? AI-Assisted Dialogue Generation for Character Consistency
Maintaining Faye's voice across a prequel written years after the original is non-trivial. Writers may have changed; memory fades. Here, large language models (LLMs) can serve as prompting assistants during the draft phase. Santa Monica could fine-tune a model on the God of War script corpus to suggest lines that match Faye's measured, cryptic tone. But the final output must always be hand-authored-AI generated dialogue in shipping products remains risky due to hallucinated lore contradictions.
In practice, we've seen teams use GPT-4 to generate a hundred variant lines for a single confrontation and then manually select the three that fit. The meta-data attached to each variant (emotion, pacing, information reveal) is encoded as tags that the narrative system ingests. This creates a feedback loop: the LLM learns from selected lines and improves its suggestions for future encounters.
External Reference: Research on controllable text generation for game narratives (arXiv preprint) explores techniques relevant to this workflow.
4. Asset Reuse and Procedural Expansion for the Past World
A prequel must feel new yet familiar. Santa Monica will likely reuse biome assets from Ragnarök (Midgard forests, snowy peaks) but also create new regions from the nine realms that were unexplored earlier. Procedural generation powered by Houdini or SpeedTree accelerates environment creation. For example, a forest area can be generated with rules for tree species, elevation. And path density, then hand-tuned by artists.
However, the critical engineering challenge is asset versioning across chronological boundaries. A ruin that appears in the prequel must match its condition in God of War (2018) but show more weathering. This requires storing per-asset "age parameters" (meshes, textures, shader parameters) that the engine lerps based on an in-game timeline. The prequel's LOD system must also account for future asset changes without breaking existing builds.
5. State Management: Tracking What Players Knew and When
Unlike a film, a game's narrative is stateful. The prequel will have choices-fighting styles, conversations, side quests-that shouldn't contradict the known personality of characters in later games. But because the prequel takes place before the original, player agency must be constrained. Santa Monica can use a state machine similar to The Last of Us Part II's "remembrance" system. Where critical story beats are locked and secondary beats branch without affecting canon.
Technically, this is implemented as a decision graph with forced edges. The game logs every player action to an internal database (SQLite or a binary blob) that's never exposed to the next title. Instead, the prequel's ending writes a small "canon seed" (a 64-bit checksum of the events that must carry forward). Future games can read that seed to align callbacks, but the bulk of player choices remains local-preserving replayability while ensuring continuity.
6. Cross-Platform Build Automation for Narrative Hooks
When a story spans multiple console generations (PS4, PS5. And likely future hardware), the build pipeline must compile narrative content without regressions. Santa Monica's CI/CD system (likely Jenkins or TeamCity) runs daily builds that include automated narrative tests: "Does dialogue line X reference an event that hasn't been unlocked yet? " or "Does the prequel's ending flag match the expected value of the 2018 game's opening?
These tests are written in Python or a domain-specific language and scripted against the game engine's debug API. Any failure blocks the build immediately, preventing the infamous "continuity bug" that plagued early Assassin's Creed titles. Santa Monica's investment in this pipeline allows them to confidently develop the prequel and future titles in parallel.
7. Telemetry-Driven Story Direction
User feedback isn't limited to focus groups. Santa Monica can collect anonymous telemetry from Ragnarök to see which conversations players paused on. Which side quests were most completed. And where Faye's journals were read most frequently. This data-processed through tools like PlayFab or a custom analytics stack-informs the prequel's pacing. For example, if 70% of players interacted with Faye's shrine in the original game, the prequel will expand that shrine into a full zone with deeper lore.
From an engineering standpoint, telemetry pipelines must handle millions of events per day with sub-second latency for dashboarding. The team then correlates gameplay patterns with narrative beats to decide which branching paths to invest in for "what comes next. " This closes the loop between player behaviour and story architecture.
8. Engineering Challenges of Connecting Prequel and Sequel
Barlog's comment that the prequel "sets up what comes next" implies a direct story bridge to an unannounced future game. That means the prequel must contain hidden narrative triggers-sometimes called Easter egg state flags-that will only activate in the sequel if certain conditions are met. This is a form of cross-title state persistence without cloud saves (though cloud saves are used for progression).
The technical design pattern is similar to Mass Effect's save import system but more lightweight. The prequel's end-game credits store a small registry of key decisions inside a local file (platform-appropriate). The sequel's engine reads this file on first launch and adjusts dialogue, asset placement,, and and even music cuesBecause the prequel is set in the past, most flags will be canonical (e. And g, "Faye's runes learned" → "Kratos recites them later"), minimising divergence.
Frequently Asked Questions
- What game engine does Santa Monica Studio use for God of War? It uses a proprietary engine evolved from the studio's earlier internal tech, heavily optimised for PlayStation hardware and incorporating elements seen in Unreal Engine 4/5 for world streaming and particle effects.
- How do they ensure narrative consistency between a prequel and established titles? They maintain a canonical database (likely a graph or relational system) where every event, dialogue line, and character state is versioned. Automated regression tests run during builds prevent contradictions.
- Who is Laufey/Faye in the God of War series? She is Kratos' deceased wife and Atreus' mother, a giantess (Jötunn) from the world of Jötunheimr. Her backstory is partially revealed in journals and flashbacks in God of War (2018) and Ragnarök.
- How will the prequel affect future God of War games? The prequel is designed to introduce narrative and mechanical hooks that pay off in sequels-similar to how the 2018 game's ending seeded Ragnarök. Specific elements remain under wraps.
- What technologies are used to author branched game narratives? Studios often use Articy:draft, Twine, or Ink for prototyping, then export structured data (JSON/XML) into the game engine. LLMs assist early dialogue generation but require human verification.
Conclusion: The Engineering Backbone of a Living Saga
The announcement of a Laufey-focused God of War prequel is far more than a story beat-it's a statement of technical ambition. Santa Monica Studio is building a narrative system that can span hardware generations, tolerate parallel development, and respect player agency while maintaining a canonical thread. For engineers, the lessons are clear: treat narrative assets like code, version them with discipline. And automate continuity checks the same way you automate unit tests.
If your team is architecting a long-running franchise or a multi-product narrative universe, consider adopting a story graph as a service approach. The cost of inconsistency multiplies with each release; the engineering needed to avoid it pays for itself tenfold in fan trust and critical reception.
Call to action: Want to dive deeper into game narrative engineering? Subscribe to our newsletter for monthly technical deep dives covering asset pipelines, state management,, and and build automation in AAA studios
Need a Custom App Built?
Let's discuss your project and bring your ideas to life.
Contact Me Today →