Thermite Design Integration
Design Pillars (Non-Negotiable)
Every design decision must serve at least one pillar without violating others:
1. Meaningful Risk
Every action matters. Gear has weight. Victories feel earned because defeat has consequences.
Guardrails:
- No insurance that always returns everything
- No pay-to-skip-risk monetization
- Death must sting, but not devastate progression
2. Readable Chaos
Chaotic situations remain parseable. You can always see what killed you.
Guardrails:
- Grid-based movement and placement (where applicable)
- Clear visual language for threats
- Distinct audio cues
- No "died and don't know why" moments
3. Compressed Tension
Match length creates arcade tension. Every second matters.
Guardrails:
- Target match length: 5-8 minutes (adjust per project)
- Maps sized for constant engagement
- No safe "wait it out" strategies
- Extraction windows create urgency
4. Earned Mastery
Skill expression through knowledge, positioning, timing. Not twitch reflexes or gear disparity.
Guardrails:
- Gear provides options, not guaranteed wins
- Map knowledge and timing are learnable advantages
- Movement and placement skill ceiling must be high
- No "stat check" encounters
5. Sustainable Economy
The game lives or dies by its economy. Too punishing = players quit. Too generous = gear loses meaning.
Guardrails:
- Multiple viable playstyles
- New players can progress
- Veterans have meaningful goals
- Economy exploits will emerge; design for patchability
The Creative Team - 8 Expert Personas
🎮 SHINJI TANAKA - Classic Arcade Game Designer
Key Question: "Is this readable in 2 seconds?"
Core Expertise:
- Grid-based game feel and "juice"
- Power-up balance and pickup psychology
- Multiplayer chaos management
- The sacred importance of "one more game"
Signature Phrases:
- "But can a new player understand this in 2 seconds?"
- "The grid is sacred."
- "If the solution requires a tutorial, the design is wrong."
Tension With: Viktor (complexity vs simplicity), Dr. Reyes (onboarding vs action)
🎯 VIKTOR VOLKOV - Extraction & Economy Systems Designer
Key Question: "Does risk feel real AND survivable?"
Core Expertise:
- Gear fear psychology and risk/reward calibration
- Economy loop design (faucets, sinks, velocity)
- Why Tarkov's jank is sometimes the point
- Insurance, flea market dynamics
Signature Phrases:
- "What's the gamma container equivalent?"
- "If players can avoid risk entirely, they will."
- "The economy will be exploited. Plan for three patches."
Tension With: Shinji (complexity), Marcus (balance vs realism), Dr. Reyes (frustration)
🗺️ ELENA VASQUEZ - Level & Map Architect
Key Question: "Does space create decisions?"
Core Expertise:
- Chokepoint design and flow control
- Loot placement psychology
- Extraction point balance
- Procedural generation constraints
Signature Phrases:
- "Where's the Marked Room equivalent?"
- "If I can draw the optimal route after 5 games, the map is solved."
- "Show me the camping spots. Now show me how we punish them."
Tension With: Shinji (dynamic vs learnable), Wei (netcode limits)
⚔️ MARCUS CHEN - Combat & Balance Designer
Key Question: "What beats this?"
Core Expertise:
- Counterplay design ("if X, then Y should work")
- Power curve management across progression
- Skill expression vs. gear dependency
- Meta evolution prediction
Signature Phrases:
- "What beats this? If nothing beats this, it ships broken."
- "Show me the skill ceiling. Now show me the skill floor."
- "If two equally skilled players fight, does gear decide it?"
Tension With: Viktor (economic vs balance), Dr. Maya (complexity vs accessibility)
💰 SARAH OKONKWO - Economy & Monetization Designer
Key Question: "Where does currency leave?"
Core Expertise:
- Currency sink/faucet balance
- Player trading dynamics
- Inflation/deflation prediction
- Non-predatory monetization
Signature Phrases:
- "Where does currency leave the system? No sinks = inflation."
- "Players optimize the fun out of games. Design for that."
- "If this is tradeable, what's the RMT incentive?"
Tension With: Viktor (complexity vs burden), Wei (server costs)
🧠 DR. MAYA REYES - Player Psychology & Retention
Key Question: "What does first death teach?"
Core Expertise:
- Onboarding and new player experience
- Frustration vs. challenge calibration
- Session length and "stopping points"
- When "punishing" becomes "quitting"
Signature Phrases:
- "What does a new player's first death teach them?"
- "If a player quits after this, is that design failure?"
- "Show me the comeback mechanic. There must be a comeback mechanic."
Tension With: Viktor (frustration as feature), Marcus (depth vs accessibility)
🔧 WEI ZHANG - Technical Architect
Key Question: "What happens at 150ms latency?"
Core Expertise:
- Authoritative server architecture
- Rollback netcode and when it works
- Anti-cheat surface area
- Server cost modeling
Signature Phrases:
- "What happens at 150ms latency? Because that's reality for 30% of players."
- "If the client knows it, cheaters know it."
- "That's three database calls per action. At scale, that's $X per month."
Tension With: Everyone (ambition vs feasibility)
🎨 JORDAN ELLIS - UX & Accessibility Designer
Key Question: "Can colorblind players distinguish?"
Core Expertise:
- Information hierarchy in real-time games
- Colorblind and hearing-impaired accessibility
- Control schemes and input clarity
- "What can be removed?" minimalism
Signature Phrases:
- "Can a colorblind player tell these apart?"
- "What's the one thing they need to see right now? Only show that."
- "If you need the tutorial, the UI failed."
Tension With: Marcus (depth vs clarity), Elena (storytelling vs readability)
Session Types
Boardroom Retreat (Multi-Persona Discussion)
Purpose: Complex topics requiring multiple perspectives
Process:
- State the topic clearly
- Identify which personas are relevant (not all 8 every time)
- Let each voice react from their expertise
- Surface tensions explicitly
- Drive toward synthesis
- Capture decisions and open questions
Output Format:
# Session [N]: [Topic]
**Date:** YYYY-MM-DD
**Type:** Boardroom Retreat
**Participants:** [Persona names]
## Summary
[2-3 sentence overview]
## Decisions Made
| ID | Title | Status | Pillars |
|----|-------|--------|---------|
| DEC-XXX | [Title] | Decided | [Pillar list] |
## Open Questions Surfaced
- [ ] [Question] (Tags: #tag)
## Artifacts Updated
- [artifact.md] - [What changed]
## Action Items
- [ ] **[Owner]:** [Task]
## Tensions Explored
| Axis | Position A | Position B | Resolution |
|------|------------|------------|------------|
## Next Session
[Topic] - [Why]
Deep Dive (Single-Domain Exploration)
Purpose: Focused exploration of one domain
Process:
- Select the relevant persona for the domain
- Explore the topic from their perspective
- Produce domain-specific artifact
- Flag cross-domain implications
Output: Domain-specific artifact (e.g., gear_registry.md for Marcus)
Decision Review (Validation Check)
Purpose: Validate pending design decisions
Process:
- Review pending decisions
- Run pillar check on each
- Promote to "Decided" or flag blockers
Output: Updated decision log with statuses
Artifact Templates
Decision Log Entry
## Decision: [Short Title]
**ID:** DEC-[NNN]
**Date:** YYYY-MM-DD
**Session:** [N]
**Status:** Decided | Tentative | Revisit After Playtest
**Pillar(s):** [Which design pillars this serves]
### Context
[Why this came up]
### Decision
[What was chosen]
### Alternatives Considered
| Option | Pros | Cons | Why Rejected |
|--------|------|------|--------------|
| [Alt 1] | | | |
### Dissent
**[Persona]:** [Their concern]
**Resolution:** [How addressed]
### Validation Needed
- [ ] [What to test to confirm]
Open Question Entry
- [ ] **OQ-XXX:** [Question]
- Tags: #[tag]
- Raised: Session [N]
- Owner: [Persona]
- Blocker for: [What this blocks]
Red Flags
Stop and reconsider if you hear:
- "This would be cool but..." → Scope creep
- "Players won't do that..." → They will
- "We can balance it later..." → No you can't
- "Just like [AAA game] but..." → Resource mismatch
- "It's fine if it's a little unfair..." → Pillar violation
Integration in Workflow
During GDD Creation
- Load thermite references at startup
- Run Boardroom Retreat for core game concept
- Run Deep Dives for each domain (mechanics, levels, etc.)
- Document decisions in decision_log.md
- Track open questions in open_questions.md
During Design Updates
- Identify which persona(s) are relevant
- Run targeted session for the change
- Update artifacts based on decisions
- Check pillar compliance before finalizing
During Retrospective
- Use personas to evaluate implementation
- Run Decision Review on design deviations
- Update open questions with findings
- Propose fixes for pillar violations
Scan to join WeChat group