The Entropy Engine: An Onboarding Guide for NPC Engineers

Fellow Traveler
Aug 15
3 min read

I. Your First Day on the Team: The Entropy Engine's Mission

Welcome to the Entropy Engine team! You've joined a unique project, and your role is one of the most vital. In this world, we don't just create deterministic characters who follow a rigid script. Instead, we build a living ecosystem. Your job is to be an architect of that ecosystem—to build the very minds of the NPCs that inhabit our simulation.

The Entropy Engine's mission is simple: to keep our world in a state of healthy balance.

This balance isn't static; it's a dynamic dance between order and chaos, which we call entropy. As an NPC engineer, you are on the front lines of this mission. Your NPC isn’t a mindless automaton—it’s an intelligent part of a team, helping the entire system stay in harmony.

II. The Core Science of Your Role

Your work isn't just about code; it's about applying real science to create a believable world. The Entropy Engine runs on measurable math:

What the EE Measures – The system tracks the rate of entropy change (ΔH/Δt) across the whole simulation.
Your NPC’s Impact – Every NPC action affects Telemetry Variables. If you build a “woodchopper,” its actions influence Resource Levels (wood). If you build a “healer,” it influences Health Metrics.
Why This Matters – Your NPC’s data feeds into the entropy calculation. That number drives the EE’s nudges back to other agents, including your NPC.

III. The Minimal Entropy Math Walkthrough

Here’s how the EE processes what your NPC does into the math that drives the system:

Telemetry Sampling The EE collects your NPC’s relevant variable(s). Example: wood_supply = 80 units.
Probability Distribution The EE estimates the probability of each state. Example:
- P(wood low) = 0.25
- P(wood medium) = 0.50
- P(wood high) = 0.25
Shannon Entropy Calculation $$H = -\sum_i p_i \log_2 p_i$$ Plugging in the example: $$H = -(0.25 \log_2 0.25 + 0.50 \log_2 0.50 + 0.25 \log_2 0.25)$$ $$H=1.5 bits$$
Rate of Change $(ΔH/Δt)$ The EE compares this entropy to the last known value, calculating the trend. If entropy rose from 1.2 → 1.5 bits in the last minute, $ΔH/Δt = +0.3 bits/min$.
Nudge Decision The node sends your NPC an EeFrame with:
- Mode bias: “reduce chaos”
- Confidence: 0.88
- Suggested action: “gather wood”

IV. Diagram: How Your NPC Fits into the Math

   ┌─────────────────────┐
   │  NPC Action Taken   │
   │ (e.g., chop wood)   │
   └─────────┬───────────┘
             │
     [Telemetry Update]
             │
   ┌─────────▼───────────┐
   │ EE Node Collects    │
   │ & Updates Variable  │
   │ wood_supply = 80    │
   └─────────┬───────────┘
             │
   [Probability Mapping]
             │
   ┌─────────▼───────────┐
   │ Shannon Entropy H   │
   │ Calculation         │
   └─────────┬───────────┘
             │
   [ΔH/Δt Computed]
             │
   ┌─────────▼───────────┐
   │ EeFrame Nudge Sent  │
   │ (confidence, bias)  │
   └─────────────────────┘

V. Your Voice in the Network: The Feedback Loop

Your NPC doesn’t just receive nudges—it also sends feedback (EeAck) after acting. This closes the loop:

Local Impact – The node updates its local model for your NPC.
Global Impact – The EE updates system-wide entropy trends.

VI. First Mission: Your NPC in the Loop

Sense: Fire in the forest → entropy spike.
Nudge: EeFrame to your NPC: “reduce chaos” (confidence 0.9).
Process: NPC logic → “gather water” instead of “chop wood.”
Action: NPC puts out fire.
Feedback: EeAck updates node → entropy recalculated.

VII. Conclusion

You’re not just coding an NPC—you’re wiring a neuron in a brain. Every decision your NPC makes affects the math that keeps the entire system in balance. Master this connection, and you’ll be building not just characters, but a living, self-regulating world.