Unit Composition, Resource Modeling, and Information Asymmetry

[YTFL]

1. Context and Objective

The current system represents each unit as a single entity with abstract health (HP). This limits realism, as it does not capture unit composition, resource constraints, or partial knowledge of enemy forces.

This upgrade introduces:

• Detailed unit composition and inventory tracking
• Replacement of HP with troop count and operational capability
• Fog-of-war intelligence model for enemy units

The goal is to make the simulation state-rich, resource-constrained, and information-limited.

---

2. Transition from HP to Troop-Based Modeling

Changes

• Replace:

  • HP = 100

• With:

  • Exact troop count
  • Unit composition breakdown

Example Structure

Unit Alpha:
  infantry: 32
  support: 6
  heavy: 2

Impact

• Losses become granular (e.g., losing 5 infantry instead of -10 HP)
• Unit effectiveness degrades non-linearly
• Enables modeling of morale, suppression, and combat capability later

---

3. Unit Composition and Equipment Modeling

Each unit will now include detailed equipment and capabilities.

Components

• Weapon systems:

  • Rifles, machine guns, anti-armor, etc.

• Ammunition:

  • Per weapon type

• Support equipment:

  • Batteries, communication devices

Example

weapons:
  rifles: 30
  machine_guns: 3

ammo:
  rifle_rounds: 1200
  mg_rounds: 600

Impact

• Combat effectiveness depends on available equipment
• Units can become combat-ineffective without being destroyed

---

4. Resource and Logistics Modeling

Introduce resource tracking at the unit level.

Tracked Resources

• Ammunition
• Fuel (for vehicles/mechanized units)
• Food (affects endurance)
• Batteries (affects sensors/communication)

Behavioral Effects

• Low ammo → reduced engagement capability
• Low fuel → restricted movement
• Low food → long-term degradation
• Low battery → reduced detection/communication

Impact

• Introduces logistical constraints
• Enables future mechanics like resupply and attrition

---

5. Combat Effectiveness as a Function of State

Replace fixed damage output with state-dependent effectiveness.

Factors

• Remaining troops
• Available ammunition
• Equipment status
• Environmental conditions

Example

effectiveness = f(troop_count, ammo, terrain, morale)

Impact

• Combat outcomes become dynamic and context-dependent
• Prevents unrealistic “full strength until 0 HP” behavior

---

Good call merging these—these three sections are really one concept: how truth vs perceived state is handled across both sides. I’ll consolidate them cleanly and incorporate your correction about dual AI roles (ally-side vs enemy-side).

---

6. Information Model, Fog-of-War, and State Separation

The simulation introduces a unified information architecture that separates true battlefield state from what each actor (user, allied AI, enemy AI) can observe. This ensures all decision-making occurs under incomplete and uncertain information.

---

6.1 Ground Truth vs Observed State

The system maintains two distinct layers:

1. Simulation State (Ground Truth — Hidden)

• Exact troop counts
• Full unit composition
• Complete resource levels (ammo, fuel, food, batteries)
• True positions and capabilities

This state is only accessible to the simulation engine and is never directly exposed.

---

2. Observed State (Perceived — Visible)

Each actor (user, allied AI, enemy AI) receives a filtered and imperfect view of the battlefield:

• Estimated troop counts (ranges, not exact values)
• Partial or inferred equipment data
• Confidence levels tied to detection quality
• Delayed or outdated information

Example

Enemy Unit (Observed):
  estimated_infantry: 20–30
  possible_weapons: [rifles, machine guns]
  confidence: medium

---

6.2 Symmetric Fog-of-War

Information constraints apply equally across all actors:

• The user does not know exact enemy data
• The enemy AI does not know exact player/allied data
• The allied AI does not have global knowledge beyond what is shared

No actor has privileged access to ground truth outside the simulation engine.

---

6.3 Dual AI Structure

The system introduces two distinct AI roles operating under the same information constraints:

1. Allied AI (Execution Layer)

• Controls allied units during continuous simulation

• Operates using only:

  • Observed battlefield data
  • Commands issued by the user

• Does not have access to full enemy state

• Acts as a decentralized executor of player intent

---

2. Enemy AI (Command Layer)

• Acts as the opposing commander

• Continuously evaluates battlefield conditions

• Receives only:

  • Estimated player/allied unit data
  • Observed intelligence (same limitations as the user)

• Plans and executes strategies to counter the player

---

6.4 Information Flow and Updates

Observed information is updated dynamically based on:

• Detection systems
• Proximity to enemy units
• Duration of observation
• Recon capabilities

This results in:

• Increasing accuracy over time with sustained observation
• Degradation or staleness when contact is lost

---

6.5 Impact on Simulation Behavior

This unified model introduces:

• Decision-making under uncertainty
• Symmetric information constraints across all actors
• Realistic misjudgments and incomplete situational awareness
• Support for deception, hidden movement, and ambush scenarios
• Elimination of artificial advantages for either side

---

This is now a clean, single source of truth section for your information model.

If you want next, I’d strongly recommend:
👉 defining a data structure for ObservedState vs TrueState, because this is where most implementations get messy fast.

---

7. Detection-Driven Information Updates

Enemy information improves based on:

• Proximity
• Recon units
• Duration of observation

Behavior

• Far distance → vague estimates
• Close observation → higher accuracy
• Continuous tracking → refined data

Impact

• Encourages reconnaissance and positioning
• Creates evolving battlefield awareness

---

8. Event-Based Resource Consumption

Resources are consumed during simulation ticks.

Examples

• Firing → ammo decreases
• Movement → fuel decreases
• Time → food consumption
• Sensor usage → battery drain

Impact

• Simulation becomes continuous and state-evolving
• Enables long-duration scenarios with attrition effects

---

9. Direction for Future Extensions

This system enables future upgrades such as:

• Supply lines and logistics networks
• Unit morale and fatigue
• Equipment degradation and maintenance
• Intelligence systems (drones, surveillance)

---

10. Communication and Command Propagation

In the current system, commands are executed instantly once issued. This does not reflect real-world operational constraints where communication introduces delays and uncertainty.

Changes

• Introduce a command transmission pipeline:

  • Command issued → transmitted → received → executed

• Each unit maintains:

  • Communication status
  • Last received command
  • Command queue

Command Flow

Command Issued → Transmission Delay → Unit Receives → Execution Begins

Impact

• Commands are no longer instantaneous
• Units may act on outdated instructions
• Enables realistic coordination challenges

---

11. Communication Delay Modeling

Command delays will depend on multiple factors.

Factors

• Distance between command source and unit
• Terrain interference (mountains, urban areas)
• Environmental conditions (weather, visibility)

Example Model

delay = base_latency + distance_factor + terrain_penalty

Behavior

• Nearby units receive commands faster
• Units in difficult terrain experience delays
• Communication is not guaranteed to be uniform across all units

Impact

• Timing becomes a critical factor in operations
• Delayed execution can alter outcomes significantly

---

12. Independent and Secure Communication Channels

Each unit operates on an independent communication channel.

Properties

• Commands are sent individually to each unit
• No instant global synchronization
• Channels are encrypted (no shared state leakage)

Future Possibilities

• Signal degradation
• Communication loss
• Electronic warfare (optional later phase)

Impact

• Removes unrealistic centralized control
• Reinforces decentralized execution behavior

---

13. Continuous Simulation Execution

The system will move away from turn-based execution toward a continuous simulation model.

Changes

• Simulation runs continuously once initiated
• No waiting for user or AI turns
• Commands can be issued at any time

Behavior

• Units continue executing current orders until new commands arrive
• Multiple actions occur simultaneously across the battlefield

Impact

• Enables real-time dynamics
• Allows overlapping operations and interactions
• Forms the basis for emergent behavior

---

14. Continuous Decision-Making for AI and User

Changes

• AI no longer waits for user input
• AI evaluates the simulation continuously and issues commands dynamically

User Interaction

• User can issue commands at any time
• Commands enter the same delayed pipeline as AI commands

Impact

• Both sides act concurrently
• Removes sequential dependency between actions
• Enables timing-based strategies

---

15. Emergent Tactical Behavior Enablement

With:

• Resource-based units
• Fog-of-war
• Communication delays
• Continuous simulation

The system enables emergent scenarios:

Examples

• Ambushes due to delayed detection
• Units moving into enemy range unknowingly
• Misaligned attacks due to communication lag
• Coordinated strikes arriving at different times

Impact

• Outcomes are no longer fully predictable
• Strategy depends on timing, positioning, and information

---

16. Integration with Resource and Detection Systems

This upgrade integrates directly with earlier Phase 2 components:

• Resource Model

  • Communication devices depend on battery levels

• Detection System

  • Information updates depend on observation quality

• Unit Composition

  • Different units may have different communication capabilities

Impact

• Communication, detection, and logistics become interconnected
• Creates a unified simulation model rather than isolated systems