Research Date: December 2024 Purpose: Foundation research for Resonance Wars poker-like bluffing tactical combat mode
1. Poker Game Theory and Nash Equilibrium#
GTO (Game Theory Optimal) Fundamentals#
Definition: GTO poker is "perfect poker" - a strategy that is unexploitable regardless of opponent strategy.
Nash Equilibrium in Games:
- When all players play GTO, no one can profit - equilibrium occurs
- In two-player zero-sum games, Nash equilibrium is where each player's choice optimally responds to the other
- True GTO is computationally impossible for complex games (10^161 states in poker)
Mixed Strategies#
Definition: When a player uses multiple actions with specific probabilities rather than always choosing one action.
Key Concepts:
- Pure Strategy: 100% probability of one action
- Mixed Strategy: Multiple actions with calculated probabilities
- Indifference: Properly balanced ranges make opponent indifferent between responses
Example (Poker River Bet):
- $200 pot, $200 bet = opponent needs 33% equity to call
- GTO: Value bet 66%, bluff 33%
- This makes opponent indifferent between calling and folding
GTO vs Exploitative Play#
GTO Approach:
- Minimizes losses against optimal opponents
- Provides baseline unexploitable strategy
- Not necessarily profit-maximizing against weak opponents
Exploitative Approach:
- Adjusts strategy based on opponent tendencies
- Higher EV against predictable opponents
- Vulnerable to counter-exploitation
Application to Resonance Wars#
Intent Balance:
- Each intent type should have counters
- Optimal mixed strategies for intent selection
- Balance such that pure strategies are exploitable
- Create indifference points where opponent can't predict
2. Simultaneous Action Selection in Fighting Games#
Frame Data and Timing#
What is Frame Data:
- Each move takes X frames (at 60fps, 1 frame = 1/60 second)
- Frame advantage: Difference between your and opponent's move duration
- Hitstun: Forced inaction after being hit
- Recovery: Vulnerable frames after move ends
Why Frames Matter:
- Determines who can act first after exchange
- Creates natural RPS interactions (fast beats slow startup)
- Punish windows for blocked/whiffed moves
Yomi (Reading the Opponent)#
Definition: Japanese for "reading" - knowing the mind of the opponent.
Yomi Layers:
- Layer 0: Discover best move, do it repeatedly
- Layer 1: Opponent expects your move, counters it
- Layer 2: You counter their counter
- Layer 3: They counter your counter to their counter
- Layer 4+: Loops back to Layer 0
Design Implication:
- All moves need counters
- Support up to Yomi Layer 3
- Layer 4 naturally wraps around
- Creates infinite depth without infinite complexity
Double-Blind Nature#
Key Insight: Fighting games are effectively double-blind due to reaction time limits.
- Human perception has inherent delay
- At moment of action, opponent's choice unknown
- Acting on outdated information
- Creates simultaneous action feel in real-time
Your Only Move Is HUSTLE (YOMI Hustle)#
Turn-Based Fighting Game Innovation:
- Game pauses every few frames for input
- All frame data visible during pause
- Removes execution requirements
- Pure strategic decision-making
- "Choreograph" fight sequences
3. Bluffing Mechanics in Board Games#
Coup#
Core Mechanic: Role claiming and challenging.
How It Works:
- Players claim to have certain role cards
- Other players can challenge claims
- Failed challenges = lose influence
- Successful challenges = claimer loses influence
Design Principles:
- High-stakes all-or-nothing bluffs
- Claims create meta-information
- Trust/suspicion drives social dynamics
- Resource management (influences) limits bluff frequency
Sheriff of Nottingham#
Core Mechanic: Declaration and inspection.
How It Works:
- Players fill bags with goods (legal and/or contraband)
- Declare contents (must be truthful about COUNT, can lie about TYPE)
- Sheriff decides whether to inspect
- Bribery and negotiation allowed
Design Principles:
- Forced confrontation (hand bag to Sheriff)
- Must look opponent in eye when declaring
- Negotiation reduces binary risk
- Set collection adds strategic depth beyond bluffing
Key Design Principles#
- Balance Risk and Reward: Stakes make bluffs impactful
- Incorporate Social Dynamics: Multiple interaction channels
- Adjust Complexity Over Time: Systems grow as players learn
- Forced Confrontation: Create moments of direct conflict
- Negotiation Options: Allow deals to modify risk
4. Rock-Paper-Scissors Design Patterns#
Balance Rules for RPS Variants#
Core Rule: No single strategy dominates.
Balance Requirements:
- Odd number of total options recommended
- Each option beats exactly half of others
- Each option loses to exactly half of others
- Even number of choices = even number of draws
- Odd number of choices = odd number of draws
Extended RPS Variants#
RPS-5 (Lizard Spock):
- Most successful extension
- Perfectly balanced
- Dramatically reduces ties
- Pop culture recognition (The Big Bang Theory)
RPS-7, RPS-9, RPS-15, RPS-25, RPS-101:
- Created by David C. Lovelace
- Increasingly complex interaction matrices
- RPS-101: 5050 possible outcomes
Why RPS Works in Games#
Core Benefit: Prevents dominant strategies.
Design Goals:
- Strategic diversity
- Dynamic decision-making
- Cyclical interactions
- No single element dominates
Adding Depth:
- Include "signals" for learning and skill
- Add asymmetric payoffs
- Create context-dependent strength
- Allow player expression within RPS framework
Application to Resonance Wars#
7-Intent System (RPS-like):
- Strike > Flank, Suppress
- Defend > Strike
- Flank > Defend
- Channel > (Path-specific)
- Suppress > Channel
- Veil Step > (Movement)
- Ward > Veil Step
Asymmetric Payoffs:
- Different damage/effect magnitudes
- Path-specific modifiers
- Unit type interactions
- Terrain effects on intent resolution
5. Yomi Game Design (David Sirlin)#
Core Philosophy#
Design Goal: Distill high-level mind games from fighting games into card game.
Key Insight: "All the complicated decisions in game theory go away if you know exactly what the opponent will do next."
Dominant Move Problem#
What to Avoid:
- Moves strictly better than all alternatives
- Moves without real counters
- Knowledge of opponent's move providing no advantage
Solution:
- Every move has counters
- Knowing opponent's plan provides counter options
- Support Yomi Layer 3 thinking throughout
Character Asymmetry#
Design Approach:
- Each character has unique abilities
- Different optimal strategies per character
- Matchup-specific considerations
- Encourages learning multiple characters
Payoff Asymmetry#
Why It Matters:
- Different payoffs for win/lose/draw based on choice
- Provides basis for reading opponent
- Context-dependent value (hand size, health, etc.)
- Dynamic equilibria shift during game
Combo System#
Additional Depth:
- Card combos create extended exchanges
- Risk/reward for going for big combos
- Resource management (hand size)
- Multiple paths to victory
6. UI Patterns for Hidden Intent Selection#
Simultaneous Action Selection Mechanic#
Core Flow:
- All players secretly plan actions
- Commitment locked in
- Simultaneous reveal
- Resolution
Benefits:
- Psychological warfare element
- Prevents reactive play
- Mitigates alpha gamer effect
- Creates memorable reveal moments
UI Design Considerations#
Selection Phase:
- Clear indication of hidden state
- Confirmation before lock-in
- Timer for decision (prevents stalling)
- Undo before commitment
Reveal Phase:
- Dramatic reveal animation
- Clear outcome visualization
- Replay/slow-motion option
- Outcome explanation
Examples in Games#
7 Wonders:
- Simultaneous card drafting
- Hidden until play
- Pass hands simultaneously
Race for the Galaxy:
- Simultaneous role selection
- Hidden role choice
- Reveal and resolve together
UI Recommendations for Resonance Wars#
Intent Selection Interface:
- Circular "intent wheel" for each unit
- Drag-to-select with confirmation
- Visual preview of possible outcomes
- Lock indicator when committed
Reveal Animation:
- Sequential unit-by-unit reveal
- Intent token "flip" animation
- Outcome calculation visible
- Chain reaction visualization
Information Display:
- Opponent's previous intents visible
- Pattern tracking hints
- Statistical summaries optional
- "Yomi layer" indicator for advanced players
7. Mixed Strategy Equilibria Calculations#
Basic Formula#
For two-player zero-sum games with three options (RPS-style):
Indifference Condition:
- Mix probabilities so opponent's expected value equals zero for all options
- If opponent can exploit a bias, mix isn't optimal
Example Calculation#
In RPS with equal payoffs:
- Optimal: 1/3 each (Rock, Paper, Scissors)
- Any deviation = exploitable
With asymmetric payoffs:
- Adjust probabilities inversely to payoff magnitude
- Higher payoff options mixed less frequently
- Creates counterintuitive optimal strategies
Game-Specific Adjustments#
Context-Dependent Mixing:
- Unit health affects risk tolerance
- Positional advantage changes value
- Resource state affects option viability
- Opponent tendencies allow exploitation
Application to Resonance Wars#
Equilibrium Factors:
- Intent type matchup results
- Path-specific modifiers
- Unit type interactions
- Current game state (health, position, resources)
Player Skill Expression:
- Read opponent patterns (exploit deviations)
- Hide own patterns (approach GTO)
- Adapt mix to game state
- Risk management based on position
8. Design Recommendations for Resonance Wars#
Intent System Architecture#
7 Core Intents: | Intent | Beats | Loses To | |--------|-------|----------| | Strike | Flank, Suppress | Defend | | Defend | Strike | Flank | | Flank | Defend | Strike | | Channel | (Path-specific) | Suppress | | Suppress | Channel | Strike | | Veil Step | (Movement) | Ward | | Ward | Veil Step | Strike |
Asymmetric Payoffs#
Damage Modifiers:
- Strike vs Flank: 150% damage
- Strike vs Suppress: 120% damage
- Defend vs Strike: 50% damage reduction
- Flank vs Defend: Full damage, ignore defense bonus
- Path abilities: Unique effects per Path
Decoy and Bluffing Layer#
Intent Token Mechanics:
- Each unit selects one intent
- Decoy tokens can be placed (limited resource)
- Reveal phase shows all tokens
- Decoy reveal creates uncertainty
Combat Resolution#
Resolution Order:
- All intents revealed simultaneously
- Ward → Veil Step interactions resolved
- Suppress → Channel interactions resolved
- Strike/Defend/Flank triangle resolved
- Path abilities trigger
- Damage applied
- Chain reactions and synergies
Yomi Layer Support#
Design Checklist:
- Every intent has counters ✓
- Knowing opponent's likely choice helps ✓
- Counter to counter exists ✓
- Layer 4 wraps to Layer 0 ✓
UI Implementation#
Selection Phase (Simultaneous):
- 30-second timer for all players
- Intent wheel per unit
- Decoy placement optional
- Lock-in button with confirmation
Reveal Phase (Sequential Drama):
- 1-second delay between reveals
- Token flip animation
- Outcome calculation overlay
- Replay available
References#
- Sirlin, David - "Yomi Layer 3: Knowing the Mind of the Opponent"
- Sirlin, David - "Spies of the Mind" (Playing to Win)
- Sirlin, David - "Balancing Multiplayer Games, Part 2: Viable Options"
- BoardGameGeek - Simultaneous Action Selection mechanic
- World Rock Paper Scissors Association - Variant rules
- MIT Course Paper - "Exploitability and Game Theory Optimal Play in Poker"
- Rempton Games - "Know Thy Enemy: Designing for Yomi in Games"
- Sheriff of Nottingham - CMON Games
- Coup - Indie Boards & Cards