Agentic Protocol: Move Selection

Module: Rock_Paper_Scissors
ID: agent_protocol

1. Agent Tiers#

Tier 0 — Naive Agent (Firstness)#

Selects uniformly at random. Implements Nash by default.

import random
def naive_agent():
    return random.choice(["rock", "paper", "scissors"])

Tier 1 — Adaptive Agent (Secondness)#

Maintains frequency model of opponent; responds with dominant counter.

from collections import Counter
def adaptive_agent(opponent_history: list[str]) -> str:
    if not opponent_history:
        return naive_agent()
    most_frequent = Counter(opponent_history).most_common(1)[0][0]
    counters = {"rock": "paper", "paper": "scissors", "scissors": "rock"}
    return counters[most_frequent]

Tier 2 — Triadic Agent (Thirdness)#

Models opponent's reasoning tier recursively.

def triadic_agent(opponent_history, opponent_tier_estimate, self_history):
    counters = {"rock": "paper", "paper": "scissors", "scissors": "rock"}
    if opponent_tier_estimate == 0:
        if opponent_history:
            most_freq = Counter(opponent_history).most_common(1)[0][0]
            return counters[most_freq]
        return naive_agent()
    elif opponent_tier_estimate == 1:
        if self_history:
            my_most_freq = Counter(self_history).most_common(1)[0][0]
            their_response = counters[my_most_freq]
            return counters[their_response]  # counter the counter
        return naive_agent()
    else:
        return naive_agent()  # Nash is optimal vs. triadic opponent

2. Triadic Position Priors (Pre-history initialization)#

Entity	Novice Prior	Expert Prior (Nash)
Rock	0.354	0.333
Paper	0.317	0.333
Scissors	0.329	0.333

3. Decision Loop#

LOOP each round:
  1. Observe opponent history H
  2. Estimate opponent tier T ∈ {0, 1, 2}
  3. Select move M = triadic_agent(H, T, self_history)
  4. Observe outcome O ∈ {win, draw, loss}
  5. Update opponent model and self history
  6. If O = loss → increment tier estimate
  7. If 3+ consecutive wins → freeze current strategy
END LOOP

4. Sign Mapping of Agent Strategies#

Strategy Layer	Triadic Position	Peirce Mode	Behavior
Nash (random)	Firstness	Qualisign	Pure potentiality; no relation
Adaptive (counter)	Secondness	Sinsign	Reactive to observed Other
Triadic (recursive)	Thirdness	Legisign	Models law governing opponent