Mechanics of Agency: Quantum Decisions

Summary

Illustrates agency principles using Matching Pennies game. Two agents simultaneously choose heads/tails: Matcher wins if both match, Mismatcher wins if different.

Agency Elements:

• Embeddedness: Agents continuously interact, dynamically adjusting strategies
• Predictive Modeling: Forecast opponent choices using internal representations
• Intentional Biasing: Deliberately choose moves to skew outcomes toward preferred states

Building Predictive Models:

1. Observation/data collection (recording historical choices)
2. Pattern recognition (identifying statistical trends/biases)
3. Probabilistic forecasting (creating probability distributions)
4. Counterfactual simulation (evaluating potential outcomes)
5. Decision-making/adjustment (maximizing expected outcomes, refining model)

Key Insights:

• Optimal play involves mixed strategies (maintaining unpredictability)
• Predictive accuracy correlates with strategic advantage
• Against true randomness: modeling provides no advantage—optimal strategy = pure randomization
• Against environment: behavior follows structured probabilistic rules, modeling becomes feasible

QBU Connection: Each decision = branching point generating multiple futures. Agency involves not only predicting but actively shaping probabilities across possible timelines.

Tags

• agency
• quantum
• decision-theory
• models
• game-theory

Cross-References

• Related: The Physics of Agency, Part 4: The Law of Control Work — Agency Costs Energy
• Related: The Quantum Branching Universe (QBU)
• Related: The Mechanics of Agency

Notes

• Concrete example grounding abstract agency theory
• Matching Pennies becomes recurring illustration
• Limits of agency: randomness defeats prediction
• Foreshadows AI alignment work (predictive modeling in strategic environments)