One-Year Monte-Carlo Simulation

Observations

  • Validator Bottleneck (around Day 90): A rapid increase in miners paired with a slower rise in validators caused AI response quality, Q, to drop by approximately 12%, while the reward pool, R, remained flat.

  • Validator Outage (around Day 180): A ten-day decline in validator activity led to an 18% reduction in rewards, even though miner numbers stayed steady.

  • Self-Healing Phase: Validator participation recovered, pushing AI response quality Q back above 90%, and the reward pool R resumed its upward trajectory.

Design Implications

  • Rewards are influenced equally by validator capacity and mining power, underscoring the importance of balanced network growth.

  • Short-term fluctuations in rewards of around ±20% are normal and expected due to dynamic participation levels.

  • The Exion DAO can adjust validator incentives, entry requirements (admission stake), or tuning parameters (α/β constants) if real-time metrics consistently fall outside predefined thresholds.

PreviousQuality of AI Responses

Last updated