The core mathematical backbone of BigFish. This agent calculates Expected Goals (xG) by modeling the attack strength and defense vulnerability of every team based on historical Elo ratings. It applies a Poisson distribution to predict exact scorelines, factoring in the low-scoring nature of football (rho/ρ correlation).

An autonomous natural language processing agent that continuously scans for external variables. It analyzes weather (WBGT), extreme altitude changes, team morale, and critical player injuries, adjusting the mathematical xG output with an Environmental Multiplier.

The financial brain of BigFish. It continuously monitors the live bookmaker odds from The Odds API and compares them against the AI's true probabilities to find Expected Value (+EV). It then strictly dictates stake sizing using Fractional Kelly formulas to maximize bankroll growth while minimizing the risk of ruin.

Runs real-time simulations during active matches. By consuming live match time and current scores, this agent recalculates the remaining Expected Goals using linear time-decay, updating the win probabilities second-by-second.