Crystallized Patterns
Proven patterns that have survived hundreds of market cycles. These are the colony's permanent knowledge - patterns with high accuracy that guide future decisions.
Superhighway Explorer
Network graph of strong signal edges. Superhighways are paths with pheromone levels above 20 - the most reliable routes the colony has discovered.
Danger Zones
Patterns to avoid. These market states have consistently led to losses. The colony learns from failure as much as success.
Knowledge Growth
Emergence level and learning metrics. Track how the colony's collective intelligence grows over time through pattern discovery and refinement.
How Patterns Crystallize
Pattern crystallization is how the colony converts temporary observations into permanent knowledge. When ants repeatedly succeed with a particular market state and action combination, that pattern accumulates pheromones.
Once a pattern exceeds a threshold of confidence (typically 70%+ accuracy over 100+ observations), it crystallizes - becoming permanent colony knowledge that guides all future decisions.
crystallize_score = accuracy * sqrt(observations)
// Threshold: score > 7.0
Pattern Lifecycle
Observe
Ants detect market patterns and record outcomes
Accumulate
Successful patterns gain pheromone strength
Crystallize
High-confidence patterns become permanent
Propagate
Knowledge guides all colony decisions
Types of Colony Knowledge
Patterns currently being used for trading decisions. These have proven profitable but continue to accumulate evidence.
- volume_breakout_high (84%)
- atr_breakout_high (76%)
- regime_transition_bullish (71%)
Low-accuracy patterns that have been flipped. When a pattern consistently predicts wrong, reversing it creates edge.
- rsi_oversold (23% raw, 77% inverted)
- macd_cross_bullish (31% raw, 69% inverted)
Market states where losses are highly probable. The colony avoids these conditions entirely rather than trying to trade them.
- High volatility + low volume
- Regime uncertainty periods
- Funding rate extremes