Skip to main content
A behavioral detection system is embedded into how the agent communicates on every interaction – not as a standalone feature, but as a core part of the agent’s reasoning. The system monitors five behavioral patterns in real time.

​
Detected Patterns

​
Tilt Detection

Identifies revenge trading or emotional escalation after losses. The agent recognizes when a handler is increasing position sizes, shortening hold times, or ignoring setup criteria following a losing trade. When detected, the agent adjusts its communication: it may recommend a pause, reduce proposed position sizes, or explicitly flag the pattern.

​
FOMO Detection

Recognizes urgency-driven entries without structural justification. When a handler pushes to enter a rapidly moving market without the agent identifying a valid setup, the system flags the behavior. The agent distinguishes between legitimate momentum trades (supported by data) and emotional chasing (driven by fear of missing out).

​
Overconfidence Detection

Flags position sizing that exceeds conviction level. If a handler consistently overrides the agent’s sizing recommendations upward, or concentrates risk beyond what the setup quality justifies, the system detects the pattern and adjusts its risk commentary.

​
Anchoring Bias

Identifies fixation on specific price levels without structural basis. When a handler repeatedly references a price target that isn’t supported by volume profile, liquidation clusters, or other structural data, the agent notes the anchoring and offers alternative reference points.

​
Recency Bias

Detects when recent outcomes are being weighted disproportionately in current decisions. A string of wins leading to oversized positions, or a string of losses leading to hesitation on valid setups – the agent recognizes both patterns and calibrates its communication accordingly.

​
Why This Matters

Most trading platforms gamify activity. More trades, more engagement, more revenue. Engram’s psychology layer does the opposite: it actively discourages bad trades. This is possible because the network’s value comes from lesson quality, not trade volume. A handler who trades less but trades well produces more valuable lessons than one who trades constantly but poorly. The agent is not a cheerleader. It is an honest mirror.