Knowledge Base
AI Features

How Tradapt's AI Pattern Detection Works

A detailed explanation of how Tradapt's AI analyzes your trade history, what patterns it detects, and how to interpret its findings.

5 min read
Last reviewed: Apr 2026

What the AI Analyzes


Tradapt's AI analyzes your complete trade history looking for statistical correlations between your behavior and trade outcomes. It does not use generic market models — it analyzes your specific trading data.


The AI looks for patterns across:

  • Emotional state at entry vs. trade outcome
  • Time elapsed since last trade vs. outcome (revenge trading signal)
  • Consecutive wins/losses vs. subsequent sizing and outcome
  • Setup adherence vs. outcome
  • Time of day and session vs. performance
  • Day of week vs. performance

How Patterns Are Detected


Each analysis compares subgroups of your data:


Example: Revenge Trading Detection

The AI identifies all trades taken within 10 minutes of a losing trade. It calculates the win rate and average R-multiple of those trades vs. your baseline. If there's a statistically significant negative divergence (your "quick-after-loss" trades perform significantly worse), it flags a revenge trading pattern.


Example: Emotional State Correlation

For each emotional state level (1–5), the AI calculates your win rate and average R-multiple. If there's a meaningful performance difference between state 1–2 (calm) and state 4–5 (elevated), it reports the correlation and quantifies the impact.


Statistical threshold: Patterns are only reported when the dataset is sufficient (minimum 15–20 trades in each compared group) and the difference is meaningful (not random noise).


The Seven Patterns Tracked


  1. Revenge Trading: Trades taken quickly after losses
  2. FOMO Entries: Late entries (beyond optimal entry window)
  3. Overtrading: Performance degradation after N trades in a session
  4. Emotional State Impact: Win rate by state level
  5. Time-of-Day Performance: Session and time window analysis
  6. Win/Loss Streak Effects: Performance after consecutive wins or losses
  7. Setup Adherence Gap: Difference between in-playbook and off-playbook performance

Interpreting the Results


AI insights show:

  • The detected pattern name
  • The supporting data (e.g., "Revenge trades have −0.7R average vs. your +0.9R baseline")
  • A specific recommendation based on your data

Treat AI findings as data for further investigation, not absolute verdicts. If the AI flags a revenge trading pattern, review the specific flagged trades to confirm they match the behavioral description.


Improving AI Accuracy


The more data you provide, the more accurate and specific the AI's analysis:


  • Log emotional state on every trade (not just memorable ones)
  • Use consistent setup type names (the AI groups by exact setup name)
  • Add behavioral tags when you notice them in real time
  • Log both winning and losing trades — analyzing only winners creates survivorship bias

After 100+ trades with consistent emotional state data, the AI's pattern confidence increases significantly.


Was this article helpful? If you have more questions, contact our support team or visit the Trading University for in-depth courses.