journal

Stage 08 | 复盘学习

in execution log + original signals -> out trade attribution + monthly retrospective report

Decision journal with full audit trail. Every trade traced back to the signal that triggered it.

Pipeline Position

01 kline -> 02 intel -> 03 signal -> 04 copilot -> 05 backtest
                                  -> 06 risk -> 07 executor -> [08 journal]
                                                                    |
                                                        feedback -> 03 signal (weights)
                                                        feedback -> 04 copilot (methodology scores)

Capabilities

• Record every decision: action, instrument, conviction, sizing, reasoning
• Full context snapshot: market data, intelligence, portfolio state at decision time
• Trade lifecycle: open -> close with exit price, PnL (USD + %), outcome notes
• Attribution: which signals and methodologies actually made money?
• Monthly retrospective report (machine-readable + human-readable)
• Feedback interface: performance data feeds back to signal scoring (Stage 03) and methodology routing (Stage 04)

Source

Adapted from tradinghouse/src/journal/journal_store.py + journal/schemas.py. Existing code has record_decision() and close_trade() with full PnL tracking.

Tech Stack

Python 3.11+ | SQLAlchemy | SQLite

Status

Scaffolding. Core journal logic exists in tradinghouse. Attribution and retrospective reports are new.