trader-memory-core

---
name: trader-memory-core
description: Use when tracking an investment thesis lifecycle — register from a screener, transition states, size a position, check review due dates, close a position, or write a postmortem.
category: Market & Macro
version: 1.0.0
tools: []
---

# Trader Memory Core

Track an investment thesis across its full lifecycle — from screening idea to closed position with postmortem — as a structured "thesis object" that captures what you thought, what happened, and what you learned across the conversation.

Phase 1 scope: single-ticker theses of five types — dividend_income, growth_momentum, mean_reversion, earnings_drift, pivot_breakout. (Pair trades and options are a future phase.)

## When to Use
- After a screening result identifies candidate tickers, to register a thesis for a candidate
- Transitioning a thesis from IDEA → ENTRY_READY → ACTIVE → CLOSED
- Attaching position-sizing details to a thesis
- Checking which theses are due for review
- Closing a position and producing a postmortem with lessons learned
- Trigger phrases: "register thesis", "track this idea", "thesis status", "review due", "close position", "postmortem", "trading journal"

## Thesis Object Fields
Maintain each thesis as a structured record with:
- **Identity**: thesis ID, ticker, created date
- **Classification**: thesis type (one of the five above), setup type, catalyst
- **Lifecycle**: current status, full status history (with dates/reasons)
- **Entry/Exit**: target entry price, actual entry price/date, stop-loss, take-profit, exit conditions
- **Position**: shares (fractional supported), position value, risk sizing
- **Monitoring**: next review date, triggers, alerts
- **Origin**: source (screener name or "manual"), screening grade, raw provenance data
- **Outcome** (once closed): realized P&L ($ and %), holding days, MAE/MFE, lessons learned

## Lifecycle Rules (Forward-Only, No Backtracking)
1. **IDEA** — thesis registered from a screener result or manual entry (fractional-share brokers, hand journaling). Required fields: ticker, thesis statement, thesis type; optional stop/target prices.
2. **IDEA → ENTRY_READY** — thesis validated, ready to size and enter, with a reason recorded.
3. **ENTRY_READY → ACTIVE** — position opened: record actual entry price, actual entry date, and share count (fractional shares are fine). This is the only path to ACTIVE.
4. **ACTIVE/PARTIALLY_CLOSED → PARTIALLY_CLOSED** (trim) — partial close: record shares sold, price, date. `shares` is the original opened quantity and stays fixed; `shares_remaining` tracks what's still open. Cumulative realized P&L = sum of all trims + final close; P&L% = cumulative P&L ÷ (entry price × original shares) × 100. Selling the entire remainder closes the thesis.
5. **ACTIVE or PARTIALLY_CLOSED → CLOSED or INVALIDATED** — final close (with exit reason: target_hit, stop_hit, thesis broke, etc.) or invalidation. Computes final P&L (fractional-share aware).

Status invariants: ACTIVE ⇒ shares_remaining = shares; PARTIALLY_CLOSED ⇒ 0 < shares_remaining < shares; CLOSED ⇒ shares_remaining = 0.

For backfilling an already-open broker position (e.g. an existing fractional-share position from IBKR, Robinhood, or similar), walk the same three steps (ingest → ENTRY_READY → ACTIVE) with backdated event dates so the history stays chronological.

## Review Workflow
Track a next-review-date per thesis; when asked "what's due for review," list every thesis whose next-review-date has passed, and record each review's outcome (OK / WARN / REVIEW).

## Postmortem
On close, produce a postmortem: what the original thesis predicted, what actually happened, realized P&L, holding period, and — where price-history data is available — Maximum Adverse Excursion and Maximum Favorable Excursion (how far the position moved against/in favor of you intraperiod, a measure of whether stops/targets were well-placed). Close with explicit lessons learned. Track win rate, average P&L%, and a breakdown by thesis type across all closed theses for a running summary.

## How to Operate This Without a Filesystem
Since state can't be persisted to files here, maintain the thesis as a structured table/summary directly in the conversation — restate the full thesis object (or the relevant slice) each time it's updated, and return updated postmortems and summaries in the chat for the user to copy into their own tracking system. Ask the user to paste back the last-known thesis state at the start of a session if continuity is needed across sessions.

## Key Principles
- **Forward-only transitions** — no backtracking through the lifecycle
- **Raw provenance preserved** — keep the original screener/manual data attached to the thesis
- **One source of truth per thesis** — track a single authoritative status and status history
- **Fractional shares supported throughout** — entries, trims, and closes all accept fractional quantities

## Ingesting From a Screener
When a screener (dividend, growth-momentum, mean-reversion, earnings-drift, or pivot-breakout style) produces candidate tickers, convert each candidate the user wants to track into an IDEA-stage thesis: carry over the ticker, the screener's stated rationale as the thesis statement, the appropriate thesis type, and any suggested stop/target levels as the starting exit plan. Preserve the screener's original grade/score in the origin field so it can be referenced later when judging whether the thesis played out as the screener predicted — this comparison is one of the most valuable postmortem outputs, since it tells the user whether that screener's signals are actually predictive over time.

## Linking Related Analysis
When later analysis (an earnings review, a technical update, a position-sizing calculation) is produced for a ticker with an open thesis, cross-reference it against that thesis by name and date rather than treating it as a standalone note — this keeps the full decision trail attached to the trade it informed, which is what makes the postmortem meaningful instead of just a P&L number.