stanley-druckenmiller-investment

Druckenmiller Strategy Synthesizer

Purpose

Synthesize outputs from 8 upstream analysis skills (5 required + 3 optional) into a single composite conviction score (0-100), classify the market into one of 4 Druckenmiller patterns, and generate actionable allocation recommendations. This is a meta-skill that consumes structured JSON outputs from other skills — it requires no API keys of its own.

When to Use This Skill

English:

• User asks "What's my overall conviction?" or "How should I be positioned?"
• User wants a unified view synthesizing breadth, uptrend, top risk, macro, and FTD signals
• User asks about Druckenmiller-style portfolio positioning
• User requests strategy synthesis after running individual analysis skills
• User asks "Should I increase or decrease exposure?"
• User wants pattern classification (policy pivot, distortion, contrarian, wait)

Japanese:

• 「総合的な市場判断は？」「今のポジショニングは？」
• ブレッドス、アップトレンド、天井リスク、マクロの統合判断
• 「エクスポージャーを増やすべき？減らすべき？」
• 「ドラッケンミラー分析を実行して」
• 個別スキル実行後の戦略統合レポート

---

Input Requirements

Required Skills (5)

Optional Skills (3)

Run the required skills first. The synthesizer reads their JSON output from reports/.

---

Execution Workflow

Phase 1: Verify Prerequisites

Check that the 5 required skill JSON reports exist in reports/ and are recent (< 72 hours). If any are missing, run the corresponding skill first.

Phase 2: Execute Strategy Synthesizer

python3 skills/stanley-druckenmiller-investment/scripts/strategy_synthesizer.py \
  --reports-dir reports/ \
  --output-dir reports/ \
  --max-age 72

The script will:

1. Load and validate all upstream skill JSON reports
2. Extract normalized signals from each skill
3. Calculate 7 component scores (weighted 0-100)
4. Compute composite conviction score
5. Classify into one of 4 Druckenmiller patterns
6. Generate target allocation and position sizing
7. Output JSON and Markdown reports

Phase 3: Present Results

Present the generated Markdown report, highlighting:

• Conviction score and zone
• Detected pattern and match strength
• Strongest and weakest components
• Target allocation (equity/bonds/alternatives/cash)
• Position sizing parameters
• Relevant Druckenmiller principle

Phase 4: Provide Druckenmiller Context

Load appropriate reference documents to provide philosophical context:

• High conviction: Emphasize concentration and "fat pitch" principles
• Low conviction: Emphasize capital preservation and patience
• Pattern-specific: Apply relevant case study from references/case-studies.md

---

7-Component Scoring System

4 Pattern Classifications

Conviction Zone Mapping

---

Output Files

• druckenmiller_strategy_YYYY-MM-DD_HHMMSS.json — Structured analysis data
• druckenmiller_strategy_YYYY-MM-DD_HHMMSS.md — Human-readable report

API Requirements

None. This skill reads JSON outputs from other skills. No API keys required.

Reference Documents

references/investment-philosophy.md

• Core Druckenmiller principles: concentration, capital preservation, 18-month horizon
• Quantitative rules: daily vol targets, max position sizing
• Load when providing philosophical context for conviction assessment

references/market-analysis-guide.md

• Signal-to-action mapping framework
• Macro regime interpretation for allocation decisions
• Load when explaining component scores or allocation rationale

references/case-studies.md

• Historical examples: 1992 GBP, 2000 tech bubble, 2008 crisis
• Pattern classification examples with actual market conditions
• Load when user asks about historical parallels

references/conviction_matrix.md

• Quantitative signal-to-action mapping tables
• Market Top Zone x Macro Regime matrix
• Load when user needs precise exposure numbers for specific signal combinations

When to Load References

• First use: Load investment-philosophy.md for framework understanding
• Allocation questions: Load market-analysis-guide.md + conviction_matrix.md
• Historical context: Load case-studies.md
• Regular execution: References not needed — script handles scoring

---

Relationship to Other Skills