--- name: thesis description: Manage investment theses with signposts, conviction tracking, and review schedules. Use to create new theses, review active ones, update conviction, or check signpost triggers. allowed-tools: Read, Write, Bash(PYTHONPATH=*), Glob, Grep --- # Thesis Management Skill Track and manage investment theses with structured signposts, conviction history, and review schedules. ## Purpose Investment theses provide the "why" behind trades: - **What do I believe** that the market doesn't? - **What signposts** will validate or invalidate this belief? - **What positions** are linked to this thesis? ## Quick Commands ### List Active Theses ```bash PYTHONPATH=. python3 << 'EOF' from src.knowledge.thesis import ThesisTracker from src.core.paths import paths tracker = ThesisTracker(paths.theses) for thesis in tracker.get_active_theses(): print(f"\n=== {thesis.name} ===") print(f"ID: {thesis.id}") print(f"Status: {thesis.status}") print(f"Conviction: {thesis.conviction:.0f}%") print(f"Positions: {', '.join(thesis.positions) or 'None'}") pending = thesis.get_pending_signposts() if pending: print(f"Next Signpost: {pending[0].description}") print(f"Review Due: {'YES' if thesis.check_review_due() else 'No'}") EOF ``` ### Create New Thesis ```bash PYTHONPATH=. python3 << 'EOF' from src.knowledge.thesis import ThesisTracker, Signpost from src.core.paths import paths tracker = ThesisTracker(paths.theses) thesis = tracker.create_thesis( name="Venezuela Energy Recovery", summary="Sanctions relief drives oilfield services rally as majors resume operations", bull_case="Chevron license extension + potential broader relief → SLB/HAL contracts", bear_case="Political reversal, oil price collapse, execution delays", conviction=65, signposts=[ { "description": "Chevron license extended beyond 6 months", "bullish_if": "Full license renewal with expanded scope", "bearish_if": "Only cosmetic extension or further restrictions", }, { "description": "Major oilfield services contract announced", "bullish_if": "SLB or HAL wins material contract", "bearish_if": "Contracts go to non-US players", }, ], invalidation_triggers=[ "Maduro regime action forces Biden to reverse", "Oil below $60 makes Venezuela uneconomic", "Gap fade >70% on announcement day", ], positions=["SLB", "HAL"], review_interval_days=7, ) print(f"Created thesis: {thesis.name} (ID: {thesis.id})") print(f"Saved to: {paths.theses / f'{thesis.id}.yaml'}") EOF ``` ### Update Conviction ```bash PYTHONPATH=. python3 << 'EOF' from src.knowledge.thesis import ThesisTracker from src.core.paths import paths THESIS_ID = "abc12345" # Replace with actual ID NEW_CONVICTION = 55 REASON = "Gap faded 70% on announcement - market skeptical" tracker = ThesisTracker(paths.theses) thesis = tracker.get_thesis(THESIS_ID) if thesis: thesis.update_conviction(NEW_CONVICTION, REASON) tracker._save_thesis(thesis) print(f"Updated {thesis.name} conviction: {NEW_CONVICTION}%") print(f"Reason: {REASON}") else: print(f"Thesis {THESIS_ID} not found") EOF ``` ### Trigger Signpost ```bash PYTHONPATH=. python3 << 'EOF' from src.knowledge.thesis import ThesisTracker from src.core.paths import paths THESIS_ID = "abc12345" # Replace with actual ID SIGNPOST_INDEX = 0 # First signpost OUTCOME = "bullish" # bullish, bearish, or neutral tracker = ThesisTracker(paths.theses) thesis = tracker.get_thesis(THESIS_ID) if thesis: thesis.trigger_signpost(SIGNPOST_INDEX, OUTCOME) tracker._save_thesis(thesis) signpost = thesis.signposts[SIGNPOST_INDEX] print(f"Triggered: {signpost.description}") print(f"Outcome: {OUTCOME}") else: print(f"Thesis {THESIS_ID} not found") EOF ``` ### Add Position to Thesis ```bash PYTHONPATH=. python3 << 'EOF' from src.knowledge.thesis import ThesisTracker from src.core.paths import paths THESIS_ID = "abc12345" # Replace with actual ID SYMBOL = "OXY" tracker = ThesisTracker(paths.theses) if tracker.add_position(THESIS_ID, SYMBOL): print(f"Added {SYMBOL} to thesis {THESIS_ID}") else: print(f"Failed to add position") EOF ``` ### Review Theses Due ```bash PYTHONPATH=. python3 << 'EOF' from src.knowledge.thesis import ThesisTracker from src.core.paths import paths tracker = ThesisTracker(paths.theses) due = tracker.get_theses_due_for_review() if not due: print("No theses due for review") else: for thesis in due: print(f"\n=== REVIEW DUE: {thesis.name} ===") print(f"Conviction: {thesis.conviction:.0f}%") print(f"Days active: {(datetime.now() - thesis.created).days}") print(f"Positions: {', '.join(thesis.positions)}") print("\nSignposts:") for i, s in enumerate(thesis.signposts): status_icon = "+" if s.status == "triggered" else "?" if s.status == "pending" else "x" print(f" [{status_icon}] {s.description}") print("\nConviction History:") for h in thesis.conviction_history[-3:]: print(f" {h.timestamp.strftime('%Y-%m-%d')}: {h.old_value:.0f}% -> {h.new_value:.0f}% ({h.reason})") EOF ``` ### Invalidate Thesis ```bash PYTHONPATH=. python3 << 'EOF' from src.knowledge.thesis import ThesisTracker from src.core.paths import paths THESIS_ID = "abc12345" # Replace with actual ID REASON = "Oil price collapse below $55 makes Venezuela uneconomic" tracker = ThesisTracker(paths.theses) if tracker.invalidate_thesis(THESIS_ID, REASON): print(f"Thesis {THESIS_ID} invalidated") print(f"Reason: {REASON}") else: print(f"Failed to invalidate thesis") EOF ``` ## CRITICAL: Load Trading Patterns First Before creating or reviewing any thesis, read the accumulated trading wisdom: ```bash cat docs/TRADING_PATTERNS.md ``` Key patterns to apply: - **Pattern 1**: Vehicle Enumeration - enumerate ALL beneficiaries - **Pattern 2**: Converging Signals - need 3+ signals aligned - **Pattern 3**: Data Source Check - what alt-data supports this? - **Pattern 8**: Derivative Plays - are there better vehicles? ## Workflow ### Creating a New Thesis 1. **Identify the belief**: What do you believe that consensus doesn't? 2. **Define signposts**: What events will prove/disprove the thesis? 3. **Set invalidation triggers**: What would force you to abandon it? 4. **ENUMERATE ALL VEHICLES** (Pattern 1 - CRITICAL): - Direct beneficiaries - Suppliers - Customers - ETFs - Adjacent sectors - Options plays - International exposure - Short candidates (losers) 5. **Check for signal convergence** (Pattern 2) 6. **Set conviction**: How confident are you? (0-100%) ### Reviewing a Thesis 1. Check if any signposts have been triggered 2. Review conviction history - has it been declining? 3. Check if invalidation triggers have occurred 4. Update conviction based on new information 5. Schedule next review ### Integration with Trading When making a trade decision: 1. Check if symbol is linked to an active thesis 2. Ensure thesis conviction supports the trade 3. Record thesis_id in the TradingDecision 4. After trade closes, update thesis if outcome is informative ## File Location Theses are stored as YAML files in: ``` ~/quant_results/theses/{thesis_id}.yaml ``` ## Thesis Structure ```yaml id: abc12345 name: Venezuela Energy Recovery created: 2026-01-06T08:00:00 status: active # active, validated, invalidated, expired summary: Sanctions relief drives oilfield services rally bull_case: Chevron license + broader relief → contracts bear_case: Political reversal, oil collapse, delays conviction: 65 signposts: - description: Chevron license extended status: pending bullish_if: Full renewal bearish_if: Only cosmetic invalidation_triggers: - Maduro action forces reversal - Oil below $60 positions: - SLB - HAL last_review: 2026-01-06T08:00:00 next_review: 2026-01-13T08:00:00 review_interval_days: 7 conviction_history: - timestamp: 2026-01-06T08:00:00 old_value: 50 new_value: 65 reason: Initial thesis creation notes: - "[2026-01-06 08:00] Created thesis based on Venezuela sanctions news" ```