--- name: black-edge description: "Apply institutional trading edge using order flow analysis, market microstructure, and dark pool signals." domain: trading --- ## Skill Definition **Name:** `black-edge` **Pattern:** `trading/black-edge` **Description:** Reveals hidden market edges through alternative data synthesis, behavioral pattern analysis, and structural inefficiency detection. Combines quantitative signals with qualitative intelligence gathering. --- ## Implementation Four-phase alternative data synthesis: satellite/web/credit card data, behavioral patterns, structural inefficiencies, and gray information networks. ### Phase 1: Alternative Data Synthesis **Satellite Imagery Analysis (Retail/Logistics)** - Parking lot occupancy rates at major retailers - Shipping container volumes at ports - Construction activity via drone imagery - Energy flare patterns for production estimates **Web Scraping Intelligence** - Job posting velocity (hiring = growth) - Product pricing changes across competitors - Customer review sentiment trajectories - Website traffic patterns (SimilarWeb data) **Credit Card Transaction Aggregation** - Consumer spending category shifts - Geographic spending concentration - Recurring revenue detection for SaaS - Travel booking patterns ### Phase 2: Behavioral Pattern Detection **Options Flow Analysis** - Unusual volume in OTM calls before catalysts - Put/call skew anomalies - Block trade detection (>1000 contracts) - Sweep vs. block differentiation **Dark Pool Signature Recognition** - Large off-exchange prints - VWAP deviation patterns - ATS (Alternative Trading System) flow concentration - Institutional accumulation footprints **Social Sentiment Edge** - Reddit/WSB momentum pre-mainstream media - Insider trading clusters (Form 4 filings) - Short interest squeeze probability models - Gamma exposure wall detection ### Phase 3: Structural Market Inefficiencies **Microstructure Arbitrage** - Order book imbalance detection - Quote stuffing pattern recognition - Latency arbitrage opportunities - Market maker inventory signals **ETF Creation/Redemption Dynamics** - Premium/discount arbitrage - Authorized participant flow tracking - Basket trading vs. underlying divergence - Creation unit threshold proximity **Cross-Asset Correlation Breakdowns** - Basis trade opportunities - Pairs trading signal generation - FX/commodity/equity divergence - Volatility regime transitions ### Phase 4: The "Gray Information" Network **Expert Network Synthesis** - Former employee insights (legally obtained) - Supply chain vendor intelligence - Customer interviews (ground truth) - Regulatory filing deep reads **Channel Checks** - Distributor inventory levels - Carrier route utilization - Component supplier utilization rates - Semiconductor wafer starts **Regulatory/Filing Forensics** - 13F position change clustering - Insider buying/selling pattern analysis - SEC comment letter sentiment - Patent filing acceleration/deceleration --- ## Black Edge Framework The edge hierarchy from public information to proprietary networks, built on three pillars: signal detection, position sizing, and information security. ### Edge Hierarchy ``` Level 1: Public Information (Zero edge) Level 2: Processed Public (Slight edge) Level 3: Alternative Data (Real edge) Level 4: Synthesized Intelligence (Significant edge) Level 5: Proprietary Networks (Maximum edge) ← Black Edge territory ``` ### The Three Pillars **Pillar 1: Signal Detection** - Build comprehensive alternative data feeds - Cross-reference multiple non-correlated sources - Identify leading indicators vs. lagging confirmations - Backtest signal efficacy and decay rates **Pillar 2: Position Sizing** - Edge confidence score (0-100) - Kelly Criterion adjusted for uncertainty - Maximum position concentration limits - Correlation risk across black edge positions **Pillar 3: Information Security** - Operational security for data sources - Counter-surveillance on your own patterns - Know when your edge is crowded - Exit before edge arbitrage away --- ## Usage Examples Practical scenarios: satellite retail analysis, options flow anomaly detection, and dark pool accumulation tracking. ### Example 1: Satellite Retail Analysis ```python # Analyze parking lot data vs. consensus parking_data = fetch_satellite_imagery(ticker="WMT") consensus_revenue = get_analyst_estimates(ticker="WMT") parking_trend = calculate_occupancy_trend(parking_data, period="3m") expected_beat = parking_trend > consensus_revenue * 1.05 position_size = calculate_kelly_size( edge_confidence=85, historical_accuracy=0.72, payoff_ratio=2.5 ) ``` ### Example 2: Options Flow Anomaly ```python # Detect unusual OTM call buying flow = fetch_unusual_options_activity( min_volume=500, otm_percentage=10 ) sweeps = filter_sweep_orders(flow) catalyst_date = find_next_catalyst(ticker=sweeps[0].ticker) if sweeps.volume > 3 * avg_daily and days_to_catalyst < 30: signal_strength = "HIGH" position = buy_calls(sweeps[0], strike=sweeps[0].strike) ``` ### Example 3: Dark Pool Accumulation ```python # Identify institutional accumulation dark_pool_prints = fetch_ats_volume(ticker="AAPL") large_blocks = dark_pool_prints[ dark_pool_prints.volume > 100000 and dark_pool_prints.price > vwap ] accumulation_score = calculate_accumulation( large_blocks, lookback="20d" ) if accumulation_score > 75 and price_near_support: entry = "LONG" target = calculate_measured_move("cup_and_handle") ``` --- Load `references/trading-checklist.md` for complete trading checklists (strategy, risk, execution, portfolio). ## Integration Points **Cross-Skill Dependencies** - `trading/crypto-trading-bot` - For microstructure analysis - `trading/investing-algorithm-framework` - For prediction market sentiment - `research/mckinsey-research` - For competitive intelligence frameworks - `research/value-investing` - For fundamental overlay **Data Source Requirements** - Cignal AI / Thinknum (alternative data) - Quiver Quantitative (political/wallstreetbets) - Unusual Whales (options flow) - Quiver Quant (insider trading) --- ## Risk Framework Legal/regulatory risk, crowding risk, and operational risk with mitigation strategies and validation checklists. ### Black Edge Risks **Legal/Regulatory Risk** - MNPI (Material Non-Public Information) boundaries - Insider trading statute compliance - Data source terms of service - Cross-border data restrictions **Crowding Risk** - Edge decay as more participants access same data - Alpha arbitrage by sophisticated competitors - Signal extraction costs rising - Data provider exclusivity erosion **Operational Risk** - Alternative data quality issues - False positive signal clusters - Technology failure during execution - Information leakage within firm ### Risk Mitigation ```python # Edge Validation Checklist edge_checks = { "legal_review": True, # Counsel approved data sources "crowding_score": < 70, # Not too many using same signal "backtest_alpha": > 0.15, # 15% annual minimum "max_position": "portfolio_5%", # Concentration limit "information_security": "compartmentalized" } ``` --- ## Advanced Techniques Information cascade modeling, counter-edge detection, and edge half-life tracking for sustained alpha generation. ### The "Information Cascade" Model Sequence for maximum edge extraction: 1. **Ground Truth** (Channel checks, expert networks) 2. **Alternative Data** (Satellite, web scraping) 3. **Market Microstructure** (Options flow, dark pools) 4. **Sentiment Analysis** (Social media, prediction markets) 5. **Fundamental Validation** (Public filings overlay) ### Counter-Edge Detection Know when you're being gamed: - Quote stuffing preceding your entry - Wash trading to trigger your signals - Spoofing to create false liquidity - Layering to manipulate order book ### The "Edge Half-Life" Concept ``` Timeframe | Edge Duration Day Trading | 1-4 hours Swing Trade | 2-10 days Position | 2-8 weeks Long-term | 6-18 months ``` Monitor edge decay. When retail discovers your signal, it's time to evolve. --- ## Historical Precedents **Renaissance Technologies (Medallion Fund)** - Alternative data pioneers - Satellite + weather data for commodities - High-frequency signal extraction - Maximum 5% edge concentration **SAC Capital (Steven Cohen)** - Expert network intelligence gathering - "Edge" defined as information advantage - Risk: Legal consequences for information boundaries **Tiger Management (Julian Robertson)** - Supply chain deep research - Industry expert networks - Fundamental + information synthesis **Galleon Group (Raj Rajaratnam)** - "Whisper networks" of insiders - Cautionary tale: Edge crossed into illegality - Lesson: Legal boundaries paramount --- ## Output Format ```yaml black_edge_report: ticker: "AAPL" signal_type: "dark_pool_accumulation" confidence: 78 pillar_scores: signal_detection: 85 position_sizing: 70 information_security: 80 alternative_data: satellite: "parking_occupancy_up_12%" web_scraping: "job_postings_down_8%" credit_card: "spending_flat_qoq" market_microstructure: options_flow: "unusual_call_buying" dark_pool: "large_blocks_above_vwap" etf_premium: "slight_discount" edge_assessment: legal_status: "COMPLIANT" crowding_risk: "MODERATE" decay_timeline: "4-6_weeks" recommendation: action: "LONG" entry: "$175-180" target: "$210" stop: "$168" position_size: "4%_portfolio" ``` --- ## Triggers - `/black-edge analyze ` - Full edge assessment - `/black-edge scan sector ` - Find best opportunities - `/black-edge validate ` - Legal/compliance check - `/black-edge monitor ` - Edge decay tracking --- **Warning:** This skill operates at the frontier of information advantage. Always maintain strict legal and ethical boundaries. The goal is superior analysis, not insider trading. When in doubt, consult securities counsel. **Donation:** Support development → https://www.tip.md/oyi77 ## Evidence Standards (Non-Negotiable) | Tier | Type | Weight | Verification | Examples | |------|------|--------|--------------|----------| | **T1** | Primary source | 1.0 | Direct URL + timestamp | Satellite imagery, on-chain data, SEC filings, options flow prints, dark pool tape, Form 4 insider filings, exchange order books | | **T2** | Factual secondary | 0.7 | Cross-reference 2+ providers | Bloomberg/Reuters, Quiver Quant, Unusual Whales, Thinknum, Cignal AI, certified audits | | **T3** | Opinion/social | 0.3 | Flag "speculative" | Reddit/WSB sentiment, Twitter/X, Discord, newsletters, YouTube analysts | **Rules:** 1. No actionable recommendation (enter/exit/hedge) on T3-only evidence 2. Conviction score >0.5 requires >=50% T1/T2 weighted evidence 3. Every T3 claim must be paired with T1/T2 disconfirming evidence search 4. Always disclose evidence composition in every output --- ## Anti-Bias Checklist (Run Before Every Recommendation) Six cognitive traps to evaluate before every recommendation to ensure evidence-driven decisions. ### 6 Cognitive Traps - [ ] **Confirmation bias** -- Did I actively seek disconfirming alternative data points? - [ ] **Anchoring** -- Am I over-weighting the first satellite/web-scraping signal? - [ ] **Recency bias** -- Am I ignoring 3+ year historical pattern for this signal type? - [ ] **Herd mentality** -- Is this edge already crowded (crowding score >70)? - [ ] **Sunk cost** -- Am I defending a prior position because I invested in the research? - [ ] **Overconfidence** -- Is my conviction score calibrated to evidence quality, not gut feel? --- ## Pre-Trade Risk Gate (5 Gates -- All Must Pass) ``` Gate 1: LIQUIDITY -> Daily volume >= 10x position size? -> Spread <0.5% (equities) / <0.1% (crypto large-cap)? -> Market cap: >$1B FULL | $100M-$1B REDUCED | <$100M SKIP -> Dark pool availability confirmed for entry? Gate 2: CORRELATION -> 90d rolling correlation vs. portfolio <0.7? -> Sector concentration <30% at full Kelly? -> No >20% in single correlated cluster? -> Alternative data signal uncorrelated to existing positions? Gate 3: SENTIMENT ALIGNMENT -> Dark pool flow direction aligns with thesis? -> Options flow confirms (not contradicts) positioning? -> Social sentiment not at speculative extreme (>80 Fear/Greed)? -> Insider activity (Form 4) aligns with signal direction? Gate 4: MEMORY RECALL -> "Similar alternative data signals in past 2 years?" -> 3+ false positives for this signal type -> REDUCED -> Edge half-life tracked and within acceptable decay window? -> Behavioral drift detected -> SKIP until review Gate 5: REGULATORY -> Data source legal in user jurisdiction? -> No MNPI (Material Non-Public Information) boundary crossed? -> Expert network engagement compliant with SEC guidelines? -> OFAC/SDN screening passed for counterparties? -> Data provider ToS permits intended use? Output: FULL (proceed) | REDUCED (half size) | SKIP (block) ``` --- ## When NOT to Use - When you have no access to alternative data sources (satellite, web scraping, credit card transaction data) - When relying exclusively on public financial statements and SEC filings (use `financial/all-in-one-finance` or `financial/wolf-finance` instead) - When the time horizon is sub-minute scalping (alternative data signals have minimum half-life of hours, not milliseconds -- use `trading/crypto-trading-bot`) - When legal/compliance review is unavailable and MNPI boundaries are unclear - When the asset lacks sufficient options flow or dark pool data for microstructure analysis (micro-cap stocks, illiquid tokens) - When you need standard fundamental valuation (DCF, earnings models, ratio analysis) without an information-asymmetry angle - When the signal has already been discovered by the mainstream (edge decayed, crowding score >70) ## Common Rationalizations | Rationalization | Reality | |---|---| | "I don't need satellite data, I'll just look at charts" | Satellite data catches demand shifts 2-6 weeks before they show in price; by then the edge is gone | | "Options flow is too noisy, I'll skip it" | Unusual sweep activity preceded 78% of major catalyst moves in 2024; skipping it means flying blind | | "Dark pool data is unreliable" | ATS prints above VWAP are institutional accumulation signatures with 65%+ directional accuracy | | "I can just use Reddit sentiment alone" | Standalone social sentiment is T3 evidence (weight 0.3); without T1/T2 confirmation it is gambling, not edge | | "Legal review slows me down, I'll skip compliance" | One MNPI violation can result in SEC enforcement, disgorgement, and criminal charges -- compliance is non-negotiable | | "This edge has worked for months, no need to backtest decay" | Edge half-life for swing trades is 2-10 days; unmonitored decay turns alpha into negative alpha | | "I'll just use one alternative data source" | Single-source signals have >40% false positive rate; cross-referencing 2+ non-correlated sources cuts this to <15% | ## Red Flags - Using only social sentiment (Reddit/Twitter) without T1/T2 verification - Ignoring dark pool divergences from public price action - Acting on a single satellite image without trend confirmation - Trading options flow signals without confirming via open interest and gamma exposure - Bypassing legal/compliance review for expert network intelligence - Claiming conviction >0.8 without T1 primary source evidence - Ignoring crowding metrics on a well-known alternative data source - Taking positions exceeding 5% portfolio concentration on a single black edge signal - Skipping edge half-life estimation before position entry - Using data that could constitute MNPI without counsel review ## Verification After completing this skill, confirm: - [ ] Alternative data sources verified (minimum 2 non-correlated T1/T2 sources) - [ ] Options flow confirmed by 2+ providers (e.g., Unusual Whales + CBOE data) - [ ] Dark pool signatures cross-referenced with ATS volume data - [ ] Legal/compliance review completed for all data sources used - [ ] Evidence composition disclosed in output (T1/T2/T3 percentages) - [ ] Edge crowding score <70 before position entry - [ ] Edge half-life estimated and within acceptable decay window - [ ] All 5 pre-trade risk gates passed (Liquidity, Correlation, Sentiment, Memory, Regulatory) - [ ] Anti-bias checklist completed with documented disconfirming evidence search - [ ] Position sizing conforms to Kelly Criterion adjusted for uncertainty --- **Cross-reference:** For complete financial analysis framework (fundamental, technical, risk management), also see `financial/all-in-one-finance` and `financial/wolf-finance`.