Back to Full Document

Technical Architecture — Due Diligence

TRADARS Analytics

An adaptive intelligence platform for institutional-grade macro analysis with a proprietary learning engine.

v2.5
Production
72
Live Instruments
15+
Data Sources
3+ Years
In Production
TRADARS Analytics — Technical ArchitectureConfidential1 / 10

System Overview

Three-Layer Intelligence Architecture

Layer 1: Data Ingestion

72 instruments from 15+ APIs with multi-source reconciliation and automated deduplication.

Layer 2: Analytical Engine

Weighted risk scoring, Pearson correlations, 20-year seasonality, oil regime detection, MECE taxonomy.

Layer 3: Learning & Memory

Outcome tracking, statistical pattern discovery, insight extraction, institutional memory accumulation.

Key Differentiator: The feedback loop. Bloomberg stops at Layer 2. TRADARS Analytics completes the cycle.

TRADARS Analytics — Technical ArchitectureConfidential2 / 10

Core Algorithm

Risk Sentiment Scoring Engine

AssetSentiment = (Δ1D×0.25 + Δ7D×0.40 + Δ30D×0.35) × Polarity // +30% Trend Consistency Bonus when all timeframes agree // COT Integration: 75% Price + 25% Institutional Positioning // Breadth Amplifier: score × (1 + (dominance-0.65) × 1.15) OverallScore = (Score_1D×0.25 + Score_7D×0.40 + Score_30D×0.35) × 2 // Scale: -10 (Strong Risk Off) to +10 (Strong Risk On)

14 weighted cross-asset signals including VIX (w=20), S&P 500 (w=15), HYG Junk Bonds (w=15), DXY (w=12), yields, gold, crypto, and dynamically-flipped WTI crude.

TRADARS Analytics — Technical ArchitectureConfidential3 / 10

Innovation Spotlight

Dynamic Oil Regime Detection

The Problem

Rising oil prices can be bullish (demand growth) or bearish (supply shock). Every other platform treats oil with a fixed polarity.

Our Solution

TRADARS Analytics compares WTI direction vs. SPX direction across 1D (40% weight) and 7D (60% weight), plus NLP keyword scanning of news headlines for supply/demand signals.

Demand Regime: Oil ↑ with Stocks ↑ → Risk ON
Supply Regime: Oil ↑ with Stocks ↓ → Risk OFF
Mixed: Inconclusive → Default polarity

Bloomberg shows oil prices. We tell you whether rising oil is bullish or bearish right now.

TRADARS Analytics — Technical ArchitectureConfidential4 / 10

Quantitative Infrastructure

Correlation Matrix & Seasonality Engine

Pearson Correlation Matrix

  • • 90-day rolling windows of daily log returns
  • • Weekend data excluded (crypto/forex normalization)
  • • Minimum 5 overlapping data points required
  • • ρ > 0.65 flagged as "strong" for risk analysis
  • • Breakdown detection triggers regime shift alerts

20-Year Seasonality Engine

  • • 26 instruments across all asset classes
  • • Monthly average, median, and win rate statistics
  • • Dual window: Full history vs. recent 5-year
  • • Risk Seasonality = SPX Median - VIX Median × 0.10
  • • Window divergence = regime shift detection
TRADARS Analytics — Technical ArchitectureConfidential5 / 10

The Neural Layer — Core Competitive Moat

Adaptive Learning Architecture

Input Layer:72 live feeds + COT + economic events + central bank rates → normalized
Hidden Layer:14-asset sentiment model + oil regime + breadth amplification → scored
Output Layer:Risk score ±10 + per-asset conviction + trade thesis → published
Backpropagation:Outcome tracking → pattern discovery → weight adjustment → improved predictions
Long-Term Memory:AidenInsight store + WayneKnowledgeBase → persistent institutional context
No Bloomberg, no TradingView, no retail platform has a feedback loop. They are static — same analysis on Day 1 as Day 1,000.
TRADARS Analytics — Technical ArchitectureConfidential6 / 10

Weekly Automated Learning

Statistical Pattern Discovery

1. Fetch resolved AnalysisOutcome records (Success/Failure) 2. Compute: win rate, avg win %, avg loss %, expectancy 3. Segment by: score threshold, direction, asset 4. LLM extracts 2-5 structured patterns 5. Patterns (N≥10) injected into future Aiden analysis // Closes the loop: outcomes → patterns → better predictions

Combined with Conversational Insight Extraction — every strategic discussion generates structured insights stored as persistent memory with market condition snapshots (DXY, VIX, risk score).

TRADARS Analytics — Technical ArchitectureConfidential7 / 10

Not a ChatGPT Wrapper

Aiden — Framework-Constrained AI

What Makes Aiden Different

  • • No RSI, MACD, Elliott Wave — prohibited
  • • No internet search — data sovereignty
  • • Live database context on every invocation
  • • Cites its own track record from LearnedPattern
  • • References past conversations from AidenInsight

Context Assembled Per Call

  • • 16 key instrument prices (multi-timeframe)
  • • Risk score + environment classification
  • • COT positioning for 15 economies
  • • Upcoming high-impact events
  • • G10 central bank rates
  • • Wayne's Knowledge Base (40+ yr institutional)
  • • Accumulated insights + learned patterns
TRADARS Analytics — Technical ArchitectureConfidential8 / 10

Accounting-Identity Guardrails

MECE Macro Data Taxonomy

Growth

GDP = C+I+G+NX

NFP, Retail Sales, PMI, GDP

Inflation

M×V = P×Y

CPI, PPI, PCE, Commodities

Liquidity

Policy → Rates → Curve

Fed Rate, OIS, 2Y-10Y Spread

External

CA + KA = 0

COT, VIX, DXY, Trade Balance

33 nodes across 4 pillars. Each leaf maps to specific data agencies and entity fields. Z-score anomaly detection at |Z| > 2.0. Cross-pillar divergence signals regime shifts.

TRADARS Analytics — Technical ArchitectureConfidential9 / 10

Why TRADARS Analytics

Summary

✓ 14-asset weighted risk engine with COT integration and dynamic oil regime detection

✓ Adaptive learning system that discovers patterns from its own outcome data weekly

✓ Framework-constrained AI grounded in 40+ years of institutional knowledge

✓ MECE macro taxonomy with accounting-identity guardrails preventing double-counting

✓ Data sovereignty — no external AI contamination, multi-source reconciliation

✓ White-label ready — CSS variable theming, days to deploy

For the full document with complete mathematical formulas: /TechnicalArchitecture

For the business case (management/sales): /BusinessCase

TRADARS Analytics — Technical ArchitectureConfidential10 / 10