Analysis Methodology &
Data Processing Architecture
A rigorous, multi-layered analytical framework that synthesizes macroeconomic liquidity signals, cross-asset capital flow dynamics, and quantitative technical analysis into actionable market intelligence. TraderPath is an analysis and research platform — not a broker, advisor, or trading service.
1. Capital Flow Intelligence Framework
Top-down liquidity analysis across four hierarchical layers
Our analytical thesis is rooted in a fundamental principle of financial markets: capital allocation precedes price action. Rather than relying on isolated technical signals, we construct a hierarchical model that tracks the flow of liquidity from central bank balance sheets through asset classes, sectors, and ultimately into individual instruments. This top-down architecture maximizes the signal-to-noise ratio by filtering opportunities through progressively granular lenses.
Global Liquidity Tracker
Macro regime classification
Ingests Federal Reserve balance sheet data, M2 money supply growth rates, DXY (US Dollar Index) momentum, VIX term structure, and yield curve dynamics to determine the prevailing liquidity regime: risk-on, risk-off, or transitional.
Market Flow Analyzer
Cross-asset rotation detection
Measures 7-day and 30-day capital flow velocity across four asset classes (Crypto, Equities, Fixed Income, Precious Metals) to identify rotation patterns and phase-classify each market as EARLY, MID, LATE, or EXIT.
Sector Drill-Down
Intra-market capital concentration
Within the identified primary market, sector-level TVL (Total Value Locked) flows, trading volume distribution, and dominance shifts are analyzed to pinpoint which vertical is attracting disproportionate capital inflow relative to peers.
Asset Analysis Engine
Instrument-level quantitative analysis
The final layer deploys our dual-engine analytical suite (7-Step Classic + MLIS Pro) on the recommended instruments, producing a structured verdict with calibrated confidence levels, directional conviction, and a risk-bounded trade plan.
Phase Detection & Rotation Model
Capital flows follow observable cyclical patterns. We measure the duration and velocity of flow into each market to classify the current phase. Statistical analysis of historical phase durations provides expected holding periods and transition probabilities.
0-30
Optimal entry zone. Capital inflow accelerating with positive velocity delta.
30-60
Trend maturing. Flow velocity stabilizing; position sizing should decrease.
60-90
Trend exhaustion imminent. Negative flow acceleration detected.
90+ / Reversal
Capital outflow confirmed. Rotation to alternative asset class underway.
2. 7-Step Analysis Engine
Sequential, gated analysis pipeline with 40+ quantitative indicators
The 7-Step engine is a sequential pipeline where each step acts as a gate. If a preceding step flags critical risk (e.g., Safety Check detects wash trading), the pipeline can terminate early with a AVOID verdict, conserving computational resources and preventing analysis on compromised data. Each step produces a sub-score on a normalized 0-10 scale, which feeds into the final weighted ensemble.
Market Pulse
Macro-regime classification via composite scoring of volatility surfaces (VIX term structure), cross-asset correlation matrices, and funding rate divergences.
Asset Scanner
Multi-timeframe technical decomposition using 40+ indicators with adaptive weighting based on current volatility regime.
Safety Check
Microstructure analysis combining order book imbalance detection, wash-trading filters, and liquidity depth assessment.
Timing Analysis
Entry optimization through confluence of support/resistance clustering, economic calendar event proximity, and intraday volume profile.
Trade Plan
Structured position construction with AI-optimized entry, stop-loss, and take-profit levels derived from ATR-adjusted support/resistance.
Trap Check
Counter-trend signal detection by analyzing liquidation heatmaps, whale position clustering, and smart money flow divergence.
Final Verdict
Weighted ensemble verdict aggregating all prior steps with confidence intervals and directional conviction scoring.
3. MLIS Pro — Machine Learning Inference System
5-layer neural confirmation engine with calibrated confidence outputs
MLIS Pro serves as an independent confirmation layer that cross-validates the 7-Step engine output. It processes raw market data through five specialized neural inference layers, each extracting orthogonal features. The final Verdict Layer uses calibrated probability estimation (Platt scaling) to produce confidence-adjusted recommendations. When MLIS Pro contradicts the 7-Step verdict, the system automatically downgrades the conviction level, significantly reducing false positive rates.
Technical Layer
Convolutional feature extraction from raw OHLCV sequences across multiple lookback windows.
Output: Pattern recognition strengthMomentum Layer
Velocity and acceleration decomposition of trend indicators with regime-adaptive thresholds.
Output: Trend persistence probabilityVolatility Layer
GARCH-family variance modeling with regime-switching detection for risk surface estimation.
Output: Volatility regime classificationVolume Layer
On-balance volume divergence analysis with institutional flow estimation from tick-level data.
Output: Accumulation / Distribution scoreVerdict Layer
Ensemble aggregation with calibrated confidence output (Platt scaling) and recommendation mapping.
Output: STRONG_BUY / BUY / HOLD / SELLDual-Engine Confirmation Protocol
The interplay between the rule-based 7-Step engine and the neural MLIS Pro system creates a robust confirmation mechanism analogous to ensemble methods in predictive modeling.
Full Confirmation
Both engines agree on direction and conviction. Confidence level elevated. Highest-probability signal output.
Partial Confirmation
Direction aligned, conviction divergent. Confidence level maintained at base. Position sizing should be reduced.
Contradiction
Engines disagree on direction. Confidence automatically downgraded. Verdict shifts toward WAIT or AVOID.
4. AI Expert Panel (VOLTRAN Synthesis)
Multi-agent reasoning system for qualitative overlay
Beyond quantitative signals, our AI Expert Panel employs four specialized agents, each trained on domain-specific reasoning, to provide qualitative context. Their independent assessments are then synthesized by the VOLTRAN aggregation module into a unified narrative.
Interprets indicator confluence, candlestick patterns, and chart formations with contextual market commentary.
Evaluates position-level and portfolio-level risk metrics including correlation risk, concentration risk, and tail-event probability.
Analyzes on-chain whale movements, exchange flow netflows, and stablecoin supply dynamics for crypto assets.
Monitors for smart contract vulnerabilities, regulatory risks, exchange solvency indicators, and market manipulation signals.
5. Data Infrastructure & Sources
Institutional-grade data feeds with multi-provider redundancy
| Provider | Data Domain | Asset Coverage | Refresh Rate |
|---|---|---|---|
| FRED (Federal Reserve) | Macroeconomic | US monetary policy, treasury yields, M2 | Daily / Weekly |
| Binance | Crypto OHLCV + Order Book | 500+ crypto pairs | Real-time |
| Yahoo Finance | Equities, Metals, Bonds | Global equities, ETFs, commodities | 15-min delayed |
| DefiLlama | DeFi TVL & Protocol Flows | 2000+ DeFi protocols | Hourly |
| CoinGecko / CMC | Tokenomics & Fundamentals | Market cap, supply metrics | 5-min |
| Finnhub | Economic Calendar | FOMC, CPI, NFP, GDP events | Event-driven |
Data Integrity Protocol
- Multi-provider cross-validation with automatic fallback chains
- Outlier detection and wash-trading filters on raw OHLCV data
- Redis-backed caching layer with configurable TTL per data source
- Prisma query timeout middleware (15s) with slow-query logging
6. Verified Performance Metrics
Outcome-tracked results from live analysis outputs
All performance metrics below are derived from verified, outcome-tracked analyses. When a trade plan is generated, the system monitors price action against the defined take-profit and stop-loss levels using Binance Klines API (crypto) or Yahoo Finance (non-crypto). Outcomes are recorded only when a definitive TP or SL hit is confirmed — pending trades are excluded from performance calculations.
7. Platform Scope & Limitations
What TraderPath is and what it is not
TraderPath IS:
- • An analytical research and intelligence platform
- • A data processing engine for market analysis
- • An educational tool for understanding capital flows
- • A transparent system with verifiable performance metrics
- • A technology product, not a financial service
TraderPath is NOT:
- • A broker-dealer or trading platform
- • A registered investment advisor (RIA)
- • A provider of personalized financial advice
- • A fund manager or asset custodian
- • A guarantee of future performance or returns
Regulatory notice: TraderPath does not execute trades, hold customer funds, or provide personalized investment recommendations. All analysis outputs are informational and educational in nature. Users are solely responsible for their own trading and investment decisions.
Explore the Methodology in Action
See how our 4-layer Capital Flow framework, 7-Step engine, and MLIS Pro confirmation system work together to produce institutional-grade market intelligence.