Veritas • Executive Methodology & Mathematical Disclosure

⚠️ Mandatory Research & Accuracy Disclosure

Academic & Quantitative Research Only

STRICT RESEARCH NOTICE • ASSUMPTION OF INACCURACY: All models, simulations, algorithmic signals, parameter values, factor regressions, and analytical outputs presented across this platform are provided strictly for quantitative and academic research purposes only. All data points, financial metrics, and market values are assumed to be unverified and inaccurate until independently audited and verified against official regulatory filings (SEC XBRL) and primary exchange trade records. Nothing on this website constitutes investment, legal, tax, or financial advice.

MSF & MSBA DUAL GRADUATE SPECIALIZATION • REMOTE CONTRACT READY

Institutional Quantitative Engineering • Fully Remote 20 hr/Week Engagement

Currently advancing dual graduate specializations in Master of Science in Finance (MSF) and Master of Science in Business Analytics (MSBA). I engineer institutional-grade financial intelligence engines bridging Wharton Research Data Services (WRDS) econometrics, heavy-tail stochastic calculus (Student's t jump diffusions), and zero-latency client-side WebUI delivery.

Engagement Target

Remote 20 hr/Wk Contract

Quant Dev • Risk Analytics • Data Pipelines

1.1 Drift-Corrected Geometric Brownian Motion (Itô's Lemma)

In standard discrete simulations, compounding raw volatility induces an artificial upward drift. Under continuous stochastic calculus, Itô's Lemma specifies that the log-price process satisfies:

Stochastic Differential Equation:

$$dP_t = \mu P_t dt + \sigma P_t dW_t \implies P_t = P_0 \exp\left(\left(\mu - rac{1}{2}\sigma^2 ight)t + \sigma W_t ight)$$

• Strict Limited Liability: $P_t > 0$ strictly holds for all $t \in [0, T]$ via exponential mapping.
• Drift Neutrality: $-rac{1}{2}\sigma^2$ removes artificial price inflation, ensuring martingale consistency under the risk-neutral measure.

1.2 Student's t Heavy-Tail Innovations vs. Gaussian Fallacy

Standard Gaussian assumptions catastrophically underestimate flash-crash probabilities. In our engine, Wiener increments $dW_t$ are replaced with Student's t jump innovations with fitted degrees of freedom $ u \in [3.0, 8.0]$:

Leptokurtic Jump Kernel:

$$f(x; u) = rac{\Gamma\left(rac{ u+1}{2} ight)}{\sqrt{\pi u}\,\Gamma\left(rac{ u}{2} ight)} \left(1 + rac{x^2}{ u} ight)^{-rac{ u+1}{2}}, \quad ext{Kurtosis} = rac{6}{ u - 4} + 3 \quad ( ext{for } u > 4)$$

Captures $5\sigma$ to $8\sigma$ tail dislocations that occur in live markets while preserving analytic tractability across 10,000 CUDA paths.

Tail Risk Metrics (VaR & CVaR)

99% Value at Risk (VaR)

$$ ext{VaR}_lpha = -\inf \{ l \in \mathbb{R} : P(L > l) \le 1 - lpha \}$$

Cutoff boundary for worst 1% 30-step tail outcomes.

99% Conditional VaR (Expected Shortfall)

$$ ext{CVaR}_lpha = \mathbb{E}[L \mid L \ge ext{VaR}_lpha]$$

Average magnitude of loss in the catastrophic tail. Strictly enforced $ ext{CVaR} \ge ext{VaR}$.

Mathematical Guarantee

Every asset in our 242-asset universe undergoes empirical kurtosis fitting. Sub-Gaussian models are rejected.

2.1 Fama-French 5-Factor + Momentum Empirical Decomposition

208 overlapping trading days between Wharton stock prices and WRDS benchmark series.

HARVEY t ≥ 3.0 THRESHOLD

Empirical Econometric Specification:

$$R_{i,t} - R_{f,t} = lpha_i + eta_{Mkt}(R_{m,t}-R_{f,t}) + eta_{SMB}SMB_t + eta_{HML}HML_t + eta_{RMW}RMW_t + eta_{CMA}CMA_t + eta_{UMD}UMD_t + \epsilon_{i,t}$$

Market Factor (Mkt-RF)

β = +1.2119

t = +18.51 (p < 0.0001)

Strong high-beta tech sensitivity. High factor significance.

Size Factor (SMB)

β = -0.4242

t = -5.22 (p < 0.0001)

Significant negative SMB loading. Confirms mega-cap dominance.

Value Factor (HML)

β = -0.5909

t = -9.08 (p < 0.0001)

Extreme negative HML loading. Portfolio is heavily tilted to pure Growth.

Momentum (UMD)

β = -0.0685

t = -1.62 (Insignificant)

Statistically insignificant (|t| < 2.0). Returns are not chasing trailing trend.

Variance Decomposition Breakdown

Systematic Factor Explained (R²) 78.15%

Idiosyncratic Alpha Variance (1-R²) 21.85%

Annualized Excess Alpha (α) +435.4 bps/yr (t = +0.39)

Institutional Reporting Transparency

We reject backtests claiming “85%+ idiosyncratic alpha.” Live market regression proves 78.15% of mega-cap tech variance is factor-driven. True alpha is +435.4 bps gross deliverable.

3.1 Execution Friction & Market Impact Waterfall

Gross alpha is a theoretical illusion without calibrated institutional market frictions. Our cost model (Core/Validation/cost_model.py) applies nonlinear square-root price impact:

Square-Root Market Impact Equation:

$$I(Q) = k \cdot \sigma \cdot \left(rac{Q}{ ext{ADV}} ight)^lpha, \quad k = 0.10, \; lpha = 0.50$$

Gross Alpha

+435.4 bps

Half-Spread

-15.0 bps

Impact I(Q)

-22.5 bps

Net Alpha

+397.9 bps

3.2 Outlier Treatment & Panel Winsorization

Across 5,695 equities and 5.09M CRSP rows, unadjusted micro-caps (e.g. reverse split squeezes like TPST) generate extreme percentage anomalies (+2,000,000%). We enforce systematic two-sided Winsorization:

1st and 99th Percentile Winsorization:

$$x_{ ext{winsor}} = \min\left(\max(x, q_{0.01}), q_{0.99} ight)$$

Protects cross-sectional covariance matrices and Sharpe rankings from penny-stock distortion while preserving genuine institutional signals.

Bailey & López de Prado Deflated Sharpe Ratio (DSR):

In our Alpha Lab (686 evolution epochs), DSR explicitly penalizes strategy Sharpe ratios for the number of failed trials $K=686$. Rejecting overfit runs with “FAILED_DEFLATOR” guarantees out-of-sample survival.

4.1 California Freight Corridor Telemetry (Caltrans Real-Time Pulses)

Alternative physical supply chain signals loaded from Data/Logistics/Caltrans/caltrans_corridor_pulses.json.

15,000+ SENSOR PULSES

San Pedro Bay z = +0.15

I-710 / I-110 container discharge from Port of LA & Long Beach.

Target: MAERSK, ZIM, FDX, UPS

Silicon Valley Transit z = -0.48

I-880 / US-101 / SR-237 hardware component freight between Oakland & South Bay.

Target: NVDA, AAPL, TSLA, GOOGL

Inland Empire Hub z = +0.15

I-10 / SR-60 mega-fulfillment inflow to Mira Loma & Ontario hubs.

Target: AMZN, WMT, TGT, COST

Otay Mesa Nearshoring z = -0.29

SR-905 / SR-125 cross-border electronics manufacturing flow.

Target: KSU, UNP, CNI, MEX_TRADE

Alpha Synthesis Mechanism: Corridor velocity anomalies provide a 3-to-6 week leading indicator for enterprise hardware backlog bottlenecks before they materialize in Compustat inventory turns.

5.1 Intellectual Property (IP) Boundary & Reverse Pressure ($R_p$)

To preserve institutional trade secrets while providing total transparency into data pipelines, our codebase maintains strict boundary isolation:

✓ Fully Disclosed & Open for Code Review:

Wharton WRDS data ingestion, CUDA Monte Carlo path generation, Fama-French regressions, Winsorization, Deflated Sharpe Ratio calculation, and the complete zero-dependency WebUI design system.

🔒 Proprietary IP Core (Abstracted / Binary Stubs):

The biophysical master equation weights and exact phase-space transition mechanics for Reverse Pressure ($R_p$) are compiled in binary C-extensions (_biophysical_core.so).

5.2 Distributed Shard Topology & Security

Shard 0 (Mac Studio M3 Ultra) Orchestrator • Telemetry • MLX

Shard 1 (CyberGamingPC RTX 4070S) 10,000-Path CUDA Swarm

Shard 3 (Synology NAS 70TB RAID) Wharton Database Vault

Dedicated Interconnect 10Gbps SFP+ Network (10.10.10.x)

Access Governance Cloudflare Zero Trust RBAC