Quantitative edge
built on live capital.

Quantitative equity strategies built on rigorous statistical methods and machine learning. Live track record with real capital. Volatility harvesting. Regime-adaptive alpha generation.

Live Performance
£100 invested at inception, Feb 2020 to present
Live · Verified
Live Strategy
S&P 500
Real capital · Broker statements available on request
Total Return
+156.9%
Live · Since Feb 2023
Sharpe Ratio
1.35
Annualised
Max Drawdown
-21.0%
Peak to trough
Track Record
3.2 yr
Live · Real capital · Verified
Strategy vs S&P 500, Since Inception
+0pp
Outperformance since February 2020, navigating the COVID crash, the 2022 rate shock, and sustained volatility regimes, with real capital from February 2023.
Strategy · £100 invested£500
S&P 500 · £100 invested£212
Sharpe vs benchmark1.35
Max drawdown vs S&P-21.0%
Analytics
Drawdown Analysis
Underwater equity curve, Feb 2020 to present
Max DD -21.0% 65.8% months positive
Trading Activity
Execution profile over the live period
Total Trades Executed
Across all live months
81
Avg Trades Per Month
High conviction, low frequency
2.1
Up Capture
% of S&P upside captured
143.9%
Capture Ratio
Up capture / down capture
2.07x
Monthly Return Distribution
Strategy vs S&P 500 · frequency of monthly returns
Positive skew +0.62 Mean +2.08% Median +2.30% Worst -14.8% Best +14.6%
Annual Returns
Year by Year, Strategy vs S&P 500
Cumulative Alpha vs S&P 500
Monthly Returns Heatmap
Live Backtest
JanFebMarAprMayJunJulAugSepOctNovDecFull Yr
Risk Profile
Extended Performance Metrics
Annualised Return
34.7%
Sharpe Ratio
1.35
Sortino Ratio
1.40
Calmar Ratio
1.66
Information Ratio
0.90
Max Drawdown
-21.0%
Win Rate (monthly)
65.8%
Avg Win / Avg Loss
1.62x
Beta to S&P 500
0.96
Correlation to S&P
0.55
Key Statistics
Best Month
+14.6%
Worst Month
-14.8%
Longest Win Streak
Consecutive months
7
Down Capture
% of S&P downside absorbed
69.5%
Annualised Alpha
vs S&P 500
+15.2%
About
Systematic strategies built with academic rigour and live capital.
Quantitative equity strategies developed at the intersection of statistical research and machine learning. The focus is on volatility harvesting, regime detection, and cross-sectional alpha, with every model validated out of sample before deployment.

All performance figures on this site represent real returns from a live, funded portfolio. Broker statements are available on request.
Academic & Research Affiliations
University of PortsmouthUNIVERSITY OF
PORTSMOUTH
Imperial College LondonIMPERIAL COLLEGE
LONDON
Queen Mary University of LondonQUEEN MARY
UNIVERSITY OF LONDON
Barts Cancer InstituteBARTS CANCER
INSTITUTE
SSRNSSRN
Publications
Working Papers & Preprints
View on SSRN →
Bootstrap Methods
Return Predictability or Risk Timing? Bootstrap Evidence from a Century of US Equity Data
Statistical framework examining whether apparent return predictability reflects genuine timing ability or compensation for time-varying risk exposure.
Working Paper
Digital Assets
Veridex: Privacy-Preserving Institutional Digital Asset Trading via Zero-Knowledge Proof Systems
Protocol design for institutional-grade digital asset trading using zero-knowledge proofs to balance transparency with execution privacy.
Working Paper
Credit
The CDS-Bond Basis: A Markov-Switching Friction Decomposition with Nonlinear Amplification
Regime-dependent decomposition of the CDS-bond basis using Markov-switching models to isolate funding, liquidity, and counterparty risk channels.
Working Paper
Factors
Currency Factor Combination and Cross-Sectional Pricing: Shrinkage Convergence, Alternative Methods, and the Identification of Volatility Risk
Empirical investigation of factor combination techniques for currency cross-sections, with focus on shrinkage estimators and volatility risk identification.
Working Paper
Portfolio Optimisation
Clustering-Based Alternatives to Mean-Variance Portfolio Optimisation
Novel portfolio construction methodology using clustering algorithms as an alternative to traditional mean-variance optimisation under estimation error.
Distributed
Machine Learning
Probability and Machine Learning Methods in Predictive Modelling: Applications to Alternative Data
Applied machine learning framework for predictive modelling using alternative data sources, with applications to sports analytics as a test domain.
Distributed
Strategy
How the edge is generated
The strategy combines multiple orthogonal alpha sources, including volatility harvesting, systematic factor exposure, and machine learning signals, each independently validated and sized through a Kelly-inspired portfolio optimiser. All signals are tested out of sample before deployment.
~
Volatility Harvesting
Proprietary signals derived from implied volatility, fixed income, and macro regime indicators. Positions adapt systematically to shifts in the broader risk environment.
0
Macro signals monitored
◇
Adaptive Allocation
Position sizing driven by clustering-based methods that dynamically group assets by statistical similarity. Weights adjust as market structure evolves.
0
Allocation updates per month
⬡
Deep Learning
Neural network models trained on high-dimensional feature spaces for cross-sectional return prediction. All models validated out of sample before live deployment.
0
Model features in production
01
Universe & Data
Large-cap equities screened for liquidity. Daily OHLCV, options chain, implied volatility surface, and alternative data ingested via automated pipeline.
02
Signal Generation
Volatility regime classifier (HMM-based), cross-sectional factor scores, and gradient-boosted ML model. Each signal z-scored and winsorised.
03
Portfolio Construction
Fractional Kelly position sizing with hard risk limits. Max 15% single-name. Dynamic hedging via index options in stress regimes.
04
Execution & Risk
Rebalanced weekly. Slippage modelled at 5bps. Stop-loss at -3% monthly drawdown. Full audit trail via prime broker.
05
Validation
Walk-forward cross-validation across 4+ years live. No lookahead bias. Out-of-sample Sharpe consistently above 1.0.