The work
Research notes, strategy write-ups, and the code behind them.
Everything here is written to the same standard: state the hypothesis, define the rule, price the risk, and show what would prove it wrong. Figures labeled illustrative are reconstructed to explain methodology — they are not live track records.
16 items / all work
01
ForgeAlpha
An autonomous multi-agent research platform: idea to validated, documented, paper-traded strategy — with every number produced by deterministic software, not a model.
System design
2026
02
Regime-aware risk allocation
Sizing by measured regime and a volatility budget, with a de-grossing ladder that acts before the drawdown.
Strategy
2026
03
Short-dated options, defined risk
A variance-premium process for 0–5 DTE index and large-cap options, sized by maximum loss.
Strategy
2026
04
How I underwrite a business
A repeatable process for equity research: quality, normalized earnings, embedded expectations, disconfirmation.
Equity research
2026
05
Engineered scorecards
Turning discretionary observations into a weighted, auditable composite signal that can be tested.
Concept
2026
06
Building in QuantConnect
Porting the research into LEAN: point-in-time data, survivorship, fill modeling, parameter stability.
Build log
2026
07
Time-series momentum, four asset classes
Sixteen ETFs, monthly, 7% volatility budget. 6.53% CAGR and 0.821 Sharpe over 2001–2026 against SPY’s 9.02% and 0.550, with a 14.3% drawdown. Walk-forward Sharpe 0.736.
Strategy
2026
08
Volatility-managed equity exposure
The S&P 500 held at a constant 8% risk budget instead of a constant weight. 0.829 Sharpe and a 22.6% drawdown against 0.550 and 55.2% for buy-and-hold.
Strategy
2026
09
Breakout trend with a barrier stop
100-day Donchian entries, 30-day exits, plus a volatility-scaled stop. Alpha of 7.30% a year against the S&P 500 with a 0.04 beta; 9.32% CAGR in the 2016–2026 holdout.
Strategy
2026
10
Cross-sectional momentum in sectors
Twelve-minus-one momentum across nine US sectors and seven country/real-asset ETFs, top five, inverse-volatility weighted. 7.64% CAGR; 2008 is the honest failure case.
Strategy
2026
11
The triple barrier, tested
Meta-labelling, barrier exits and barrier stops applied to four books. The classifier has no out-of-sample skill (AUC 0.51); the stop leg is the only part that earns its keep. A negative result, documented.
Research
2026
12
Micron Technology — HOLD, $940
The most profitable year in memory history at 6.6× sales, with a forward P/E of 3.9× that assumes the cycle does not turn. We are not sellers; we are not adding either.
Equity research
2026
13
UnitedHealth — BUY, $470
Operating margin fell from 8.1% to 4.2% in one year. A return to a mid-6% margin on $460B of revenue is worth $23 of EPS; at 20× that is $470.
Equity research
2026
14
Nike — SELL, $33
Three years of turnaround without a margin. Revenue is flat, free cash flow has fallen 67%, and the 4.3% dividend is not covered by it.
Equity research
2026
15
ExxonMobil — HOLD, $175
The best cash-return story in integrated oil, priced like it after a 43% year. A 3.5% free-cash-flow yield and a 2.6% dividend; we would buy at $145–150.
Equity research
2026
16
Academic archive
Fourteen college work samples — equity research, stock pitches, industry studies and group presentations from the University of Alabama, 2021–2025.
Academic
2025
No work in this category yet.
On performance figures
Charts and tables on these pages are illustrative reconstructions used to explain a method — the shape of a drawdown, the geometry of a payoff, the sensitivity of a valuation. They are not audited results and should not be read as a track record. Live research, full backtests and code are available on request.