Methodology
The systematic strategy workflow, linked to the research papers.
This page is the table of contents for how the portfolio process is built: collect data, create features, fit the model, construct the portfolio, then prepare implementation. Each step points to the dedicated insight paper that explains it.
Map the whole process
How to build a systematic strategy
Start here if you want the plain-English overview of how a systematic strategy moves from idea to live portfolio.
Collect clean data
Systematic strategy step 1: data collection
Universe selection, point-in-time data, missing values, survivorship bias, and why data quality matters more than model complexity.
Engineer features
Systematic strategy step 2: feature engineering
Turn raw prices, liquidity, volatility, fundamentals, and horizons into normalized inputs the model can compare across names.
Fit and validate the model
Systematic strategy step 3: fit and validation
Estimate whether signals predict returns without leakage, overfitting, or mistaking a lucky backtest for a durable edge.
Construct the portfolio
Systematic strategy step 4: portfolio optimization
Translate forecasts into weights while controlling risk, costs, turnover, liquidity, sector exposure, and leverage.
Build it with AI
My first quant strategy with AI
A practical end-to-end blueprint for using AI to assemble data, features, beta fits, risk terms, optimization, metrics, and automation.
Prepare broker execution
How to turn target positions into broker trades
Turn target positions into broker-ready files, pre-trade checks, paper-trading workflows, and reconciliation steps.
Signal and risk research
The evidence behind the workflow
These papers are not separate from the methodology. They explain the signal ideas, risk controls, and market effects that feed into feature engineering, model validation, and portfolio construction.
Systematic momentum: what to look at
Why the model starts with relative trend and cross-sectional ranking.
Short-term reversal: real but hard to trade
A useful but fragile signal that needs careful execution and cost control.
Why diversification kills momentum
Why the strongest momentum names often cluster by industry or sector.
Sector risk: the noise hedge funds remove
Why neutralizing broad sector noise can make the prediction problem cleaner.
US sector seasonality (XL ETFs)
Calendar effects as a simple example of measuring unconditional return patterns.
Fundamental inputs
Company-quality papers used in feature design
These explain the fundamental dimensions that can become model features or watchlist rationale: valuation, growth, profitability, balance-sheet health, and shareholder returns.
How to assess company quality
A beginner-friendly checklist for judging a company: valuation, growth, profitability, balance-sheet health, shareholder returns, and management.
Company quality step 1: valuation
How to think about price versus value, why cheap can be dangerous, and why great companies can still be bad investments at the wrong price.
Company quality step 2: growth
How to separate healthy growth from low-quality growth by looking at revenue, earnings, cash flow, expectations, and sustainability.
Company quality step 3: profitability
Why margins, returns on capital, cash conversion, and industry context matter when judging whether a business is truly high quality.
Company quality step 4: financial health
How to read balance-sheet risk, debt, cash flow coverage, and refinancing pressure before a business problem becomes a survival problem.
Company quality step 5: shareholder returns
How dividends, buybacks, acquisitions, reinvestment, and management incentives reveal whether capital is being used wisely.