Methodology

Transparent by design,
rigorous by default

We believe educational research is only valuable when you can verify and understand the process behind it. Here is exactly how our reports are produced.

Data Sources & Coverage

Price Data

End-of-day OHLCV data from major exchanges. Adjusted for splits, dividends, and corporate actions.

HKEX (Hong Kong)
NYSE / NASDAQ (US)
SSE / SZSE (China)
TSE (Japan)

Market Indices

Benchmark indices for performance comparison and regime classification.

Hang Seng Index
S&P 500
NASDAQ Composite
CSI 300

Alternative Data

Supplementary datasets for multi-factor analysis and regime detection.

Volatility Indices (VIX)
FX & Commodities
Bond Yields
Macro Indicators

Backtesting Engine

Architecture

Our engine processes historical data bar-by-bar in chronological order, ensuring no future information leaks into past decisions.

Event-Driven
Bar-by-bar processing, no look-ahead bias
Configurable
Custom transaction costs, slippage, position sizing
Reproducible
Same inputs always produce same outputs

Execution Model

Default Slippage 0.1% per trade
Commission Model Exchange-specific
Fill Assumption Next-bar open
Minimum History 5+ years required

Bias Controls & Safeguards

Backtesting is only useful if the results are honest. Here are the specific controls we apply to avoid the most common pitfalls.

Look-Ahead Bias

Using information that wouldn't have been available at the time of the trade decision.

Our Control

Strictly event-driven architecture. Signals generated using only data available at each bar's close.

Survivorship Bias

Only testing against stocks that still exist, ignoring delisted or bankrupt companies.

Our Control

Datasets include delisted securities. Index constituent lists are point-in-time, not current.

Overfitting

Tuning parameters to fit historical noise rather than genuine patterns.

Our Control

Minimal parameter tuning. Out-of-sample validation. Walk-forward testing where applicable.

Selection Bias

Only publishing strategies that performed well, hiding the failures.

Our Control

We publish both successful and unsuccessful strategies. Learning from failure is equally educational.

Metrics & Statistical Standards

Every report includes a standardized set of performance and risk metrics, calculated using industry-standard formulas.

Metric Description Why It Matters
Sharpe Ratio Risk-adjusted return vs. risk-free rate Higher = better return per unit of risk
Max Drawdown Largest peak-to-trough decline Worst-case loss scenario
Win Rate % of trades that are profitable Strategy consistency indicator
Sortino Ratio Return vs. downside deviation only Penalizes only harmful volatility
Calmar Ratio Ann. return / max drawdown Return efficiency vs. worst loss
Profit Factor Gross profits / gross losses >1 means profitable overall

Important Disclosure

All research published on Trading.hk is for educational and informational purposes only. Backtest results are historical simulations and do not guarantee future performance.

Our reports are not personalized investment advice. The same report is published to all subscribers. We do not recommend buying or selling any specific securities.

Always consult a licensed financial advisor before making investment decisions. Past performance is not indicative of future results.

See the methodology in action

Browse our report library to see how these principles translate into real educational research.