Quantitative Strategy Research & Backtesting

Clear learning outcomes.

• ✓ Hypotheses, variables and decision rules sits inside the module 'From Narrative to Falsifiable Hypothesis'. The analytical objective in this lesson is measurement discipline. Start by defining the object being measured, the timestamp at which it becomes observable, and the unit in which it will be compared. Do not move from an observation to a directional claim unless the rule explicitly defines that inference and the evidence supports it.
• ✓ Pre-registration and research questions sits inside the module 'From Narrative to Falsifiable Hypothesis'. The operational objective is repeatability. Another analyst using the same data, rule version and sequence should be able to reproduce the classification. Ambiguous cases are not forced into a preferred category; they are tagged unresolved and retained in the evidence set.
• ✓ Null models and baseline comparisons sits inside the module 'From Narrative to Falsifiable Hypothesis'. The review objective is robustness. The method is evaluated across ordinary, adverse and boundary cases. A result that appears only after hindsight selection, favorable fill assumptions or post-outcome rule changes is treated as weak evidence.
• ✓ OHLCV structure, timestamps and sessions sits inside the module 'Data Integrity and Time Alignment'. The analytical objective in this lesson is measurement discipline. Start by defining the object being measured, the timestamp at which it becomes observable, and the unit in which it will be compared. Do not move from an observation to a directional claim unless the rule explicitly defines that inference and the evidence supports it.
• ✓ Corporate actions, rolls and missing data sits inside the module 'Data Integrity and Time Alignment'. The operational objective is repeatability. Another analyst using the same data, rule version and sequence should be able to reproduce the classification. Ambiguous cases are not forced into a preferred category; they are tagged unresolved and retained in the evidence set.
• ✓ Lagging features to prevent leakage sits inside the module 'Data Integrity and Time Alignment'. The review objective is robustness. The method is evaluated across ordinary, adverse and boundary cases. A result that appears only after hindsight selection, favorable fill assumptions or post-outcome rule changes is treated as weak evidence.

Structured modules.

1. 01From Narrative to Falsifiable HypothesisLessons, practical work and knowledge checks
2. 02Data Integrity and Time AlignmentLessons, practical work and knowledge checks
3. 03Backtest Engine DesignLessons, practical work and knowledge checks
4. 04Transaction Costs and Fill ModelsLessons, practical work and knowledge checks
5. 05Performance Measurement and UncertaintyLessons, practical work and knowledge checks
6. 06Out-of-Sample and Walk-Forward TestingLessons, practical work and knowledge checks
7. 07Robustness, Sensitivity and Monte CarloLessons, practical work and knowledge checks
8. 08Research Governance and Final ReportLessons, practical work and knowledge checks