Free module · Systematic strategy development
Strategy Research, Backtesting & Robust Optimisation
Turn an idea into a controlled experiment, and keep parameter search from becoming a search for luck.
point-in-time data · purging · nested tuning · multiple trials · stress tests
The building blocks
A backtest is an experiment with many opportunities to accidentally help the strategy. Start with one decision timestamp and one untouched future period.
- Check what was knowable at each decision
- Separate fitting, tuning and final evaluation
- Account for search, dependence and deployment changes
Lessons in this module
- Write the experiment before the strategy
- Walk-forward validation, overlapping labels and purging
- Hyperparameter optimisation without an unrestricted search
- Multiple trials, false discoveries and selection diagnostics
- Dependent returns, block bootstrap and realistic stress tests
- Fine-tuning, retraining and the research-to-production decision
Practice and apply
- Align turnover and return — Weight changes from 0 to .5; next return is 1%; cost is 10 bps per unit turnover.
- Purge a future-dependent label — Label ends [96,98,101]; validation starts 100; gap 2; retain only end<98.
Work through the practice exercises · Quant development tools