Nothing Enters the Portfolio
by Accident.

Broad parameter spaces are explored to identify regions of stability rather than isolated combinations producing exceptional historical results.

Candidates are tested on historical data excluded from their development.

Shortlisted systems are rerun with every-tick modelling and inspected for data quality.

Chronological trade histories are combined and evaluated for correlation, concentration and drawdown interaction.

Relative strategy exposure is set at portfolio level. User-selected multipliers change risk from this reference point.

New execution and behaviour are compared with research assumptions — not presented as proof of future performance.