An adversarial stance toward our own ideas.
Every paper claims an edge somewhere. Most evaporate under honest cost accounting, a single regime change, or a different random seed. Our working assumption is that every candidate signal is noise until proven otherwise — and we spend more effort trying to break an idea than to defend it.
Promotion to live capital requires surviving both a chronological out-of-sample tail and a cascade-strict OOS regime — at full cost, on raw forward returns, on data the model never touched. The few strategies that clear this bar are deployed with real money, sized small, and held to the same metrics they earned in research. Drift from those metrics is the exit condition, not the review trigger.