Systematic Investment Research
We research markets
the way science
researches nature.
Heurys is a systematic investment research company. We develop hypotheses, subject them to multi-layered validation, and deploy only the models that prove statistically durable — with our own capital.
What we are
We are a technology company that conducts systematic investment research. We do not sell signals. We do not sell software. We build hypotheses, put them through rigorous verification, and use only what the evidence supports — measured, monitored, and held to the standard of a laboratory rather than a trading desk.
Principles
The discipline behind the returns is the product.
Research culture
Every idea begins as a question, not a conviction. We reward disproof as much as discovery, because a hypothesis that survives honest attempts to break it is worth far more than one that was never tested.
Data engineering
Signal is only as trustworthy as the data beneath it. We build clean, point-in-time datasets — free of survivorship and look-ahead bias — so that what we measure is what was actually knowable at the time.
Hypothesis methodology
We formalise each idea into a testable statement with a clear economic rationale. Structure comes before search, so that we are testing beliefs about markets — not fishing through noise.
Validation
Findings pass through multiple independent layers — out-of-sample windows, regime splits, cost and capacity modelling. Only models that remain statistically durable across all of them earn capital.
Risk management
Risk is a first-class design constraint, not an afterthought. Position sizing, correlation, drawdown and liquidity are modelled deliberately so that no single idea can define the outcome of the whole.
Edge health
Markets adapt, so edges decay. We continuously monitor the live behaviour of every model against its research baseline — an immune system that retires a strategy the moment its evidence weakens.
The method
From question to capital, one disciplined path.
Nothing reaches live capital by intuition alone. Each stage is a filter — and most ideas are meant to fail one of them. That is the point.
Hypothesis
An economic idea, stated precisely enough to be wrong.
Evidence
Point-in-time data, engineered to remove bias.
Validation
Multi-layered testing across samples and regimes.
Deployment
Sized by risk, deployed only when durable.
Monitoring
Live edge tracked against its research baseline.