Edge Lab is where we publish our trading-system research — the analysis, the development, and the validation techniques behind it. Each study takes a trading idea and tests it the way Edge S2 tests everything: methodically, against a baseline, with the method shown in full so a reader can see how it was measured and where it stands.
Edge Lab runs every system through a staged funnel of gates — cheap, event-level checks first, then the expensive portfolio-level work. The discipline is consistent across studies. See the published studies ↓
| Survivorship-bias-free universes | Point-in-time membership that includes the names later delisted or acquired, on split- and dividend-adjusted prices. |
| The benchmark that matters, first | Every long-only equity study leads with buying and holding the index over the same window. |
| A random baseline threaded through every number | We replace only the signal with a seeded random pick through the identical engine, so each result is reported as edge versus random. |
| Robustness and cost | Out-of-sample holdouts or parameter sweeps for fitted systems, and sensitivity to turnover and trading costs. |
| Reproducibility | Each published study is backed by a package that re-runs to the numbers on the page. |
We stay neutral on outcomes. A study states what was measured, in what window, against what baseline, and lets the numbers carry the conclusion. Open questions are named as the next tests rather than dressed up or hidden. Many ideas do not clear every gate — testing a large number of candidates to find the few that hold up is the work, and a non-advancing result is a finding worth publishing.
Live now at /lab/pub
Edge S2 is currently in private beta. To follow the research or ask about the platform, get in touch.
Browse the Research