Research
A laboratory,
not a trading desk.
Our edge is not any single model. It is a culture — a way of asking questions about markets and refusing to believe the answers until the evidence is overwhelming.
01 — Culture
Most of what looks like a pattern in financial data is noise wearing a costume. The central task of research is to tell the two apart — reliably, repeatedly, and without fooling ourselves.
So we organise around the scientific method rather than around markets or products. Ideas are hypotheses. Backtests are experiments. Live performance is the final, unforgiving referee. The people who work here are researchers first, and the culture protects that above all else.
We measure ourselves not by how many strategies we launch, but by how ruthlessly we retire the ones that no longer earn their place.
02 — Tenets
Four commitments we do not compromise.
Disproof is progress
A result that fails to replicate has told us something true about the world. We design experiments to break our own ideas first, so the market does not have to.
Economic rationale first
Before a single backtest, an idea must explain why an edge should exist — who is on the other side, and why they persist. Statistics without a mechanism is coincidence.
Point-in-time truth
Our datasets reconstruct exactly what was knowable at each moment in history. No revisions, no survivors, no hindsight leaking backwards into the past.
Independence of tests
Validation layers are designed to fail independently. Passing one says little; surviving all of them, across regimes we did not choose, is the bar.
“The market is the most honest reviewer there is. Our job is simply to make sure our work is ready for it — and to be honest with ourselves long before it is.”