We trade enforced structural relationships. We publish our methods.

Axoquant looks for relationships that a mechanism obliges to hold — rebalance mandates, defended currency bands, share-class identity, contract convergence — across currencies, equities, futures, and crypto. Every candidate is tested against matched control ensembles before capital touches it: an idea must beat a placebo built to null exactly the mechanism it claims.

Most candidates do not survive. That is the point.

FIG. 1 — THE SPECIFICITY GATE

Of 2,029 strategies authored to date, 265 accumulated enough independent test windows to face the specificity gate — forward replay against matched control ensembles that null the claimed mechanism. Each candidate entering the gate was itself the surviving parameterisation of a several-thousand-trial optimisation. 19 survived (Benjamini–Hochberg false-discovery control at pre-registered thresholds, consistency required across independent windows). The remainder were refuted and archived.

exchange feeds
4 asset classes, daily bars

3.8M+

parameterisations backtested
across 16,577 optimisation runs · recorded floor

2,010

strategies refuted
and archived with their evidence

Published against ourselves

Recently refuted

Carry as alpha G10 carry basket passed survivability (Sharpe 1.6, stable across windows) and was refused at the specificity gate: p = 0.62 against a beta-adjusted null. It is the carry factor, not an edge. We publish it as one.
Cross-sectional mean reversion, US equities Negative after real per-share costs in every test window. The fee floor is the finding.
Our own first equity flagship Voided 48 hours after passing, by our own fills-level audit — the instrument ledger showed the backtest had traded a different venue than its label claimed. The gate that caught it is now a standing pre-flight check.