Approach

Conviction is earned through falsification, not asserted.

ASQIOS treats research as an adversarial process. The default verdict on any idea is rejection; an edge has to survive a defined sequence of tests, in a fixed order of priorities, before it is believed.

The decision order

Priorities resolve lexicographically.

Order is not a preference — it is a rule. A lower priority is only ever considered once every higher priority is fully satisfied.

[ 1 ]

Shariah compliancePermissibility is the gate. A non-compliant strategy does not advance, regardless of its measured performance.

[ 2 ]

SurvivalAvoiding ruin precedes pursuing return. The system protects its own continuity before it optimises anything else.

[ 3 ]

Geometric growthCompounding over the long run — not single-period averages — is the objective once compliance and survival are secured.

[ 4 ]

RobustnessAn edge must hold across regimes and survive correction for multiple testing before it is trusted.

[ 5 ]

EfficiencyCost, turnover, and implementation friction are minimised last — after everything above is settled.

// Distrust acceptances. Trust rejections.

Validation gates

Verdict criteria are pre-registered before any data is touched.

A strategy moves through defined gates. Crucially, the criteria for each verdict are committed in advance — so the trial ledger functions as evidence, never as a story written after the fact.

G1
Candidate package

The mechanism, universe, and hypothesis are defined and documented.

DEFINED
G2
Shariah ruling

Qualified scholars rule on permissibility before research proceeds.

RULED
G3
Pre-registration

Minimum detectable effect, required sample, and verdict thresholds are fixed.

COMMITTED
G4
Reproducible backtest

Run through the reproduction harness and logged to the immutable trial ledger.

LOGGED
REJECTED NOT PROVEN REQUIRES R-2 DATA
Data honesty

Free data can reject. It can never accept.

ASQIOS currently researches on open data — regulatory filings, public end-of-day prices, and Shariah-universe proxies. The known contaminations of free data, such as survivorship bias and non-point-in-time fundamentals, almost all flatter long-only strategies.

That asymmetry is turned into a discipline rather than hidden. A strategy that fails on free data is genuinely failing. A strategy that succeeds on free data has proven nothing institutional — acceptance is structurally reserved for higher-grade data that the platform does not yet use.

The R-2A stance

Valid rejections are available now. Valid acceptances are gated behind point-in-time institutional data — and that gate stays closed until it is properly opened. The platform never lets a flattering result on weak data masquerade as a real one.

◦ FREE DATA → VALID REJECTION ONLY

See how the method is structured into a platform.