Skip to content
Salarium Classical Profile MarkSALARIUMOPEN QUANTITATIVE RESEARCH
1.0.0-RC1/RELEASE CANDIDATE
Salarium Classical Profile MarkSALARIUMOPEN QUANTITATIVE RESEARCH

A stock model, four paper portfolios, and the results and rules behind them.

For research and education. No live trading or personal investment advice.

NAVIGATE

Four paper portfoliosResearch candidatesResearch recordMethod and limitsDisclosures

EVIDENCE

Four paper accountsForward paper snapshotRelease snapshot20D release rankingsCandidate snapshot

PROJECT

GitHub repositoryModel cardRelease notes
COMMIT 7c730e84c409MODEL RELEASE SNAPSHOT AUG 11, 2026, 11:14 PM UTC© 2026 NIALL GILLEN · MIT LICENSE

MODEL COURTROOM / CASE 001

Oil cut the drawdown.
It failed the recovery test.

In this historical test, a 20% oil sleeve reduced the largest loss but missed the required improvement in recovery time. It remains experimental.

Not promoted8 / 9 recorded gates passed

THE CASE FOR

The portfolio fell less.

Maximum drawdown improved by 15.8 percentage points against the cash-yield comparator. Net Sharpe improved by 0.162.

Drawdown improved in 6 evaluated years. The candidate also cleared the published holdout and cost-stress gates.

THE CASE AGAINST

It did not recover much faster.

The longest recovery shortened by 3.6%, below the protocol’s 20% requirement. Annualized return drag was 2.1 percentage points.

Oil is not a general crisis hedge. ETF proxy behavior in inflationary conditions does not establish protection during deflationary or liquidity shocks.

The recorded acceptance tests

Maximum drawdownPASS
Expected shortfallPASS
Recovery timeFAIL
Return dragPASS
Sharpe improvementPASS
Yearly consistencyPASS
Holdout drawdownPASS
Holdout SharpePASS
Transaction-cost stressPASS

Verdict computed from all nine published gate booleans. Missing or non-true gates fail closed. Clearing gates does not automatically change the release. Arguments are editorial summaries of the stored research, not independent AI testimony.

Timing and evidentiary limits

The artifact claims a frozen protocol, but its generated timestamp (2026-08-12T02:58:36.700868+00:00) precedes its stated freeze time (2026-08-12T03:00:00Z). These timestamps alone do not independently establish preregistration.

Integrated evidence is historical simulation across 139 rebalances. ETF proxies are not contract-level futures execution. No live hedge was deployed.

Source commit: f9f5d61dd621ed4eb474df4c1cd1b82c1a2b689c
Configuration SHA-256: 04e28d4b32a694592d780e635bbed98cbc1d73cf01b9f871d3e426eb05ab36ef

Read the complete case evidence ↗
Download the research bundle ↓