COT Crude Oil — Killed April 2026

Hypothesis: NYMEX WTI crude oil commercial hedger extreme short positioning (z below −2, rolling two-year weekly) is a contrarian buy signal. Analysis extended from 1986 to 2022.

Battery result: 0 of 4 tests passed. Specificity ratio: 1.11×.

Test 1 — Monte Carlo null: Hit rate is not statistically different from 50/50. The p-value does not reject the null at any horizon.

Test 2 — Blind era-split replication: The signal fails to replicate in the held-out era. This is a recurring pattern across COT crude research: results that appear robust in backtests collapse when the test window is blinded.

Test 3 — Specificity vs negative controls: Specificity of 1.11× means the signal performs marginally better than random negative controls — essentially noise. A genuine signal would show specificity ratios above 3× in this test framework. 1.11× is not distinguishable from random variation.

Test 4 — Mechanism sensitivity: The structural premise — that commercial hedgers in crude oil systematically over-hedge at price floors, creating a recoverable signal — is not supported. WTI commercials include large integrated oil majors whose hedging is tied to business operations and CapEx cycles, not sentiment extremes.

Failure mode: PURE NOISE. The 1.11× specificity figure is the most informative number here: even with a generous definition of “works better than chance,” the margin is trivial and does not survive blinding. The crude oil COT signal joins a documented pattern of COT-based systems that show apparent historical fit but fail formal specificity testing.

What survives: The broader energy market context remains relevant to other Observatory signals. The 2029–2031 chain incorporates food-energy price coupling via structural mechanisms unrelated to COT positioning.