File of record
Forecasts
Time-bounded claims frozen at issuance. A warning’s live rung can move during the window; the scored card does not. Resolution is binary on the warned outcome inside the window.
Kyle 2026-09-23: W7 is the first real-world test of GUARDIAN STAR predictive capabilities. A miss at window close is a process failure, not “likely misses sometimes.” Miss → retool (after-action and framework changes in the same turn), not a log line and a lowered rung.
Open a card when a warning is issued at likely or higher with a
window of 30 days or less, or when Kyle marks a call as scored.
After F1 (2026-10-24): do not open a card on a straight adoption of FO
“likely within N days” when the cabinet is split or the implementing
secretary is on the record against the act. Haircut first (see
references/scored-forecasts.md); score only if the GS rung after the
haircut is still likely or higher.
Open
(none)
Resolved
F1 — W7 US diesel export embargo inside 30 days
Frozen: 2026-09-23
Claim: The US administration imposes a full or partial ban, embargo,
or equivalent export restriction on diesel fuel by proclamation, IEEPA /
DOE / Commerce order, or a completed floor vote (W7 warned outcome /
W7-I4) on or before 2026-10-23.
Issued likelihood: likely, moderate confidence
Source of the call: FO Early Warning 23 Sep (R.C.), adopted as W7
Window: 2026-09-23 through 2026-10-23, America/Los_Angeles (FO 23 Sep:
“likely within the next 30 days”; WARNINGS.md window).
Hit: W7-I4 fires inside the window. Resolve immediately; do not wait
for close. No retool from this test.
Miss: 2026-10-23 ends without I4. Congressional calls, cabinet
studies, the president’s urging, voluntary refiner cuts, SPR / heating-oil
actions, and a gasoline-only restriction are not a hit.
On miss (Kyle 2026-09-23): retool prediction frameworks. Apply W7-D3.
Do not treat as within-error-bar. Write the after-action and the
framework changes in the same turn (skill: scored-forecast resolution).
Status: MISS
Resolved: 2026-10-24 (window closed 2026-10-23 23:59:59 PDT; clock
2026-10-24 07:00 PDT). Kyle 2026-09-23 scored-test rule.
What was checked (this run):
- Federal Register API, presidential documents, publication 2026-09-20
through 2026-10-24: term diesel export ban OR embargo → count 0.
Presidential diesel document in the window: EO 14435 “Emergency Tax
Relief on Diesel Fuel” only (FR 2026-20855, published 2026-10-09;
White House, 5 Oct, https://www.whitehouse.gov/presidential-actions/2026/10/emergency-tax-relief-on-diesel-fuel ) — dyed-diesel tax/penalty
relief, not an export restriction.
- FR API diesel + export, same dates: LNG-plant authorizations
(2026-10-20), SAFE Vehicles Rule III (2026-09-30), Safe Explosives Act
(2026-09-25). None is a diesel-export restriction.
- White House proclamations/actions indexes this session: JS shells;
no diesel-export-ban instrument retrieved.
- Congress.gov: H.R. 10423 (temporary prohibition on diesel-fuel export
through 31 Dec 2026) and H.R. 10422 (Diesel Price Reduction Act of 2026)
— both introduced 16 Sep 2026, referred to House Foreign Affairs,
status Introduced; no floor vote (https://www.congress.gov/bill/119th-congress/house-bill/10423 ; https://www.congress.gov/bill/119th-congress/house-bill/10422 ).
- Reuters, Bose/Psaledakis/Shakil, 2 Oct: Trump “will not authorize an
export ban on diesel” after the G7 reserve release (https://www.reuters.com/business/energy/trump-not-going-be-doing-diesel-export-ban-2026-10-02/ ).
- Reuters, Renshaw/Douglas, 23 Oct: farm aid, dyed-diesel EO, Russian-fuel
GL, European stock release; “a threat to restrict US diesel exports”
as past pressure, not an imposed restriction; analysts “little diesel
price relief before midterms” (https://www.reuters.com/business/energy/trumps-diesel-price-relief-efforts-grow-costlier-election-nears-2026-10-23/ ).
- INDICATORS.md: no W7-I4 tag this window. Weekend sweep 24 Oct already
logged I4 unfired through close and left D3/F1 to this job.
Instrument: none. W7-I4 unfired through 2026-10-23 PDT.
After-action (F1 miss — Kyle 2026-09-23 retool)
Assumption that failed: midterm diesel-price pressure would outweigh Wright/Burgum opposition and refiner pushback inside 30 days, so an executive restriction was likely.
FO’s R.C. versus what occurred: - 23 Sep: “a diesel export embargo is likely within the next 30 days” (adopted; W7 opened at likely; F1 frozen here). - 24 Sep: Wright’s “not considering a ‘flat ban’” read as imminent non-flat (GS likely → very likely). - 6 Oct: “almost certainly exhaust other options before resorting to a diesel export embargo” after the 5 Oct dyed-diesel EO (GS very likely → likely). - 14 Oct: “A U.S. diesel export embargo is increasingly unlikely ahead of the 2026 midterms” after DPA/Delek (GS likely → unlikely). Live rung was already off likely before close. - Occurred: no I4. Substitutes executed instead — G7/IEA stock release (2 Oct), dyed-diesel EO 14435 (5 Oct), OFAC GL 135 Russian-origin diesel import authorization (9 Oct), DPA/Delek Tyler expansion (13 Oct), NEHHOR heating-oil exchange (21 Oct). AAA national diesel 24 Oct $6.3374 versus series high $6.5276 on 22 Sep (https://gasprices.aaa.com ).
Should GS have opened at likely given a split cabinet? No. On 23 Sep the implementing secretary (Wright) was already public that a ban “definitely doesn’t work”; Burgum had said an export ban would not lower prices; Bessent was “examining”; Trump had urged staff. That is a split, not a decision path. Opening at likely was a straight adoption of FO’s time-bounded Outlook. The call should have opened at roughly even chance (FO wording still quoted; GS rung haircut one step for named opposition from the implementing secretary). F1 should not have been scored: after the haircut the GS rung would not have been likely.
What would have discriminated earlier: - Wright 23 Sep public rejection plus the White House denial of Politico’s 90-day blanket plan — held as not D1 (denial of a flat/blanket plan, not of any restriction). Correct on D1’s wording; insufficient as a rung haircut. - Trump, South Lawn, 2 Oct: “will not authorize an export ban on diesel” (Reuters). Held as a Wright-class hedge on “the export ban” / flat ban, not D1, because FO 24 Sep imminent was unretracted. That over-parsed a dated presidential “no” to keep FO’s time-bounded likely. D1’s wording (“a dated ‘no’ from the president”) was met. The live rung should have dropped on 2 Oct without waiting for FO 6 Oct / 14 Oct.
Retool applied this turn: analysis skill, FO ingest reference, and this register. FO still wins on wording until Kyle says otherwise. Time-bounded likely adoption is no longer automatic when the cabinet is split.