Alpha Metallurgical Resources filed Q1 2026 10-Q on May 8. The defined kill trigger on the long position entered its trigger zone (capture 54.73% vs 55.00% threshold, by 30bps), and the 10-Q disclosed an ASC 360 impairment test on Aracoma and Kingston/Mammoth complexes for the first time we have it on record at these complexes. This memo resolves the long. It does not re-derive a short — that requires fresh evidentiary work, including an open research gap that this filing did not close.

What the filing says

Q1 capture at the trigger boundary. Coking-only Met segment realization $128.40/mt (10-Q MD&A, Note 4) divided by AMR-disclosed PLV Q1 2026 average $234.67/mt = 54.73%. Two alternative builds — Met-segment-blended ($124.39 / $234.67 = 53.0%) and Met-coking-with-byproduct — give 53.0% to 55.5%. All defensible builds sit at or below the 55.0% threshold. The trigger as defined was reached. The margin is small; treat this as the trigger zone, not a clean break.

ASC 360 impairment test. Note 1, Asset Impairment: "As of March 31, 2026, due to declines in pricing forecasts for U.S. High-Vol. metallurgical coal products, the Aracoma and Kingston/Mammoth mining complexes were tested for impairment." Both passed undiscounted-cash-flow recoverability — no charge. The test was not disclosed in the preliminary 8-K (2026-04-24); this is 10-Q-only disclosure. The new information is the disclosed test, not the test mechanics — companies run step-1 tests whenever a triggering event flag hits. Management's internal HVA pricing forecast has deteriorated enough that this complex pair tripped the flag.

Other items. Q1 FCF -$24.6M before buybacks ($29M operating cash, $40.7M capex, $13.4M DTA contribution). Black lung quarterly periodic cost $4.034M = $16.1M annualized vs $12.2M FY2025 (+32%). Customer concentration: India + Turkey each >10% of revenues; top 10 = 82% (vs 73% Q1 2025). Kingston Wildcat first commercial tons slipped from Q1 to Q2 2026. 45X credit Q1 $7.2M = $28.8M annualized — low end of $30-50M guidance range, ≈25% of current EBITDA run-rate. Buyback $22.9M Q1; $343.8M authorization remaining; no 10b5-1 plan.

Cohort cross-check is suggestive but conditional

CNR Q1 2026 10-Q (filed 2026-05-07): zero matches on long-lived asset impairment / recoverability / pricing forecast language. HCC Q1 2026 10-Q (filed 2026-04-30): zero long-lived asset impairment matches. METC Q1 2026 not yet filed (expected mid-May).

The conclusion "AMR is the only producer in this cohort to disclose the impairment test in Q1 2026 10-Q" is supported. The inference "AMR's internal HVA forecast is uniquely deteriorated" is not yet supported. CNR completed the Consol-Arch merger August 2024; fresh-start fair-value accounting writes PP&E to current carrying values, which structurally reduces the probability that recoverability tests trip regardless of internal forecast. We have not yet reviewed CNR's PP&E footnote per equivalent reserve ton vs AMR's. The cohort silence may be an accounting artifact at CNR specifically. Until that disambiguation is closed, the AMR-idio bear reading should be held conditionally, not as a settled finding. This is the load-bearing item.

What the market thinks

AMR closed $184.11, down 4.82% on the print, down 8.5% from initiation (2026-03-19). Sell-side mean target $196, zero buys / three holds; B. Riley downgraded Buy→Neutral Feb 4 at $203. IV rank 43rd percentile, P/C 0.71 — directional bullish positioning lingers in options. Short interest 19.1%. The market is pricing a slight recovery tilt; ≈0% expected return is implied by current price across the scenario distribution.

Why the gap exists (such as it is)

Three reasons, each modest. (1) The impairment test disclosure is 10-Q-only and buried in Note 1. (2) The cohort comparison requires three 10-Qs read within a nine-day window for the same disclosure type — sell-side covers each name in isolation. (3) AMR has more realized HVA capture compression than HCC (54.7% vs 70.1%), though capture conflates HVA loading with mix differences (HCC's Mine 4/7 LV cushion). The directional claim — that AMR has more HVA exposure than HCC — is supported by regression evidence in worldview. The pair-trade construction (long HCC / short AMR) follows from this directional reading; the gross EV (+2.3% over 6 months) does not survive borrow cost (≈4% on AMR over 6 months) at current prices.

Engaging the contradiction: Courtis insider buying

W.B. Courtis purchased $42.5M+ in AMR equity at $160-188 levels through 2025. Three readings, all defensible: (a) Courtis is on a multi-year cyclical horizon longer than the kill trigger; (b) Courtis is wrong; (c) the buying predates the HVA forecast deterioration captured in the Q1 10-Q. We don't observe his trading since the most recent Form 4. This memo does not reconcile the contradiction. It registers it. The kill-trigger thesis is shorter-duration than Courtis's apparent horizon; both can be right at different time scales. A short re-derivation must engage with this more directly.

Risks (ranked by impact)

1. F2 normalizes from the supply side. Queensland force majeure status unresolved. If Australian PLV breaks below $200 on supply normalization, AMR capture lifts mechanically without HVA recovery being required.
2. CNR fresh-start accounting explains cohort silence. If CNR carrying values are <50% per equivalent reserve ton vs AMR, the structural finding compresses materially.
3. Courtis bull case is right. Multi-year cyclical recovery in seaborne met coal absorbs Q1 weakness as transient.
4. 45X tax credit reduction or repeal. $28.8M annualized = ≈25% of EBITDA run-rate. Modest probability, large impact if it materializes.
5. Sectoral met-coal risk-on. China stimulus or India BF additions lift cohort regardless of relative positioning.

Catalysts

• METC Q1 2026 10-Q (~mid-May). Completes the 4-name cohort silence test. (P=0.75 silent on impairment)
• AMR Q2 2026 10-Q (~Aug 7-14). Tests Q2 capture (P=0.70 capture <58%) and impairment escalation (P=0.78 no charge taken).
• Kingston Wildcat first commercial tons Q2 2026. (P=0.65)
• CNR PP&E footnote review. Internal task — resolves the load-bearing gap in the cohort analysis.

What would change our mind

(1) METC discloses its own first 10-Q-disclosed HVA impairment test — cohort split, AMR-idio compresses to sectoral. (2) CNR PP&E review shows carrying values <50% per equivalent reserve ton vs AMR — fresh-start accounting explains silence, structural finding evaporates. (3) HCC takes a long-lived asset impairment charge in Q2 — Blue Creek cushion thesis breaks. (4) PLV-HVA spread closes below $40 by Q3 2026 (currently $72) — F2 normalizes organically; the pair-trade construction inverts.

Position management

The Q1 print satisfies the kill trigger criterion as defined six months prior. Forward 6-month probability-weighted EV is approximately -2%. Fresh short or pair construction is not warranted by this memo — it requires resolving the CNR fresh-start gap AND one of: AMR rally to $200-210, IV rank below 25%, METC silent, sell-side capitulation post-Q2.

The thesis was revised on Feb 23 (long-only spread normalization at narrowed alpha) and again here on May 9. Two revisions in eleven weeks is close to the line of fitting triggers to data. The current revision is supported by the kill trigger as defined six months prior, not a new construction.

Evidence