Fit

Fit

Does not fit the framework (P1 not met)

EQT clears the universe and trips no exclusion, but the year-10 durability gate — the one pure gate — does not clear, so the overall read is Does not fit the framework. Confidence is medium: the jury agreed across families on every gate criterion, and the basis the tally records is that probability spreads were at most 0.25 or one non-load-bearing criterion was contested. No flags are raised — not watchlist-only, no exclusion hits, no China-sensitivity flag, and no prior-driven-risk flag.

Universe and exclusions — unsoftened

EQT passes both universe lines cleanly. It is US-domiciled common stock on the NYSE (ticker EQT, SEC CIK 33213) — not an ADR and not a Chinese issuer — so the geography line holds; the identity rests on the SEC-sourced record because live web re-verification was unavailable this run. Market capitalization is about $32.8B (615.717M diluted shares [1] at the 2026-07-31 close of $53.29), clearing the $10B floor by 3.3x. Here is the plain-words point on that pass: the $32.8B is not organic scaling — it sits on roughly $8.6B of gas sales because the share count more than tripled from 167M (2016) to 615.7M (2025) [2]. The universe test is about size, and on size it passes.

No exclusion hit fires. The checks and what each found:

• Auto OEM (X1) — clean. EQT is a natural-gas production and midstream company; it manufactures no vehicles [3].
• Promotion pattern (X2) — checked, not a hit. Both prongs must be evidenced. On delivery, most resolved commitments were met or beaten (deleveraging from ~$13.7B toward $5B; $4.7B of asset sales closed a year early), with one miss — net debt landing ~$7.7B versus a ~$7B target, to fund the Olympus acquisition. Founder-CEO Toby Rice holds a nine-figure stake (2,454,427 shares plus 1,000,000 options), but all 17 directors and officers together hold under 1% of the class [4]. The counter-fact in the same breath: escalating language such as "transformation now complete," paired with sub-1% insider ownership and an ever-rising share count, is a real governance caution [5] — but delivery held, so the exclusion does not trigger.
• Structural decline (X3) — checked, absent. The disqualifier (three consecutive years of high-single-digit revenue decline) does not fire, and production volume rose 2,016 → 2,228 → 2,382 Bcfe (FY2023–FY2025). The counter-fact: revenue is violently cyclical (down 58% in FY2023), but the declines are price-driven and non-consecutive, not a structural bleed [6].
• Consensus-saturated story (X4) — clean. EQT trades at ~6.9x EV/EBITDA, ~16x earnings and ~3.8x sales — cyclical-commodity multiples, the opposite of an extreme multiple-to-sales darling. EV of $40.5B ($32.8B market cap + $7.69B net debt) over FY2025 EBITDA of $5.85B (operating income $3,249.6M [7] plus depreciation, depletion and amortization of $2,600.4M) is 6.9x. The counter-fact: an "AI data centers and LNG" demand narrative is attached to the name and appears in EQT's own strategy language [8], which bears watching for sentiment even though the current multiple does not reflect it.
• China dependence (S1) — clean, no flag. All reserves, wells and pipelines are in PA/WV/OH [9] and revenue is domestic wholesale gas; 94% of employees sit in PA/WV/TX/OH [10]. LNG exports give indirect global price exposure but no direct China revenue or asset concentration.

Pattern match

EQT fits none of the four recognition setups cleanly. It is not a large bank at the bottom of a credit cycle; it is not a high-dividend-yield case (the ~1.2% dividend is immaterial); it is not a healthcare/insurance forecasting error; and it is not a quality tech monopoly on a fear dip. The nearest adjacency is the fourth setup — a large, essential business on a fear-driven dip — but the two checks that setup demands both fail here: there was no capitulation (the fall carried no volume spike) and there was no discrete adverse event (see the Dislocation treatment). What EQT actually is, in framework terms, is a commodity cyclical whose drawdown tracked a softening gas strip rather than a repricing of a temporary earnings anchor — closer to the "cheap on FCF yield" trap the framework is built to screen out than to any of the four setups it hunts.

The pillar ledger

Year-10 durability (P1) — the gate — not met

This is the decisive point. The gate asks whether year-10 revenue and adjusted free cash flow will both be higher than today's with very high conviction. The volume leg is credible: 28,046 Bcfe of proved reserves (93% Marcellus), roughly 30 years of inventory, and production rising three years running [11]. The price leg is where the gate fails. Revenue equals that volume times a commodity price EQT is a price-taker on and states it is "unable to predict" [12]; it "typically receive[s] market-based pricing" at a structurally negative Appalachian basis [13]. That price ranged $2.65–$9.86/MMBtu in 2025 alone and drove revenue from $12,114M in FY2022 [14] to $5,045M in FY2023, a 58% one-year fall [15]. There is also no market-structure moat: EQT calls itself the "only large-scale, integrated" US gas producer [16], yet at ~6% of US output it is a price-taker in a fragmented field, and Expand Energy claims the "largest independent" title outright [17]. The strongest surviving counter-fact sits in the same treatment: production is rising, 4.5 MTPA of LNG offtake is signed for 2030–2031, and integrated midstream defended a record $2.84B FCF in 2025 — so if gas merely holds near mid-cycle, both legs clear comfortably [18]. The gate is binary by construction, and genuine doubt on the price leg resolves it to not met (probability 0.53, spread 0.10, cross-family agreement). Full treatment: Durability. Reference line: revenue and adjusted FCF both higher in 2036 with very high conviction — the position is short of it on the price leg.

FCF consistency (P2) — not met

The framework's stability metric is adjusted FCF's rolling 5-year average; it is not computable here because stock-based compensation is missing for every year 2016–2025 in the feed. Reported FCF is genuinely unpredictable rather than a stable annuity: it swung from −$23M (FY2018) to $2,838M (FY2025) [19]. The counter-fact in the same breath: the reported rolling 5-year average does trend up ($184M → $1,449M across FY2020–FY2025), and the single negative year coincided with $3.0B of peak growth capex, not an operating failure. Full treatment: Durability.

Dislocation and yield (P3) — event and capitulation not met; yield short of bar; forward path met

P3a (identifiable event) is not met: the 28% fall from the 25 March 2026 all-time high carried no guidance cut, miss, or regulatory shock — Q1 2026 was a record $1.83B FCF and an 11% EPS beat [20] — and instead tracked a softer gas strip ($2.89/MMBtu in Q2) [21]. P3b (capitulation) is not met: peak 20-day average volume through the fall was 0.98x the pre-peak median, and the busiest single session reached only 1.64x — orderly selling, not a wash-out. P3c (yield vs bar) is not met: on adjusted FCF (reported FCF − SBC − 5-year average cash acquisitions), EQT yields 5.6% on FY2025 and 1.6% on the three-year average against the 10% default line — 440 bps and 840 bps short. The FY2025 arithmetic: reported $2,837.5M − SBC $60.8M − 5-year-average acquisitions $943.1M = $1,833.7M, over $32,811.6M = 5.59% [22]. The bar is the 10% moderate line because net debt/EBITDA is 1.31x ($7.69B / $5.85B) [23], not the softer 8–9% fortress line where the 8.6% reported yield would nearly clear. P3d (forward path) is met (probability 0.56, spread 0.05): consensus reported FCF yield reaches 11.7% by FY2028 and the adjusted basis crosses ~10.7% around then as the 2021–2024 acquisition dock ages off the trailing average — but only if EQT stops making large cash acquisitions, which it has not (Olympus closed July 2025) [24]. Full treatments: Dislocation and Yield. Note that adjusted FCF, its yield, balance-sheet class and yield baseline all returned not_computable in the deterministic feature file (SBC missing every year, EBITDA missing FY2025); the figures above are reconstructed from the filed 10-K cash-flow statements and flagged as such.

Balance sheet and self-help (P4) — outlast met; repurchase engine not met

P4a (outlast + allocation headroom) is met: EQT is investment-grade with a laddered senior-note schedule (~$7.5B total, largest single year $1.65B in 2029) [25], a $3.5B undrawn revolver [26], and net debt cut toward a $5.0B target [27]. The vote split here is worth stating: three seats read this met, one not met, so cross-family agreement is absent — the counter-fact is that in FY2024–FY2025 allocation went entirely to debt paydown and asset sales with zero buybacks, the forced-toward-debt pattern the framework flags. P4b (repurchase engine) is not met: the engine runs backwards. Shares rose 3.7x from 166.978M (2016) to 615.717M (2025) on serial all-stock acquisitions, while cumulative buybacks since 2021 were just $622M and zero in both FY2024 and FY2025 [28]. The counter-fact: management now frames aggressive countercyclical buybacks as its next lever and is willing to hold "up to a few billion dollars" of cash for them [29], but the framework rewards a demonstrated habit, not a stated one, and the levered exception (≥25% yield plus a multi-year share-count reduction) fails on both legs. P4c (dividend cover) is not applicable: the ~1.2% dividend ($0.66 annualized, covered ~7x by FY2025 FCF) is not part of the return case [30]. Full treatment: Self-Help.

Diagnosis (P5) — met (temporary lean), but a live permanent tail

The adversarial trial put the probability the impairment is temporary at 0.71 (mean 0.66, spread 0.16, per-judge 0.56–0.72, order-stability gap 0.07), recorded as not contested. The driver that broke is simply the realized gas price ($2.89/MMBtu in Q2 2026), which mean-reverts on the LNG/power/data-center demand calendar [31]. The counter-fact in the same treatment is genuinely permanent and price-independent: proved reserves per diluted share fell 32%, from ~66.8 Mcfe (FY2023) [32] to ~45.6 Mcfe (FY2025), as the share count rose with no repurchases to offset it [33]; one judge sat at 0.56, and the tally logs a temporary-side quote-check failure, so the 0.29 permanent tail is substantive rather than nominal. P5 is met on the framework's threshold but does not offset the gate. Full treatment: Damage Math.

Instrument context (I1) — not verifiable

Long-dated listed options exist (January 2028 LEAPS, ~17 months of tenor) and 30-day implied volatility was 27.1% as of 2026-07-31 — well inside the framework's ~50–55 acceptable line. But precise current open interest could not be verified from an accessible citable source, so the criterion is recorded not_verifiable. This never blocks the pillar verdicts; it would only drive a watchlist-only overlay if the pillars otherwise fit. Full treatment: Clock.

What a 3x-in-3-years would require

The tally's re-rating math is unavailable, and this is rendered exactly as it stands: the applicable bar and the normalized adjusted FCF are both missing, because adjusted FCF, adjusted-FCF yield and balance-sheet class all returned not_computable in the deterministic feature file (SBC missing every year, EBITDA missing FY2025). No price-at-bar-yield on normalized adjusted FCF can be computed from the record of this run.

The closest available arithmetic comes from the Damage Math tab's DCF-lite, which is calibrated to the tape rather than to the framework bar: a 10% no-growth capitalization of ~$3.3B normalized FCF gives ~$33.0B, essentially the current $32.8B market cap — so the price already embeds roughly mid-cycle cash with no growth. For base-rate context, EQT's own history (the Clock tab) shows four drawdowns of ~28–43% in the post-2020 pure-play era, recovering from trough to prior peak in ~3 to ~28 months depending on the gas cycle; the current 28% episode is the shallowest of that set and has already retraced ~9%. A framework-style 3x is not supported by this arithmetic: the entry yield is below the bar, the dislocation is mild, and the re-rating lever is an undated commodity price rather than a scheduled catalyst.

Contested and undetermined

Nothing was contested and nothing was cannot-determine. The tally records no contested criteria and no criterion resolved to cannot_determine. Two criteria carry an internal split without changing the aggregate: P4a met on a 3–1 seat split (no cross-family agreement), and P4a's masked-probe seat read not_met while the open seats read met — noted here for completeness, not as a contested verdict. I1 is recorded not_verifiable (open interest not citable), which is a data limitation rather than a contested reading.

Provenance

Source: the report's deterministic verdict tally and jury records.

Two independent model families supplied four voting seats plus a name-masked probe; they agreed on every gate criterion, and masking the company's name barely moved any probability (gap 0.01), so the verdict is not an artifact of the models recognizing EQT. The one load-bearing probability with meaningful disagreement is the diagnosis (spread 0.16), and even there the order-stability gap of 0.07 shows the trial ruling did not depend on which brief the judges read first.

The falsifier ledger

These are the standing what-would-change-this conditions carried verbatim from the tally — the framework's five templates first, then the name-specific thresholds, directions, and windows the jury and trial recorded (several overlap because more than one seat contributed):

Data gaps

The run could not resolve the following, from the tally's list:

• Adjusted FCF (FCF − SBC − 5-year-average acquisitions) is not_computable: SBC is missing for every fiscal year 2016–2025 in the feed, so the framework's core FCF measure, its yield, yield baseline, float-retirement years and FCF stability all return not_computable; the figures on this tab are reconstructed from the filed 10-K cash-flow statements.
• Balance-sheet class returned "unknown" (EBITDA missing for FY2025), so net-debt/EBITDA leverage class could not be computed deterministically; EBITDA was computed here as operating income plus depreciation, depletion and amortization (1.31x, moderate band).
• Live, dated web verification of current market cap, exact float, and sell-side consensus positioning was unavailable (web-research preload failed on provider credits); universe facts rest on the feature file (priced 2026-07-31) and the SEC identity record.
• Peer valuation multiples (EV/EBITDA for AR, RRC, CNX, CTRA, EXE, CRK) could not be computed — peer financials were empty in this run — so the not-darling read is anchored to EQT's own multiples and revenue shape rather than a peer table.
• Precise current listed-options open interest / liquidity depth was not verifiable from a citable source (option pages blocked automated access); LEAPS existence and dated IV levels were verifiable, but open interest is stated qualitatively only.
• Reported short interest, borrow pressure, and public net-short disclosures are all unavailable for EQT in this run, so short-side seller composition and any crowding cannot be characterized.
• The framework's acquisition dock is cash-based only; EQT's largest deals (Alta, Tug Hill, Equitrans, Olympus) were substantially stock-funded and are invisible to the FCF adjustment — captured only via the share-count trend.
• Natural-gas forward-strip and winter-2026-27 weather/storage assumptions that would date the near-term commodity catalyst are not in the corpus; the near-term re-rating timing is bounded but not pinned.