2026-10-23 17:00 ET — Commodity-linked large caps: oil majors (XOM/CVX/COP fade-shorts) vs gold miner NEM (fade-long) after the Brent-up / gold-below-$4,000 Friday

Slot: Fri 2026-10-23, 17:00 ET hourly research (after the close). Status: FINAL (whiggity consult added in §10).

TRAI data used (all returned 200 at ~16:45 ET):

• /api/latest: run 2026-10-23_postclose_2026-10-23T20-34-13-980Z (4:34 PM ET; daily stats = Oct 23 session). Regime transitional: VIX 19.45 (+0.42), VIX9D 18.70, VIX3M 20.48, VVIX 109.7 (mild contango); SPY −0.52%, QQQ −0.01%, SMH +0.74%.
• /api/events (asOf 2026-10-23), /api/correlations (60 sessions to 2026-10-22), /api/scorecard (replay 2026-04-23 → 2026-10-21, outcomes through 10-22), /api/macro (asOf 2026-10-22), /api/regime, /api/options.

Other data (my pulls and calcs): Yahoo chart API daily bars (2016-01 → 2026-10-23 close) via yfinance, including futures (CL=F, GC=F, BZZ26.NYM); earnings-date history from Yahoo via yfinance. "o→c" = today's open to close. "My calc" = computed by me from those bars; not a published figure.

1. Focus and why

• Focus: the four commodity-linked large caps in today's postclose buckets: XOM +2.79%, CVX +2.88%, COP +3.55% (all mean_reversion fade-short) and NEM −4.59% (mean_reversion fade-long) (TRAI /api/latest, 4:34 PM ET).
• Why this hour: these are Friday-close signals, so the first possible trade is Mon 10-26 after a weekend with live Iran/Saudi oil headline risk. XOM and CVX report Fri 10-30 BMO and are already inside TRAI's 5-day "elevated" gap-risk window. NEM's move was its own earnings reaction (Q3 released 10-22 after the close), which TRAI's event layer missed (S-008).
• Rotation: today's earlier reports covered semis (09:00), software (10:00), telecom/SpaceX + oilfield services SLB/HAL (11:00), AI power (12:00), payments (13:00), mega-cap cloud (14:00), alt managers/SPOT (15:00) and staples (16:00). The oil majors, E&Ps and gold miners have not been covered.
• One-line read: the energy fade-shorts are fades of an oil move with no company news, and oil has not turned (Brent settled at its highest since Apr 29, Iran strike talk into the weekend). TRAI's own replay says energy fade-shorts in transitional regimes have lost money (n 22, −0.97%/trade, t −3.28; §6). NEM's fade-long is a fade of an earnings + downgrade + gold-below-$4,000 day; history after NEM's big down reactions is roughly flat (§5.4).

2. What the businesses are and what moves them

• XOM (ExxonMobil): integrated oil & gas: Upstream, Energy Products (refining/fuels), Chemical Products, Specialty Products. The SEC registrant is now "ExxonMobil Holdings Corporation" (CIK 2115436) (XOM 8-K, 2026-10-16).
• CVX (Chevron): integrated; upstream-heavier than XOM (Permian, TCO/Kazakhstan, Australia LNG, Venezuela).
• COP (ConocoPhillips): pure upstream E&P (Lower 48, Alaska, Norway, LNG). No refining, so the most direct crude beta of the three.
• NEM (Newmont): largest gold miner (Nevada JV with Barrick, Australia (Tanami, Cadia, Boddington), PNG (Lihir), Peñasquito (Mexico), Ghana). Revenue moves with gold. Costs move with oil/diesel, AUD and royalties.
• Drivers in numbers (Yahoo daily, my calc):
• Daily-return ρ with WTI front month (CL=F, roll days excluded): 2026 YTD XOM 0.58, CVX 0.57, COP 0.63; last ~60 sessions 0.63 / 0.64 / 0.66. Beta to CL=F in 2026: XOM 0.28, CVX 0.25, COP 0.34. So a 4% oil day "explains" ~1.0–1.4% of their move.
• SPY–CL=F ρ: 2026 YTD −0.40 (2019–2025 range +0.03 to +0.35). Oil is trading as a risk-off shock this year. That is why XOM/CVX/COP show negative 60-day beta to SPY (TRAI beta60 −0.97 / −0.78 / −0.85).
• NEM ρ with gold (GC=F): 2026 YTD 0.74, last ~60 sessions 0.84. Beta to gold 1.24 (GDX 1.41, AEM 1.32). NEM–CL=F ρ −0.17 in 2026 (+0.06 to +0.27 in 2021–2025). NEM–SPY ρ 0.47 in 2026 (0.17–0.32 in 2020–2025).
• YTD total return to 10-23 close (Yahoo dividend-adjusted close, my calc; COP price-only +61.1%): COP +64.6%, XLE +58.1%, CVX +57.3%, XOM +55.3%, NEM +9.7%, GDX −6.5%, gold (GC=F) −7.6%, SPY +12.3%, WTI (CL=F) +80.1%.

3. Today's drivers (macro)

3.1 Oil up for a third week

• Brent (Dec) settled $116.72, +$3.55 (+3.14%), highest since Apr 29; WTI $103.27, +$3.80 (+3.8%). Drivers: continued Houthi attacks on Saudi Arabia and a tanker oil spill off Umm Al Quwain (UAE). Brent +6% on the week, WTI +2.3% (Reuters via Investing.com, 2026-10-23).
• Expectations that the US resumes strikes on Iran; Iran's army chief talked of "asymmetric warfare"; Brent +6.2% on the week, its third weekly gain (CNBC, 2026-10-23).
• ING: its Q4 Brent forecast of $80 "appears unattainable" (WSJ, 2026-10-23 3:54 PM ET). WSJ gives WTI +4.3% on the week vs Reuters' +2.3%. The difference is unresolved (probably different contract/reference points).
• Implied Oct 22 settles (Reuters figures above, my arithmetic): Dec Brent $113.17, WTI $99.47. Aramco ordered evacuation of its Dhahran HQ on Oct 22 (same Reuters link).
• Contract note: Yahoo BZ=F has rolled to the Jan-27 contract and shows −1.33% today. That is a roll artifact; the Dec-26 contract (BZZ26.NYM) closed +3.30% on Yahoo. Use the Reuters settle above (see energy note).
• XOM-specific (refining): XOM shut a 50 kbpd catalytic reformer at its 275 kbpd Joliet, IL refinery on Oct 20 ("working to resolve an operational issue"); a September power outage had already caused two weeks of downtime (Reuters via BOE Report, 2026-10-21). Chicago-area gasoline is ~$4.82 vs ~$4.27 national (CBS Chicago, 2026-10-23; headline seen via MarketBeat, article not opened). No restart date. Not quantified for Q3/Q4.
• No XOM, CVX or COP 8-K, guidance change or company release found for 10-23. The move is the oil factor.

3.2 Gold below $4,000

• Spot gold −0.6% to $3,998.49 at 2:27 PM ET, below $4,000 for the first time since mid-July; −4% on the week, the worst week since June. Dec futures settled −1.3% at $4,007.90. Driver: oil-led inflation worries lifting rate-hike odds (CME FedWatch: ~92% hold at Oct 27–28 FOMC, ~78% chance of a December hike), plus margin calls from the tech selloff (CNBC, 2026-10-23). Also Anadolu, 2026-10-23.
• Intraday 3-month low ~$3,975; support cited at the July 17 low $3,959 and the YTD low $3,941 (FXStreet, 2026-10-23). The 52-week high is $5,477.79 (USA Today, 2026-10-23 8:13 AM; secondary).
• Today, oil and gold moved in opposite directions through the rate channel. That hits NEM twice: lower gold revenue and higher diesel/consumables costs (§4.4).

4. Company prints, guidance and consensus (primary where possible)

4.1 NEM: Q3 2026, released Thu 10-22 after the close (call 5:30 PM ET)

Source: Newmont release, 2026-10-22 (also 8-K, 10-Q).

• Q3: attributable gold 1,302 koz; CAS $832/oz; AISC $1,454/oz; realized gold $4,317/oz (−$97 q/q). Adjusted NI $2.3B ($2.20/sh); reported NI $2.5B ($2.35); adjusted EBITDA $4.0B; operating cash flow $2.8B; FCF $1.9B. Dividend $0.26. $1.6B bought back since the last call; $2.8B left of the $6.0B program. Cash $8.8B; net cash $3.2B before a $1.95B payment to Barrick in Q4 (Nevada Gold Mines amendment covering Fourmile, Fiberline, Mike).
• FY26 guide: production 5.26 Moz; CAS improved to $925, AISC to $1,520.
• The problem is Q4: production 1,365 koz (strongest quarter), but CAS $1,250 and AISC $1,900/oz, sustaining capital $600M, development capital $520M. The release cites higher sustaining capital, lower co-products (Peñasquito silver ~half of Q3), and inflation "including higher oil prices". Lihir El Niño drought risk could hit Q4 production and costs.
• 2027 preview: production up low-to-mid single digits; unit costs higher (inflation above the 3% assumption, diesel, AUD, royalties, lower Peñasquito co-products); development capital up ~$500M. Red Chris block cave FID in early 2027 ($4.2–4.8B, 100% basis; first production 2033). Tanami Expansion 2 commercial production H1 2027.
• Assumptions and sensitivities (release table): guidance assumes gold $4,500 and Brent $70. Gold ±$100 → $505M; Brent ±$10 → $60M direct + $50M indirect (new indirect line). The table does not state the period.
• _My arithmetic on the table's own basis (not a company figure):_ Brent at $116.72 is ~$47 above the assumption, which is ≈ 4.7 × $110M ≈ $510M of cost. Gold at ~$4,000 is ~$500 below the assumption, which is ≈ 5 × $505M ≈ $2.5B less revenue/margin.
• Consensus: adjusted EPS $2.05 (LSEG); _correction (§10 item 1): the $6.36B (+15%) revenue in this Reuters piece is the reported figure (sales $6,363M vs $5,524M a year earlier), not consensus; revenue consensus was $6.24–6.28B (below), so revenue also beat_; shares +1.5% after hours on 10-22; CEO said Newmont is not interested in bidding for Northern Star (Reuters via Investing.com, 2026-10-22). Other sources give $2.11 / $6.278B (Benzinga, 2026-10-23 7:15 AM), ~$1.95 / ~$6.24B (Investing.com, 2026-10-23) and $1.94 (TIKR, 2026-10-23). Range $1.94–2.11: a beat against every source.
• Downgrade: Freedom Broker cut NEM to Hold from Buy, PT $130 → $116, citing Australian cost inflation and Lihir/Cadia risk; premarket −1.5% at $112.40 (Benzinga above). The same firm cut AEM and Barrick the same day (Investing.com above). 52-week high $135.29 (Investing.com above).
• Reaction (TRAI /api/latest + Yahoo): gap −3.36% (0.94 ATR on ATR% 3.56), o→c −1.28%, close $108.87, −4.59%, relVol 1.45. Residual vs GDX (−1.82%) ≈ −2.8 pp. The +1.5% after-hours move reversed into a −4.6% day. The market read Q4 AISC and the gold/oil squeeze over the Q3 beat.
• TRAI flags: NEM has next = null and only 3 past events in /api/events. The Oct 22 print is missing, so the event layer did not mark today as NEM's reaction day (S-008, first flagged at 09:00).

4.2 XOM: Q3 on Fri 10-30, release ~6:30 AM ET, call 9:30 AM ET

• Timing: results Friday Oct 30 ~5:30 AM CT, call 8:30 AM CT (XOM IR, 2026-10-16).
• 8-K "earnings considerations" (3Q vs 2Q, after-tax, explicitly not an estimate) (XOM 8-K Ex. 99.1, 2026-10-16):
• liquids price −$1.3B; gas +$0.3B; industry margins +$4.4B (Energy Products +4.5, Chemical −0.6, Specialty +0.5); scheduled maintenance +0.2; Middle East volume disruptions 0.0; external impacts on operations −0.6; day effect +0.3; other −0.9.
• Identified-items/timing memo: divestments +0.7, other reserves −0.2, timing effects −2.5 (−0.7 Upstream, −1.9 Energy Products).
• ~4,140M diluted shares. 2Q26 base: GAAP $14.5B, adjusted $14.7B (Upstream adj. 9.2, Energy Products 4.1).
• 2Q26 EPS $3.52 adjusted (missed $3.56); Q3 dividend $1.03. Q3 consensus EPS $4.10 / revenue $105.43B; beat in 6 of the last 8 (MarketBeat, accessed 2026-10-23 ~17:10 ET). An early-October preview had $3.40 / $94.88B (ScanX, 2026-10-02, pre-8-K; likely stale). Consensus moved up a lot after the 8-K; the LSEG/FactSet figure was not checked.
• TRAI /api/events: 10-30 BMO, elevated (5 trading days). Last 4: avg |gap| 1.26% (0.59 ATR), fill 75%, continuation 0%, avg day-1 in gap direction −0.86%. All 4 gaps were negative (−0.62, −1.26, −1.11, −2.06%); close-close −0.29, +0.63, −1.02, −0.97%.

4.3 CVX: Q3 on Fri 10-30 BMO, call 11:00 AM ET

• Call time from Chevron advisory via Business Wire/Yahoo, 2026-10-01. BMO per TRAI and MarketBeat ("confirmed").
• Consensus EPS $4.86 / revenue $61.33B; Q2 2026 $6.06 vs $5.55; beat 7 of 8 (MarketBeat, accessed 2026-10-23 ~17:10 ET). Others: $4.66 / $56.75B (Hudson Labs, 2026-10-06); $4.32 (ScanX, 2026-10-02). Range $4.32–4.86; the newest is highest. Note: Hudson Labs lists Q2 EPS as $5.60 vs MarketBeat's $6.06 (different adjustments); unresolved.
• FY26 context (Hudson Labs preview, secondary): capex $18–19B (low end); production +7–10%; $3B structural cost cuts achieved early; Venezuela target 600–700 kbpd by 2031; TCO 3GP; Project Kilby power (2.67 GW, Microsoft PPA, FID late 2026).
• TRAI /api/events: last 4 avg |gap| 1.20% (0.63 ATR), fill 50%, continuation 100%, avg day-1 in gap direction +1.25%; close-close +2.74, +3.34, −1.39, +2.35%.

4.4 COP: Q3 on Thu 11-05 BMO, call 12:00 PM ET

• Timing: ConocoPhillips IR, 2026-09-24.
• Consensus EPS $3.05 / revenue $18.56B; Q2 $3.24 vs $2.90; beat 7 of 8 (MarketBeat, accessed 2026-10-23 ~17:10 ET). Single source; not cross-checked.
• TRAI /api/events: 9 trading days out (not yet elevated). Last 4 avg |gap| 2.15% (0.77 ATR), fill 50%, continuation 50%; close-close −2.33, −2.43, −1.93, +1.50%; drift days 2–5 −4.39, −5.57, +8.67, +6.65%.

4.5 Peers and next-week calendar (verified)

5. Today's tape and the TRAI picture

5.1 Tape (TRAI /api/latest for c-c and signals; Yahoo daily for gap and o→c)

\* Not in the TRAI universe. On 10-22, AEM −5.98%, B −5.18%, GDX −5.80% (Yahoo), so the gold group was already down hard before NEM's print. NEM fell −2.62% that day. Refiners (MPC/PSX) fell intraday as crude rose (input-cost squeeze).

5.2 Linkages (TRAI /api/correlations, 60d to 10-22)

• Cluster 2 (avg ρ 0.713): MPC, PSX, XOM, CVX, COP, OXY, EOG, DVN. XOM–CVX 0.86, XOM–COP 0.82, CVX–COP 0.81, COP–EOG 0.88. XOM/CVX/COP are one trade: the "1 per cluster" cap allows only one of the three fade-shorts.
• Cluster 16 (avg ρ 0.645): FCX, NEM. NEM's next-highest: RIVN 0.50, GS 0.48, SOFI 0.47, HIMS 0.46, MARA 0.45, RIOT 0.43, HOOD 0.42. In 60-day data NEM trades with speculative risk-on names, not as a defensive.
• NEM vs the energy shorts: NEM–XOM −0.11, NEM–CVX −0.16, NEM–COP +0.05. A NEM fade-long next to an energy fade-short is not a hedge pair on 60-day data, but both legs lose if oil keeps rising and gold keeps falling (the 10-23 pattern).
• Sector matrix: Energy intra 0.62; Energy vs Semis −0.12; Materials intra 0.13 (FCX/NEM are the only two names).

5.3 Macro context (TRAI /api/macro, asOf 10-22)

• 10y 4.96%, 2y 4.65%, DXY 101.56; XLE 20-day relative strength vs SPY +8.36% (60-day return +15.2%), "leading"; XLB +4.47% rel20d.

5.4 Does history favour the fades? (my calc)

• Energy after a big oil day (Yahoo, 2016 → 2026-10-23): after days with XLE ≥ +2% and CL=F ≥ +3% (n 103), next-day o→c: XOM −0.29% (40% up, t −1.87), CVX −0.17% (t −0.99), COP −0.55% (t −2.26), XLE −0.34% (t −2.08). Next-day gaps were positive on average (+0.18% to +0.32%). Long history shows a small in-session fade.
• But 2026 only (n 12): XOM +0.02%, CVX +0.47%, COP −0.21%. Monday after a Friday event (n 18): XOM −0.01%, CVX +0.09%, COP −0.44%. No fade this year, and none on Mondays.
• TRAI replay, energy MR fade-shorts (net of 5 bp, my calc from /api/scorecard `replayTrades`; same convention as TRAI's own ranking): all n 29, −0.34% (hit 34%); transitional regime n 22, −0.97%, hit 23%, t −3.28. Clustered by day: 10 signal days, day-mean −0.94%, t −2.72, 8 of 10 days negative. Majors/E&Ps only (ex-refiners): n 11, −0.50%. On 7 of the 10 transitional signal days WTI (CL=F) had risen ≥ +2.7%, and on 7 of 10 it rose again the next day (Yahoo, my calc).
• Friday-signal (Monday-trade) slices, all sectors (TRAI replay, my calc): MR fade-shorts n 36, −0.30%, t −1.15; in transitional n 24, −0.71%, t −2.48. MR fade-longs n 33, −0.52%, t −1.52; transitional n 8, −0.76%.
• NEM after its own big down reactions (Yahoo, 43 prints 2016–2026, my calc): when the reaction day was ≤ −3% (n 15), next-day o→c +0.63% (median +0.59%, 53% up), residual vs GDX +0.24%, days +1..+5 +0.61%. Small and not significant: no reliable bounce or drift. All prints: next-day o→c +0.06%.
• TRAI replay, NEM/FCX fade-longs: n 6, +0.90% but 33% hit (one +4.40% NEM day on 2026-06-10). Materials/Utilities fade-longs n 9, +0.85%. Too small to use.
• AEM print → NEM (43 AEM prints, Yahoo, my calc): NEM's absolute move on AEM's reaction day is 2.20% vs 1.64% baseline, but same-sign only 56% and residual ρ (vs GDX) −0.13. AEM moves NEM through the sector factor, not a specific read-through.

5.5 Data-integrity correction to the 16:00 report and S-054 (found this hour)

• /api/scorecard replayTrades[].oc is already the trade's signed return (checked: COP fade-short signal 2026-09-15 has oc +4.25%, and COP actually fell −4.25% open→close on 09-16). TRAI's own ranking confirms it: mean_reversion_short n 322, expectancyNet +0.097%, t 0.77.
• The 16:00 report multiplied oc by side again, which flipped the sign of every short slice. Corrected (net of 5 bp, my calc): all MR fade-shorts n 322, +0.10%, t +0.77 (not −0.26%); Thursday-signal fade-shorts n 99, +0.76%, hit 69%, t +3.30 (not −0.92%, t −4.02); staples fade-shorts n 15, +0.53% (not −0.69%). Thursday fade-longs n 28, +0.95%, t 3.18 (the sign was right). Long-only and TRAI-native figures (S-002 short_momentum, S-013 transitional fade-short −0.05%) were not affected.
• So today's +1.31% live fade-short result agreed with the replay rather than contradicting it. An erratum is added to the 16:00 report §4 and S-054 is corrected (strategy.md). This report uses the corrected convention throughout.

6. Analysis

• Energy fade-shorts are fading the driver, not the stock. XOM/CVX/COP residual z-scores (1.75–1.85) are modest; the move is mostly oil beta × an oil shock plus a negative SPY beta on a down tape. Brent settled at a 6-month high with Iran strike talk into the weekend (§3.1). TRAI's replay shows energy fade-shorts in transitional regimes lost on 8 of 10 signal days (§5.4), and most of those were also oil-up days. This is S-013 ("check that the driver still holds") with numbers attached.
• The 2016+ next-day fade (COP −0.55%, XLE −0.34%) is real but small and absent in 2026. It does not offset the weekend headline risk or the replay result.
• Earnings proximity: XOM/CVX are 5 sessions from their prints, with small typical gaps (~0.6 ATR). XOM's pre-announced margin tailwind (+$4.4B) vs timing effects (−$2.5B) sets up a noisy headline beat/miss. CVX's last 4 reactions continued 100%. A Monday fade-short does not carry into the print (TRAI is flat by 3:55), so the print matters only for the Thu 10-29 → Fri 10-30 session.
• NEM fade-long: the gap was 0.94 ATR, below S-006's 2-ATR exclusion, so the fade is allowed by current proposals. It is still an earnings + downgrade + commodity day (S-025's "catalyst" case), and NEM's history after −3% reactions is flat (§5.4). The fundamental headwind (Q4 AISC $1,900; oil ~$47 above plan; gold ~$500 below plan) persists into Monday, and FOMC Wed 10-28 / AEM Wed AMC are the next catalysts. The fade-long has no edge I can find, but also no strong evidence against it beyond the generic Friday-signal slice (fade-longs −0.52%).

7. Actionable for today / the next session (explicit)

• Today: none. TRAI is flat by 3:55 PM; these are postclose signals.
• Mon 10-26 (ACTIONABLE as a caution, pending the 6 PM session):
• If Monday's premarket run again lists XOM/CVX/COP as fade-shorts, treat them as one cluster-2 bet (cap allows one). The evidence leans against taking it unless crude has reversed: energy fade-shorts in transitional regimes n 22, −0.97%/trade (8 of 10 days negative), Friday-signal fade-shorts in transitional −0.71%. Check Dec Brent (BZZ26; Dec Brent expires Fri 10-30) and WTI vs Friday's settles ($116.72 / $103.27) before 10:00. Skip if crude is flat or higher (S-059).
• Weekend gap risk: US–Iran strike headlines (CNBC, §3.1). Oil-headline gaps hit cluster 2 in the same direction, and NEM in the opposite direction through gold/rates.
• NEM fade-long: no edge found (§5.4). If taken, it fades an earnings reaction (S-025) into FCX Tue BMO (cluster 16), FOMC Wed and AEM Wed AMC. Size per normal rules; do not treat a NEM long as a hedge for an energy short (ρ ≈ −0.1; both lose on oil-up / gold-down).
• Tue 10-27: FCX BMO; NEM is the only same-cluster name (S-004 peer flag).
• Thu 10-29 → Fri 10-30: XOM + CVX BMO the same morning; cluster-2 peers (COP, OXY, EOG, DVN, ρ 0.75–0.86) carry peer-event risk on Fri 10-30 (S-004). Also Dec Brent's last trading day.

8. Strategy implications (proposals; see strategy.md)

• S-059 (PROPOSAL): commodity-driver check for fades on commodity-linked names. Skip MR fade-shorts on cluster-2 energy names when the signal-day move came with WTI/Brent ≥ +2.5% and front-month crude has not reversed below the signal-day settle by 10:00 ET. Mirror rule for gold miners vs GC. Evidence: §5.4 (n 22, −0.97%, t −3.28; 10 days, t −2.72).
• S-060 (PROPOSAL, data): add commodity context symbols (Dec/Jan Brent with an explicit roll calendar, WTI front, GC, GDX, AEM, B) and an AEM/B → NEM and XOM/CVX → cluster 2 peer-event map. Extends S-024 / S-004.
• S-061 (PROPOSAL, data quality; extends S-008): backfill missed prints (NEM 10-22) so the reaction day is tagged; until then, TRAI treated NEM's earnings day as a plain fade-long.
• S-062 (NEEDS DATA): oil–gold "double squeeze" on miners. When oil is up ≥ 2% and gold down ≥ 1% on the same day (as on 10-23), do gold-miner fade-longs underperform? NEM's own guidance quantifies the cost link (§4.1).
• S-063 / S-064 / S-065 (NEEDS DATA, from the whiggity consult, §10): gold-miner "beat but sold off" drift (pooled GDX top holdings), pre-print 52-week high → earnings gap (universe-wide), and an oil-factor (WTI/crack/SPY) residual spread for COP vs XOM.
• S-054 correction: sign error fixed (§5.5). Thursday-signal fade-shorts are positive in replay (n 99, +0.76%, t +3.30).

9. Unknowns / open questions

• XOM/CVX/COP consensus comes from MarketBeat, Hudson Labs and ScanX (secondary). They disagree (XOM $3.40–4.10; CVX $4.32–4.86) and were not checked against LSEG/FactSet.
• The Joliet reformer's restart date and its earnings impact are unknown.
• WTI weekly change: Reuters +2.3% vs WSJ +4.3% (unresolved).
• NEM sensitivity table: the period (annual vs remaining-year) is not stated; my $510M / $2.5B arithmetic uses the table as printed.
• CL=F continuous series includes roll jumps (I excluded |daily move| > 15%); correlations are close-to-close and do not separate overnight from intraday.
• The replay uses a daily-bar VWAP proxy and today's universe (survivorship), per /api/scorecard assumptions. Energy fade-short days overlap (several names per day), so the per-trade t-stat overstates; the day-clustered t (−2.72, n 10 days) is the fairer figure.
• No intraday bars: I can't say whether the 2016+ next-day energy fade happens before or after 10:00 ET.

10. Research partner (whiggity) notes

I gave Kieran's Monica research bot whiggity this report's draft summary after the close. I treat its reply as input from a second researcher, not as a source. Its citations were generic quote pages with no dates, so none of them is cited here. Items are tagged [whiggity]. A point is marked from whiggity, verified only where I confirmed it against a public source I opened myself (linked, accessed 2026-10-23 ~17:30–18:30 ET) or against my own Yahoo/TRAI calcs. Conflicts with my findings are listed separately. Everything else is an open question and is not used as fact.

Verified (from whiggity, verified)

1. NEM Q3 headline numbers: from whiggity, verified. Adjusted EPS $2.20, sales $6,363M ($6.36B), free cash flow $1,912M (~$1.9B) (Newmont release, 2026-10-22; also 8-K).
2. Fix to §4.1 found while checking: §4.1 listed "revenue $6.36B (+15%)" as consensus. $6.36B is the reported figure (Q3 2025 sales were $5,524M, so +15% y/y; same release). Revenue consensus was $6.278B (Benzinga) or ~$6.24B (Investing.com), both cited in §4.1. So Q3 beat on revenue too. §4.1 and the NEM note are corrected. I could not reopen the Reuters/Investing.com page (bot check), so the LSEG revenue consensus is unknown.
3. NEM FY26 AISC "improved ~10% to ~$1,520/oz": from whiggity, verified. FY26 gold by-product AISC guidance went from $1,680/oz (unchanged at Q2; Newmont Q2 release, SEC, 2026-07) to $1,520/oz on 10-22 (−9.5%). FY CAS went from $1,055 to $925. The release attributes the cut to "cost discipline and higher realized co-product pricing" (Newmont release above). This does not conflict with the Q4 $1,900/oz guide. See conflict C1.
4. CVX new 52-week high on 10-23 above a $226.68 prior close: from whiggity, verified (intraday price differs slightly). TRAI /api/latest (4:34 PM run): close $233.21, +2.88%, so prior close $226.68 ✓. Yahoo daily: 10-23 high $233.74, low $228.02. The previous 52-week high was $229.04 (10-22), which was also the highest daily high in Yahoo data since 2016. TRAI's rangePos 0.907 reproduces exactly with Yahoo's high/low, so TRAI agrees with $233.74, not whiggity's $233.53 (probably a different feed or a pre-close snapshot; small).
5. COP leadership transition: from whiggity, verified (but it has already happened). On Aug 6 the board elected Andy O'Brien (then CFO) President & CEO and board member, effective Sep 1, 2026. Ryan Lance retired as CEO and became Executive Chair. Kontessa Haynes-Welsh became CFO the same day (COP 8-K Item 5.02, filed 2026-08-11). 11-05 will be O'Brien's first quarterly call as CEO, which is a modest guidance/capital-framework risk on top of the usual print.
6. Dates: from whiggity, verified. They match §4.2–4.5: XOM Fri 10-30 (XOM IR), CVX Fri 10-30 BMO (Chevron advisory + TRAI), COP Thu 11-05 BMO (COP IR).
7. E&Ps have a higher crude beta than the integrateds: from whiggity, verified (beta, not correlation). Daily beta to WTI (CL=F, roll days excluded; Yahoo, my calc):

Correlations are much closer (2026: 0.57–0.66 for all seven). So E&Ps move more per unit of oil, but are not much more tied to oil.

1. SLB is the international/offshore name, HAL the North American name: from whiggity, verified (with a caveat for HAL). SLB Q3: international $6,993M of $9,449M (74%), North America $2,401M (SLB release, 2026-10-23). HAL Q3: North America $2,160M of $5,703M (38%), down 5% q/q on lower U.S. land pressure pumping; international $3.5B (HAL release, 2026-10-20). Caveat: HAL is now majority-international. It has more NA-land exposure than SLB, but it is not a pure NA-land play. In 2026, HAL–WTI ρ is 0.31 and SLB–WTI ρ is −0.01 (my calc), so neither tracks crude well this year.
2. NEM, AEM and Barrick are GDX's top three weights: from whiggity, verified. As of 10-22: NEM 11.72%, AEM 11.30%, B 8.05% (VanEck GDX holdings, as of 2026-10-22). The "drag via ETF flows" mechanism is not verified; see conflict C4.
3. AEM outperforms NEM on NEM-specific selloffs: from whiggity, verified, but only because AEM tracks GDX. On NEM's earnings-reaction days where NEM lagged GDX by > 2 pp (n 14, 2016–2026; Yahoo, my calc), AEM beat NEM on 14 of 14, by +4.88 pp on average. But AEM − GDX averaged only +0.31 pp. AEM was not a NEM-specific winner; it simply did not share NEM's idiosyncratic drop. 10-23 fits the pattern: NEM −4.59%, AEM −1.25%, GDX −1.82%.
4. Crack spreads fell on 10-23 while crude rose: from whiggity, verified for today; the Q3 picture is the opposite. NYMEX front-month 3-2-1 crack (RB/HO vs CL; Yahoo, my calc; contract months not aligned, so treat as approximate): $51.39 on 10-23 vs $58.52 on 10-22 (−12%); Oct-to-date average ~$62 vs ~$66 in Sept and ~$66 in July–Aug. MPC −0.45% and PSX −0.51% on a +4% crude day (§5.1) agree. XOM's 8-K shows Q3 industry margins +$4.4B vs Q2 (§4.2), so compression is a Q4 risk, not a Q3 one. On 10-23 XOM (+2.79%) and CVX (+2.88%) lagged COP (+3.55%) and DVN (+4.49%), which fits, n = 1.

Conflicts with my findings (listed explicitly)

• C1. NEM AISC $1,520 vs $1,900: both are right, and they measure different periods. Both are gold by-product AISC from the same 10-22 release (the guidance table has a FY26 column and a Q4 2026E column). FY26 = $1,520/oz (improved from $1,680). Q4 2026E = $1,900/oz. This is not a by-product vs co-product difference: Q3 co-product AISC was $1,963 vs $1,454 by-product, and the release gives no co-product guidance. Consistency check (my arithmetic on the release's own numbers): 9M by-product AISC $4,966M on 3,638 koz sold; FY guide $7,450M on 4,900 koz. That implies Q4 ≈ $2,484M / 1,262 koz ≈ $1,970/oz, which is within ±5% of the $1,900 Q4 guide. The FY figure improves because Q1–Q3 averaged $1,365. Q4 rises on higher sustaining capital (Q4 $600M of the $1,950M FY), Peñasquito silver at ~half of Q3, and oil. The report's §4.1/§6 framing (Q4 cost step-up is the problem) stands. whiggity's "improved" is also right, for the full year.
• C2. What drove NEM's selloff. whiggity says Red Chris block-cave capex plans. The facts are verified: development capital $4.2–4.8B (100% basis, real terms, net of the CAD$500M Canadian grant; FID early 2027; commercial production 2033), and 2027 development capital +~$500M (Newmont release). The attribution is not supported. The dated coverage I opened puts the move on gold sliding toward $4,000, the Q4 AISC of $1,900 and Freedom Broker's downgrade (TIKR, 2026-10-23; Benzinga in §4.1). TIKR mentions Red Chris as a secondary spending point. The tape also points to a sector factor: GDX −1.82%, gold −1.25% (§5.1). Kept as open question O1.
• C3. COP YTD return: ~66% (whiggity) vs my 64.6%. My §2 figures are total returns (Yahoo dividend-adjusted close, 2025-12-31 → 2026-10-23). COP: +64.6% total, +61.1% price only. Neither matches 66%. A different start date or source could explain it; not used. §2 now says "total return".
• C4. "XOM/CVX publish 8-K earnings considerations in the first week after quarter end." Wrong on both counts for Q3 2026. XOM's considerations 8-K was filed 10-16, 16 days after quarter end (XOM 8-K Ex. 99.1). CVX filed no earnings-considerations 8-K this quarter. Its October 8-Ks are (a) 10-05, officer changes: CFO Eimear Bonner becomes President, Oil, Products & Gas, and Jeff Gustavson becomes CFO, both effective 2027-01-01 (CVX 8-K, 2026-10-05); and (b) 10-06, divestiture of its Hess Midstream interests and DJ Basin crude midstream assets, with an expected one-time after-tax loss of ~$3–4B on closing to be treated as a special item (CVX 8-K, 2026-10-06). The Hess Midstream loss is new to this report. It should be excluded from adjusted EPS. Watch whether closing falls in Q4 and how GAAP vs adjusted EPS headlines read on 10-30. The causal claim that the 8-K "keeps earnings-day moves modest" is untested (O3).
• C5. Earnings-day move sizes. whiggity says implied ±1.5–2.5% for XOM/CVX and ±4–7% for miners. Realized moves since 2016 (Yahoo, my calc, close-to-close on the reaction day): XOM mean |move| 2.04% (median 1.59%, n 43; non-earnings baseline 1.23%); CVX 2.54% (median 2.52%, n 42; baseline 1.19%). So the XOM/CVX range is about right for realized moves. Miners: NEM 4.09% (median 3.29%, n 44); AEM 3.75% (median 3.41%, n 43). Typical miner moves are ~3.3–4%, below whiggity's 4–7%. NEM's 4.59% on 10-23 sits near its mean. I did not verify implied moves (no dated options source).
• C6. "NEM earnings drops drag GDX/AEM via ETF flows." On the 14 NEM idiosyncratic-down reaction days: GDX − GLD −1.49 pp, AEM − GLD −1.18 pp, NEM − GLD −6.05 pp. NEM's 11.7% weight explains only ~0.7 pp of GDX's lag, mechanically. The remainder is consistent with a sector move (miners also fell on 10-22, before NEM's print; §5.1). Flows are not shown, and I found no evidence that NEM drags AEM specifically (item 9, and the AEM→NEM check in §5.4: residual ρ −0.13).

Open questions (not verified; not used as facts)

• O1. Did Red Chris capex drive any part of the 10-23 drop? This would need the call Q&A or analyst notes. Was the $4.2–4.8B range an increase over a prior estimate? The release says it reflects "a more complete view of the investment requirements" but gives no prior figure.
• O2. "CVX at a 52-week high a week before earnings = stretched positioning." The high is verified (item 3). "Stretched" is not measured (no positioning data). See the quick look in Q3 below.
• O3. Do pre-announcement 8-Ks dampen XOM's reaction? n is small, and CVX gives no such 8-K (C4).
• O4. "Miner-to-gold correlation breaks on earnings days." Plausible by construction, since earnings add idiosyncratic variance (NEM's mean |move| is 4.09% on print days vs 1.64% otherwise). The correlation itself was not computed.

whiggity's testable questions: quick looks (Yahoo daily 2016–2026, my calc)

• Q1. NEM beat + day-0 drop > 4%: GLD-hedged t+1..t+5 drift. Beat per Yahoo's EPS-surprise field. n 5 (2018-10-25, 2024-02-22, 2024-07-25, 2025-02-21, 2025-10-24): NEM − GLD from the day-0 close to t+5 was mean +2.05%, median +0.23%, 3 of 5 up (−2.73% to +8.11%). Too small to use. Filed as S-063 (pool the top GDX miners for n).
• Q2. NEM print → long AEM / short GLD at the close → next-day alpha. All NEM prints: +0.03% (n 43, t 0.09). After NEM ≤ −3% reactions: −0.16% (n 15, t −0.52). No edge; not filed.
• Q3. 52-week high within 5 sessions before the print → day-0 gap (CVX/XOM). CVX n 4: gaps +0.84, −3.97, +1.51, −1.23% (all-print mean gap −0.05%, |gap| 1.54%). XOM n 7: mean −0.38% (−4.42% to +1.78%) vs all-print mean −0.43%. No pattern in n ≤ 7. Filed as S-064, universe-wide, so it can be tested with n in the hundreds.
• Q4. Rolling 60-day regression on ΔWTI, Δcrack, ΔSPY; trade the COP − XOM residual spread at 2σ. Not run (needs a clean crack series by contract month). Filed as S-065.

Did the consult change the verdict?

No. §7 stands: energy fade-shorts lean against unless crude reverses (S-059), and the NEM fade-long has no edge found. The consult added:

• a cleaner framing of NEM's guide (C1): FY better, Q4 worse, and the market priced Q4 plus gold;
• a revenue beat for NEM (the §4.1 consensus label is fixed);
• the CVX Hess Midstream $3–4B special-item loss and 2027 CFO change for the 10-30 print, and COP's first call under the new CEO on 11-05;
• CVX at a multi-year high into the print: a context flag, not a signal (Q3 sample too small).

_Screening/research, not investment advice._