1. Bottom Line & Directional Bias
We are long COMEX December copper (HGZ26) with invalidation at a daily close below 6.6607 (pivot S2) [settle 2026-10-14]. Three reasons. (1) Trend intact: the 10-14 settle of 6.6605 is up 2.33% over 20 sessions and sits at the 45.2% position of the 20-day channel 6.4395–6.9285; the last completed weekly bar (2026-10-05–10-09) closed at 6.692, +2.18% w/w, so the current pullback is a retracement, not a reversal. (2) Physical tightness outside China: LME warehouse stock fell 1,625 MT d/d to 229,650 MT as of 2026-10-13 and is down 19,575 MT over 20 reports, while COMEX registered added only 1,100 short ton to 471,300 as of 2026-10-12 — a slow rebuild that does not yet offset the drawdown. (3) Positioning has room: managed-money netPct dropped to 20.11% (crowding 61.66 percentile) as of 2026-10-06 from 27.36% a fortnight earlier, so the long is not crowded. The risk is the SHFE warrant surge of 14,050 MT to 23,457 MT as of 2026-10-14, and the Asia print of 6.6168 at 07:00 on 2026-10-15 already sits below S2 — if that holds into the 10-15 daily close, the long is wrong. Upside: R1 6.7683, then R2 6.8107. Downside: S1 6.6933, S2 6.6607, then the 20-day low 6.4395.
2. Price Action & Technical Analysis
The prior session settle is 6.6605 [settle 2026-10-14], down 0.97% on the day, up 0.17% over 5 sessions and up 2.33% over 20 sessions. ATR14 is 0.1226, i.e. 1.84% of price as the full expected daily range, and RV20 is 17.8% annualized from settled daily bars. The 20-day channel runs 6.4395–6.9285 with the settle at the 45.2% position; the 52-week range is 4.9045–6.96, so price is in the upper fifth of the annual range. The classic pivot ladder from 2026-10-13 H/L/C governs the 10-14 session: P 6.7357, R1 6.7683, R2 6.8107, S1 6.6933, S2 6.6607. The 10-14 settle of 6.6605 finished essentially on S2 and below P, a weak close that hands the near-term burden of proof to the bulls. The last five settled bars show the shape: 10-08 C 6.5685 (H 6.763/L 6.5475), 10-09 C 6.692 (H 6.72/L 6.5115), 10-12 C 6.747 (H 6.783/L 6.694), 10-13 C 6.726 (H 6.778/L 6.703), 10-14 C 6.6605 (H 6.789/L 6.599) — the 10-14 bar broke below the 10-13 low of 6.703 and closed near its own low, a bearish-session signal inside an uptrend. The last completed weekly bar (2026-10-05–10-09) opened 6.573, high 6.763, low 6.5115, closed 6.692, +2.18% w/w; the current week from 2026-10-12 is unfinished and no weekly-close conclusion is drawn from it. In early Asian trade on 2026-10-15 (07:00), HG=F last printed 6.6168, down 0.66% vs settle, with an intraday high of 6.633 and low of 6.6145 — that is below S2 6.6607 and below the 10-14 settle, so the invalidation level is being tested intraday, not yet on a daily close. A break below S2 targets the 20-day low 6.4395; a reclaim of P 6.7357 opens R1 6.7683 and R2 6.8107. View: long, but only while the daily close holds above 6.6607.
3. Supply-Demand Balance & Fundamental Drivers
The physical balance is mixed but net-supportive outside China. COMEX registered stock stood at 471,300 short ton as of 2026-10-12, up only 1,100 short ton d/d (+0.23%) from the 2026-10-09 reading — a marginal rebuild that does not materially loosen the US deliverable pool. LME warehouse stock was 229,650 MT as of 2026-10-13, down 1,625 MT d/d (-0.7%) and down 19,575 MT over the last 20 reports, a persistent drawdown that signals continued consumption of exchange inventory. The offset is China: the SHFE warrant jumped 14,050 MT to 23,457 MT as of 2026-10-14, a +149.36% d/d increase from 9,407 MT on 2026-10-13 — a single-day doubling that argues for softer Chinese spot demand or warrant registration ahead of delivery, and it is the clearest bearish datapoint in the block. No WASDE or ICSG balance sheets are available, so no global surplus/deficit figure is cited. On the macro transmission side, the 10-year Treasury yield is 5.19% and the dollar index is 102.53, up 0.34% on 2026-10-14; a firmer dollar and high carry cost are a headwind for dollar-denominated copper, but the copper/gold ratio at the 100th 1-year percentile says the market is still pricing pro-growth demand. Term structure is in contango: M1-M2 is -0.0315 (-0.47%) with a roll yield of -5.62% and slope 0.0342 — contango is a carry/roll cost for longs, not a price cap, and it is consistent with adequate nearby supply rather than prompt tightness. Seasonality over the next 20 sessions (last 15 years, same calendar start) shows a mean of +0.42%, median +1%, up 8 of 15 years, best 2016 +19.17%, worst 2015 -10.43% — a mild positive tilt, small sample, context only. Headlines note copper price forecasts lifted for 2026 with a tariff-collapse scenario still on the table, and end-users restocking at high prices (2026-10-14) — supportive of near-term physical offtake. View: tight ex-China inventories and a pro-growth cross-asset tape support the long; the SHFE warrant surge is the risk to watch.
4. Positioning & Fund Flows
CFTC positioning has been de-risking, which is constructive for a long entry. Managed-money net length fell from 82,522 contracts (netPct 27.36%, crowding 69.05 percentile) as of 2026-09-22 to 78,058 (netPct 25.92%, crowding 68.51) on 2026-09-29 and to 61,134 (netPct 20.11%, crowding 61.66) on 2026-10-06, with the 10-06 week showing a -16,924 weekly change in net. Open interest was 304,033 on 2026-10-06 versus 301,657 on 2026-09-22 — roughly flat — so the net decline came from long liquidation (longs fell from 96,421 to 76,836) rather than a short build (shorts 13,899 to 15,702). This is a washout, not a divergence: price held its 20-day uptrend while specs cut length, leaving the CTA proxy at 98 (max trend-following conviction) and hedge ratio at 56.9%. Crowding at the 61.66 percentile of the 3-year window is moderate — the trade is not crowded, so there is room for trend followers to re-add. On volatility, RV20 is 17.8% while the metals implied-vol complex is calm: GVZ at 22.09 (1Y percentile 9%) and OVX at 48.77 (1Y percentile 42%) as of 2026-10-13, with VIX at 16.87 (1Y percentile 46%) on 2026-10-14 — optionality is not pricing a copper-specific event, so upside option premium is relatively cheap versus the realized trend. Retrospective risk metrics: 52-week max drawdown 13.49%, 20-day drawdown 4.36%, Sharpe30 0.9838, VaR95 -2.75% — these describe the past and are not the reason for the trade. View: positioning is light enough to support a long, with the CTA proxy at 98 flagging trend-following demand if price reclaims P 6.7357.
5. Cross-Asset Relative Value
Cross-asset ratios as of 2026-10-14 confirm a pro-growth tape. The copper/gold ratio is 1.62, at the 100th 1-year percentile and 63rd 3-year percentile (a parallel 1.61 print sits at the 100th 1-year and 62nd 3-year percentile) — copper is outperforming gold at the top of its one-year range, a classic late-cycle growth signal that supports the long. The gold/silver ratio is 73.52, at the 87th 1-year percentile and 29th 3-year percentile — a high ratio means gold is relatively rich and silver is lagging, so the metals complex is not uniformly bid; the copper strength is specific, not a broad precious-metals move. The oil/gold ratio is 0.0231, at the 97th 1-year percentile and 53rd 3-year percentile, and the 3:2:1 crack spread is 65.16 USD/bbl — energy is firm relative to gold, consistent with a reflationary backdrop. DXY is 102.55 and the 10-year yield is 5.19% [2026-10-14]; a strong dollar is the main cross-asset headwind, but copper has outperformed despite it. The contango (M1-M2 -0.0315, roll yield -5.62%) is a roll cost for longs, not a signal. View: the copper/gold ratio at the 100th 1-year percentile validates the long; the high gold/silver ratio warns the bid is growth-specific, so the trade is a copper long, not a metals long.
6. Historical & Seasonal Patterns
The seasonality block covers the next 20 sessions from the same calendar start, sampled over the last 15 years: mean return +0.42%, median +1%, with an up-hit rate of 8 of 15 years (53%). The distribution is skewed by outliers — best year 2016 at +19.17%, worst 2015 at -10.43% — so the median, not the mean, is the better central estimate, and the sample is small enough to treat as context only, not as a standalone signal. Two observations matter for the current long. First, the 8-of-15 hit rate is only mildly positive, so seasonality neither confirms nor contradicts the bullish call; it simply does not fight it. Second, the wide dispersion (a 29.6-point spread between best and worst) warns that this window has produced genuine drawdowns even in up-years, which is consistent with the 52-week max drawdown of 13.49% and argues for the hard stop discipline in section 8 rather than a seasonality-based averaging-down. The current 20-day gain of +2.33% [settle 2026-10-14] already exceeds the full-window median, so the seasonal tailwind is partly front-loaded; the remaining 20-session edge is modest. Net: seasonality is a mild positive overlay (median +1%, 53% hit rate) that supports holding the long into year-end but provides no justification for adding risk if the 6.6607 invalidation triggers.
7. Scenario Analysis (Base / Bull / Bear)
Base case — Bullish (probability 55%). Trigger: the 10-15 daily close holds above 6.6607 (pivot S2), reclaiming the level the Asia print of 6.6168 (07:00, 2026-10-15) is currently testing, and price reclaims P 6.7357 within five sessions. Target: R1 6.7683 first, then R2 6.8107; the validated strategy target 6.8444 sits just beyond R2 and is the full-objective level. Action: hold the validated long at entry 6.6605 with stop 6.5379; add nothing until P 6.7357 is reclaimed on a daily close, at which point the CTA proxy (98) implies trend-following demand can extend the move. This case agrees with the section 1 call and leans on the ex-China inventory drawdown (LME -19,575 MT over 20 reports) noted in section 3.
Bull case (probability 20%). Trigger: a daily close above R2 6.8107 with the copper/gold ratio holding at or above its 100th 1-year percentile, and LME warehouse stock extending its drawdown below 229,650 MT. Target: 6.9285 (20-day high), then the 52-week high 6.96. Action: hold the full 3-contract cap and trail the stop from 6.5379 to just under P 6.7357 once R1 6.7683 is taken out, locking in a positive-carry exit while the trend persists.
Bear case (probability 25%). Trigger: a daily close below 6.6607 invalidates the long; a sustained break under the 20-day low 6.4395 flips the bias to bearish, with the SHFE warrant surge (+14,050 MT to 23,457 MT as of 2026-10-14) as the confirming fundamental. Target: 6.4395 initially, with the 52-week drawdown context (13.49%) implying room toward the mid-6.3s if the break sustains. Action: exit the long on the 6.6607 daily close; do not re-enter until price reclaims S1 6.6933 and P 6.7357 on consecutive daily closes. Probabilities sum to 100%.
8. Trading Strategies & Risk Management
The validated strategy, quoted verbatim: LONG entry 6.6605 / stop 6.5379 / target 6.8444 / horizon 1-2w / R:R 1:1.5, with the identical 1-5d variant carrying the same entry, stop, target and R:R. The stop at 6.5379 sits 0.1226 below the 6.6605 entry — exactly 1× ATR14 (0.1226) per the risk math — placing it one full expected daily range beneath the entry and just under the 10-14 bar's low of 6.599, so a normal intraday retracement does not trigger it while a genuine break of the 6.6607 invalidation zone does. Contract multiplier is 25,000, giving risk per contract of 3,095 USD including fees and slippage; a 10,000 USD risk budget therefore caps size at 3 contracts. Scaling: enter the full 3-contract allocation only if the 10-15 daily close holds above 6.6607; if the Asia weakness (last 6.6168 at 07:00) persists into the close, stand aside entirely rather than average down. Scale out one-third at R1 6.7683, one-third at R2 6.8107, and trail the remainder to the 6.8444 target with the stop raised to breakeven once R1 is taken out. The 1:1.5 R:R means the 0.1226 stop risks 0.1839 of upside, consistent with the 6.8444 objective. Do not widen the stop to accommodate the SHFE warrant headline; the 1×ATR placement is the program's validated figure. If stopped, re-entry requires a daily close back above S1 6.6933 and a reclaim of P 6.7357, with the same 6.5379 stop re-armed from a fresh 6.6605-area entry.
9. This Week's Data Calendar
All times BJT | ET. Thursday 2026-10-15: 20:30 | 08:30 — Core PPI m/m (F 0.3%, P 0.2%, surprise if outside F±0.1%), Core Retail Sales m/m (F 0.6%, P 1.4%, ±0.8%), PPI m/m (F 0.6%, P 0.4%, ±0.2%), Retail Sales m/m (F 0.3%, P 1.2%, ±0.9%); all flagged USD/HIGH, transmitting to GC, SI, DXY. 11:30 | 23:30 — Fed Chairman Warsh Speaks (USD/HIGH). Sunday 2026-10-19: 10:00 | 22:00 — China Fixed Asset Investment ytd/y, Industrial Production y/y, Retail Sales y/y, Services Production Index y/y (CNY/HIGH, direct HG impact). Monday 2026-10-20: 09:15 | 21:15 — 1-y LPR (P 3%) and 5-y LPR (P 3.5%); 09:30 | 21:30 — GDP y/y. The 10-15 US data cluster is the immediate DXY risk to the 6.6607 invalidation test.
This report is generated automatically from public quantitative and macro data for research and market tracking only. It does not constitute investment advice or a recommendation to trade.