I. Bottom Line & Directional Bias
Directional conclusion: bearish. SHFE Copper CU2611 [10-16 settle] 109,450, 5-day -0.23%, 20-day -2.09%, at the 62.7th percentile of the 20-day channel 106,650–111,770; ATR14 is 1,140.71 (1.04%), RV20 is 16.1%. The core rationale is threefold: first, China's supply side is marginally loosening, with SHFE warrants [10-16] at 36,386 tonnes, up 1,453 tonnes (+4.16%) on the day, while LME inventory [10-16] at 225,425 tonnes fell 1,175 tonnes on the day and has drawn a cumulative 29,675 tonnes over the last 20 readings, yet the international destocking has not transmitted to China; second, the positioning structure is bearish, with top-20 seats holding 341,518 lots long and 370,989 lots short, a net short of 29,471 lots, widening by 5,777 lots versus the prior day as shorts actively add; third, the main contract shows an open-interest-cut rally (short covering), with notional open interest of CNY 97.15bn and a daily change of -CNY 6.669bn, so the advance lacks confirmation from incremental longs. The transmission chain is: China warrant accumulation + widening net short at seats → near-month pressure → price testing below pivot P 109,550. Invalidation conditions: a daily close above 111,100 (below the 20-day high of 111,770), or a large sustained LME inventory build alongside a continued sharp rise in SHFE warrants.
II. Price Action & Technicals
Technicals are anchored to the [10-16 settle] of 109,450. The 5-day change is -0.23% and the 20-day change is -2.09%; price sits at the 62.7th percentile of the 20-day channel 106,650–111,770, with ATR14 at 1,140.71 (1.04% of price, full intraday range) and RV20 at 16.1%. Classic pivots are derived from the [10-15] high/low/close for the [10-16] session: P 109,550, R1 110,190, S1 108,300. The settle of 109,450 is slightly below P 109,550, a gap of about 100 points, less than half an ATR (570 points), so P does not constitute an effective trading level. R1 110,190 is 740 points from the settle, about 0.65× ATR, and can serve as an upside reference; S1 108,300 is 1,150 points from the settle, about 1.01× ATR, and is an effective downside support. The last five daily bars (C = close): 10-12 closed 110,630, 10-13 closed 110,990, 10-14 closed 111,080, 10-15 closed 108,940, 10-16 closed 109,860; after the 10-15 single-day drop, 10-16 saw a modest recovery but failed to reclaim the 10-14 high of 111,080. The last completed weekly bar (2026-10-08 to 2026-10-09): open 109,740, high 111,390, low 108,770, close 110,110, weekly change +0.39%; the current week (from 2026-10-12) has not yet closed and is tentatively at 109,860 (-0.23%), so no weekly conclusion is drawn. China night session recap (prior session [10-16] 21:05 to [10-17] 01:00): open 110,030, high 110,150, low 109,440, close 109,480, -0.35% versus the day session close and +0.03% versus the settle; the night session center of gravity moved lower, testing 109,440 intraday before recovering.
III. Supply-Demand Balance & Fundamental Drivers
On supply and demand, inventories show a pattern of falling abroad and rising at home. LME copper inventory [10-16] is 225,425 tonnes, down 1,175 tonnes on the day (-0.52%), with a cumulative draw of 29,675 tonnes over the last 20 readings, indicating continued tightening of international visible stocks; COMEX copper registered warrants [10-15] are 470,900 short tons, unchanged on the day (+0%), with a cumulative -5,400 short tons over the last 20 readings, a mild move. In China, by contrast, SHFE copper warrants [10-16] are 36,386 tonnes, up 1,453 tonnes on the day (+4.16%), versus the prior reading [10-15] of 34,933 tonnes (up 11,476 tonnes on the day); Chinese delivery warehouses have accumulated for two consecutive days, raising near-month supply pressure. Yangshan copper premium [10-16]: bill of lading average USD 130/t (low 122, high 138), warrant average USD 137/t (low 124, high 150), with bills up USD 1/t versus the prior day, so import costs show no clear weakening. On macro transmission, the US 10-year Treasury yield [10-16] is 5.1% (+0.18%) and the dollar index DXY [10-16] is 101.84 (-0.16%); higher Treasury yields impose a valuation drag on USD-denominated copper, partly offset by the modest dollar pullback. On supply-side news, Peru's August copper output fell 5% y/y to 230,676 metric tonnes (Cailian Press [10-17]), so international mine disruptions persist; however, the union at Antofagasta's Centinela copper mine reached an agreement with management on a new labor contract ([10-17]), easing strike risk. On term structure, the international curve is in CONTANGO, with M1-M2 at -0.0295 (-0.45%), RY at -5.39%, and slope 0.0344; the near-month discount structure persists, with no sign of a squeeze.
IV. Positioning & Fund Flows
The positioning structure is bearish. Top-20 seats (aggregate across the contract, [10-16]): longs 341,518 lots (daily change -8,924), shorts 370,989 lots (daily change -2,754), net short 29,471 lots, widening by 5,777 lots versus the prior day; shorts exceed longs by 29,471 lots in absolute terms, and the long reduction (-8,924 lots) is larger than the short reduction (-2,754 lots), a clear sign of active long liquidation. Main contract notional open interest is CNY 97.15bn, with a daily change of -CNY 6.669bn, open interest -12,140 lots and price change +0.1%, characterized as an open-interest-cut rally (short covering), i.e., the price rise was driven by short covering rather than active long additions, leaving the rally's momentum in question. International COT (last 4 weeks): [10-13] managed money net long 61,058 lots (Δ-76), [10-06] net long 61,134 lots (Δ-16,924), [09-29] net long 78,058 lots (Δ-4,464), [09-22] net long 82,522 lots (Δ+17,416); managed money net longs have fallen for three consecutive weeks from a late-September high of 82,528 lots to 61,058 lots, consistent in direction with the 20-day price decline of -2.09%, with no clear divergence. COT crowding (as-of [10-06], CFTC latest report date 12 days ago): netPct 20.11%, crowding 61.66 (3-year percentile), CTA 98, hedge 56.9%; the CTA trend-following proxy is in the extreme zone at 98, and while crowding has not reached the |z|>1.5 extreme line, the elevated CTA reading implies crowded trend strategies and a higher risk of a liquidation stampede on reversal. Implied versus realized volatility: copper has no direct IV data, so referencing ^GVZ (gold implied volatility) [10-16] at 22.58 (1Y percentile 12%) and ^OVX (WTI implied volatility) [10-16] at 50.74 (1Y percentile 48%), while SHFE copper RV20 is 16.1% and LME copper RV20 is 16.7%; cross-asset IV percentiles show metal options are priced cautiously overall, with no significant event-risk premium evident.
V. Cross-Asset & Onshore-Offshore Value
On cross-asset relative value, the copper/gold ratio [10-16] is 1.57, at the 92.46th 1Y percentile and 54.23rd 3Y percentile, a 1-year high; copper is relatively strong versus gold, suggesting growth expectations are already well priced, and further upside requires fundamental support. The oil/gold ratio [10-16] is 0.0243, at the 100th 1Y percentile and 55.29th 3Y percentile, with crude at a 1-year extreme versus gold, a strong inflation-transmission signal from energy to copper. The gold/silver ratio [10-16] is 72.96, at the 87.7th 1Y percentile and 29.37th 3Y percentile, with silver relatively weak versus gold. International versus China market comparison (like-for-like): SHFE.CU versus LME.CU shows a China tax-adjusted premium of +1.28% [10-16] (-0.03% five days earlier), flipping from a discount to a premium, with the China screen relatively expensive versus the international market and the import window marginally open; the Yangshan bill average of USD 130/t ([10-16]) is the import cost anchor, and the positive premium combined with stable bill premiums means import P&L estimates are for reference only and do not form a conclusion. On term structure, the international curve is in CONTANGO with M1-M2 at -0.0295 (-0.45%), a near-month discount that corroborates the China warrant accumulation. Risk metrics (retrospective): DD52w 13.49%, DD20d 4.36%, Sharpe30 -0.3858, VaR95 -2.75%; the 20-day max drawdown is 4.36% and the 30-day risk-adjusted return is negative, so the short-term risk/reward is weak. Seasonality (same-day start, 20 trading days, last 15 years): average +0.44%, median +0.77%, up in 8 of 15 years, best 2016 at +16.76%, worst 2015 at -8.51%; background only, limited sample.
VI. Historical Statistics & Seasonality
Seasonal statistics for the same date plus the subsequent 20 trading days over the past 15 years show: average gain of +0.44%, median of +0.77%, up in 8 of 15 years, for a win rate of approximately 53.3%; the best year was 2016 at +16.76%, the worst was 2015 at -8.51%. The sample window is limited, so this serves only as background reference and does not constitute a directional basis. From a distribution standpoint, the median exceeds the mean, indicating that a few extreme years (e.g., 2016 at +16.76%) pull the average upward, while most years' 20-day performance clusters in a moderate range; however, the -8.51% tail drawdown in 2015 likewise cautions that seasonality cannot rule out significant down years. Combined with this report's bearish directional conclusion, the seasonal signal is inconsistent with the current direction: the historical same-period win rate leans slightly bullish, whereas current drivers (SHFE warehouse warrant accumulation, top-20 seats net short 29,471 lots, main contract rising on open interest cuts) point to pressure. Accordingly, this report does not use seasonality as a trading basis, but only as auxiliary background for the bullish scenario in the scenario analysis. Note that the seasonal statistics are based on a historical same-period window, do not incorporate the current year's inventory cycle or positioning-structure differences, and the 15-year sample is small with limited statistical significance; investors should not build positions solely on this basis.
VII. Scenario Analysis
Base case (probability 55%, bearish bias, consistent with Section I): Trigger conditions are SHFE warehouse warrants continuing to accumulate ([10-16] already up to 36,386 tonnes, +1,453 tonnes d/d) and top-20 seats net short maintained or expanded (currently net short 29,471 lots); price tests pivot S1 at 108300 from the [10-16 settlement] of 109450, and after breaking below S1 at 108300 targets 107660 (the established short target). If LME inventories shift from drawdown to consecutive builds and SHFE warehouse warrants keep rising sharply, the bearish-invalidation condition is triggered (see Section I) and the base case yields to a neutral stance. Operationally: short per the established strategy, entry 109860, stop-loss 111001, target 107660, horizon 1-5 days, risk/reward 1:1.9; or entry 110190, stop-loss 111331, target 107910, horizon 1-2 weeks, risk/reward 1:2.0.
Bullish scenario (probability 20%): Trigger condition is a daily close above 111100 (below the 20-day high of 111770), i.e., one of the invalidation conditions described in Section I is triggered, while LME inventories continue to draw down (cumulative -29,675 tonnes over the last 20 periods) and SHFE warehouse warrants fall back; in that case price opens upside space above 110190 (R1), and after breaking 111100 targets 111770 (the 20-day high level), with further reference to the 20-day high excluding the current day at 112330. Operationally: exit all bearish positions and move to neutral; wait for a daily close confirmed above 111100 before assessing any reversal; do not chase the rally.
Bearish scenario (probability 25%): Trigger condition is an effective break below S1 at 108300, with SHFE warehouse warrants still rising and LME inventory drawdown slowing or turning to builds; downside space then opens, with price approaching 108300 and targeting 107660 (the established target). If a resonance emerges of large consecutive LME builds and sustained sharp increases in SHFE warehouse warrants, the bearish-invalidation condition is triggered and the directional view must be reassessed. Operationally: on top of the base-case short position, add only per established rules if price breaks below 108300 and the position risk budget allows; do not add other entry levels. The three scenario probabilities sum to 100%.
VIII. Trading Strategies & Risk Management
Established strategy (quoted verbatim): Short entry 109860 / stop-loss 111001 / target 107660 / horizon 1-5d / risk-reward 1:1.9; short entry 110190 / stop-loss 111331 / target 107910 / horizon 1-2w / risk-reward 1:2.0. Stop-loss logic: both 111001 and 111331 sit below the 20-day high of 111770, and 111100 is the invalidation trigger level described in Section I; placing stops above this zone ensures that once the daily-close break above 111100 invalidation condition is met, the position has already been exited, avoiding enlarged losses from a directional misjudgment; meanwhile, the stop distance matches ATR, keeping noise-trigger probability manageable. Risk conversion (quoting program output): bearish 109860, stop distance 1141 = 1×ATR (ATR14 1140.71), contract multiplier 5, risk per lot CNY 5,815 (including fees and slippage), single-trade risk budget CNY 66,853 corresponding to a reference cap of 11 lots; bearish 110190, stop distance 1141 = 1×ATR (ATR14 1140.71), contract multiplier 5, risk per lot CNY 5,815 (including fees and slippage), single-trade risk budget CNY 66,853 corresponding to a reference cap of 11 lots. Position sizing and add/trim rules: the single-trade risk budget of CNY 66,853 is the cap; combined risk of the two shorts must not exceed the established multiple of that budget; after entry, if price runs toward target and unrealized profit covers more than 1×ATR (1141 points), the stop may be moved to entry per established rules; do not add before S1 at 108300 is broken, and after the break add only within the risk budget per established rules; exit immediately upon hitting the stop, no averaging down, no adding to losers. Horizon management: for the 1-5 day trade, close if target is not reached by expiry; for the 1-2 week trade, halve the position if price has not moved out of the cost zone (within ±0.5×ATR) after 5 trading days.
IX. This Week's Data Calendar
Key data this week (Beijing time): 10-19 09:30 new home prices m/m; 10-19 10:00 energy production, fixed-asset investment ytd/y, household income, industrial capacity utilization, industrial value-added y/y, real estate development investment ytd/y, retail sales y/y, services production index y/y; 10-20 09:15 1-year Loan Prime Rate (prior 3%), 5-year Loan Prime Rate (prior 3.5%); 10-20 09:30 GDP y/y. These releases affect ferrous metals, nonferrous metals, and chemicals, with fixed-asset investment, industrial value-added, retail sales, and GDP flagged as high-priority items.
This report is generated automatically from public quantitative and macro data for research and market tracking only. It does not constitute any investment advice.