I. Bottom Line & Directional Bias
Directional conclusion: bullish. SHFE Gold AU2612 saw active long open interest build at the 21.8th percentile of the 20-day range; [10-12 settlement] CNY 907.78, +0.71%, main contract open interest 228,792 lots (+334 lots), top-20 seats net long 164,243 lots (daily +3,846 lots), with the driver shifting from short covering to active long positioning. Internationally, GC=F settled at USD 4,190.6, Asian early session USD 4,193.37 (+0.07%), China tax-adjusted premium +0.4%, onshore and offshore aligned in direction. GVZ sits at the 12th percentile of its 1-year range, IV−RV at +5.3 volatility points, with options pricing event risk as US CPI, core CPI and retail data land densely over the next two days. Core rationale: first, price sits at the 21.8th percentile of the 20-day range of 889.82–961.88, and the rise on open interest build after a 20-day decline of −6.26% reflects active longs rather than short covering; second, total open interest is at the 95th percentile of the past year, top-20 seats net long 164,243 lots with a daily increase of 3,846 lots, so flows and price repair are aligned; third, the international Asian early session at USD 4,193.37 is above the settlement of USD 4,190.6, with a China premium of +0.4% and no basis divergence between international and China markets. Invalidation: a daily close below 901.83 (S1) turns the view neutral, with downside to 889.82 (20-day low); an effective break above 911.11 (R1) opens upside to 926.
II. Price Action & Technicals
Technicals are anchored to the [10-12 settlement] of CNY 907.78: close CNY 905.52, +0.71% versus prior settlement, 5-day +0.1%, 20-day −6.26%, 20-day range 889.82–961.88, position at the 21.8th percentile, ATR14 at CNY 14.85 (1.64%, full intraday range), RV20 at 18.5%. Pivot P is 907.41, R1 is 911.11, S1 is 901.83; the settlement of 907.78 sits slightly above P, but the gap between R1 and settlement is only CNY 3.33, less than half an ATR (CNY 7.43), so R1 is not treated as a trade level and only S1 serves as the stop-loss reference. The last completed weekly candle (2026-10-08 to 10-09) opened at 893.1, high 911.22, low 889.82, close 908.64, weekly change −0.21%; this week (from 2026-10-12, one session so far) is not yet complete, tentatively 905.52 (−0.34%), so no weekly conclusion is drawn. The last five daily candles show a low of 889.82 probed on 10-08 followed by two consecutive rebounds; on 10-12 the high was 913 but it closed at 905.52, with a relatively long upper shadow. China night session (10-12 21:05 to 10-13 02:30) opened at 901.28, high 903.9, low 900.26, close 902.9, −0.29% versus the day session close and −0.54% versus settlement, stabilizing after retesting above S1. Internationally, GC=F [10-12 settlement] USD 4,190.6, 1-day −0.61%, 5-day +0.81%, 20-day −3.71%, ATR14 USD 80.19 (1.91%), RV20 17.5%, 20-day range 4091.2–4439.8 at the 28.5th percentile, 52-week range 3901.3–5586.2; Asian early session (10-13 07:00) USD 4,193.37 (+0.07%), high 4199.9, low 4193, with the international market stabilizing above the lower end of the 20-day range, in line with the pace of China's repair.
III. Supply-Demand Balance & Fundamental Drivers
On supply and demand, inventories show no pressure: SHFE Gold warrants stand at 116,028 kilograms [10-12], daily change 0 (+0%), prior [10-09] also 116,028 kilograms, flat for consecutive sessions; COMEX Gold registered warrants at 15.07 Moz (468,730 kilograms) [10-08], daily change 0 (+0%), prior [10-07] unchanged. On the ETF side, GLD holdings are 33,858,916 ounces [10-09] versus 33,778,539 ounces 20 trading days ago, a change of +0.24%, with allocation accounts adding modestly as prices fell, consistent with the headline direction of “increased China buying after price declines”. On basis, China's tax-adjusted premium to GC=F is +0.4% [10-12] versus +0.53% five days earlier; the premium has narrowed but remains positive, the import window is inclined to be open, and China is relatively expensive versus the international market without divergence. On term structure, the international M1-M2 is −19.2 (−0.45%), RY is −5.44%, slope is 17.677, with the front month at a discount and a positive curve slope, showing no sign of a squeeze in the cost-of-carry structure. Macro transmission: the US 10-year Treasury yield is 5.24% [10-12], daily change −0.08%, and the dollar index is 102.57 [10-12], daily change +0.35%; dollar strength and lower yields offset each other, limiting net pressure on gold. Over the next two days, US CPI, core CPI, PPI and retail data land densely; expectations are for core CPI at 2.5% y/y (prior 2.4%) and CPI at 3.6% y/y (prior 3.4%); an inflation upside surprise would pressure SHFE Gold via the dollar channel, while a downside surprise would reinforce the low-end repair logic. On seasonality, over the past 15 years the same-date 20-trading-day average is +0.39%, median +0.91%, up in 9 of 15 years, best 2016 at +4.36%, worst 2014 at −5.34%; this is background only, with a limited sample.
IV. Positioning & Fund Flows
Positioning is bullish but crowded. Top-20 seats [10-12]: total open interest 222,611 lots (+3,458), longs 58,368 lots (−388), shorts 164,243 lots (+3,846), net direction is net short 105,875 lots (daily +3,846 lots), with the expanding net long scale aligned with the price repair. Main contract notional open interest is CNY 207.18bn, daily change +CNY 302m, open interest +334 lots and price +0.71%, a rise on open interest build (active longs) rather than short covering. The trigger is open interest at the 95th percentile of the past year; high positioning means volatility will amplify once direction is confirmed, while a break below S1 could accelerate declines as positions are unwound. Internationally, COT over the last four weeks: 10-06 total open interest 396,109, longs 127,101, shorts 16,114, net long 110,987 (Δ−9,331); 09-29 net long 120,318 (Δ−7,071); 09-22 net long 127,389 (Δ−5,727); 09-15 net long 133,116 (Δ−1,856), with managed money net positions declining for four consecutive weeks, aligned with the 20-day price change of −3.71% and showing no persistent divergence between price and positioning. On crowding, 10-06 netPct is 28.02%, crowding 90.8 (3-year percentile), CTA 62, hedge 14.45%; crowding has exceeded 92 for four consecutive weeks, CTA is stable at 62, and trend-following accounts remain long but are cutting at the margin. On volatility, GVZ is 22.79 [10-09], daily change +0.11 points, 1-year percentile 12%; GC=F RV20 is 17.5%, IV−RV is +5.3 volatility points (IV/RV 1.3), with options pricing event risk and implied volatility significantly above realized volatility; if IV falls after the data, it will pressure the cost of long positions.
V. Cross-Asset Relative Value & Global vs China Markets
On cross-asset relative value, the gold-silver ratio stands at 69.25[10-12], at the 77.38th percentile over 1 year and the 25.79th percentile over 3 years; the ratio sits at the high end of its 1-year range, with silver relatively weak. Under an alternative measure, the gold-silver ratio is 69.07, at the 74.6th percentile over 1 year and the 24.87th percentile over 3 years, leading to the same conclusion. The copper-gold ratio is 1.61[10-12], at the 99.6th percentile over 1 year and the 61.51st percentile over 3 years; the copper-gold ratio is at an extremely high 1-year percentile, reflecting strong growth expectations and diverting relative allocation away from gold. The oil-gold ratio is 0.0224[10-12], at the 95.24th percentile over 1 year and the 50th percentile over 3 years; the oil-gold ratio is at a 1-year high, with energy relatively strong versus gold. The Crack 3:2:1 crack spread is USD 64.55/bbl[10-12], with no direct transmission to gold, serving only as macro context. On the dollar and rates: DXY is 102.58[10-12], +0.35% d/d; the 10-year UST yield is 5.24%[10-12], −0.08% d/d; VIX is 15.21[10-12], +0.37 points d/d, at the 16th percentile over 1 year, with risk appetite stable. On global vs China markets, SHFE.AU trades at a +0.4%[10-12] premium to GC=F on a China tax-adjusted basis, versus +0.53% five days earlier; the premium is positive and narrowing, with the China market relatively expensive versus the international market and the import window tending to open; freight and premium/discount are not included, and this serves only as a proxy for import profit/loss. Internationally, the Asian early-session price of 4,193.37 USD is above the settlement of 4,190.6 USD, with the international and China markets moving in the same direction and no basis divergence. Risk indicator review: DD52-week is 23.87%, DD20-day is 6.42%, Sharpe30 is −3.336, VaR95 is −2.94%; recent risk-adjusted returns are negative, indicating that drawdown risk remains during the low-level repair process.
VI. Historical Statistics & Seasonality
Seasonal statistics for the same date over the past 15 years, covering the subsequent 20 trading days, show: an average gain of +0.39%, a median gain of +0.91%, up years in 9 of 15 years, and a win rate of 60%; the best year was 2016 at +4.36%, the worst was 2014 at −5.34%, with a sample range of nearly 10 percentage points and high dispersion. By way of contrast, AU2612's current 20-day change is −6.26%, a significant deviation from the seasonal median path, with the price at the 21.8th percentile of the 20-day range of 889.82–961.88, marking a statistically low-level repair starting point rather than the middle of a trend. In terms of distribution shape, the median (+0.91%) exceeds the mean (+0.39%), indicating that a few deep-decline years (such as 2014) pulled down the mean, while the typical year shows a mild uptrend; this pattern offers weak support for the current “low-level bullish” view, but the sample is only 15 years and a single window, so it has no significance-test value. On positioning, this issue's trigger condition is open interest at the 95th percentile over the past year, with the top 20 seats net long 164,243 lots (+3,846 lots d/d); high open interest combined with a seasonally positive bias means that if the direction is confirmed, the volatility amplification effect will be stronger than normal; conversely, if the daily close breaks below 901.83 (S1), seasonal support offers no protection, and the 20-day low of 889.82 becomes the next observation level. Internationally, GC=F's 20-day change is −3.71%, at the 28.5th percentile, in line with the China market's pace; the seasonal conclusion serves only as background and is not a standalone basis for opening positions.
VII. Scenario Analysis
Base case (probability 55%, bullish): the trigger is a daily close holding above 901.83 (S1), with the 10-14 US core CPI y/y landing in the 2.4%–2.6% range (i.e., within the expected 2.5% ± 0.1% threshold), and the dollar index and 10-year UST yield (currently 102.58 and 5.24%) not rising in tandem; the target is an effective break above 911.11 (R1), then 927.79, about 2.2% above the [10-12 settlement] of 907.78, with the structure of a night-session pullback to 900.26 followed by stabilization and a main-contract open-interest build on the rise (open interest +334 lots, change +0.71%) continuing. Operate per the established strategy: go long, entry at 905.52, horizon 1–2 weeks. Bull case (probability 20%): the trigger is core CPI y/y below expectations (actual ≤ 2.4%, i.e., more than 0.1 percentage point below expectations), with the dollar and UST yields falling, combined with increased China buying after the price pullback (10-12 headline direction consistent) and continued allocation flows with GLD holdings +0.24% over 20 days; the target is a break above 927.79, then 961.88 (20-day high), about 7.5% above the [10-12 settlement]. Operate by adding after an effective break of 911.11 (R1), moving take-profit up in tranches, and closing if 901.83 (S1) is breached. Bear case (probability 25%): the trigger is core CPI y/y above expectations (actual ≥ 2.6%) pushing the dollar and UST yields higher, or a daily close below 901.83 (S1); the target is 889.82 (20-day low), and if lost, the next downside is below 889.18 (the strategy stop-loss level). Operate by not shorting actively, executing stop-loss only on existing longs; if IV falls after the event (GVZ currently at the 12th percentile over 1 year, IV−RV at +5.3 volatility points), the cost pressure on long holdings rises, so reducing positions should take priority over adding. The three scenarios sum to 100%.
VIII. Trading Strategies & Risk Management
Established strategy (quoted verbatim): go long, entry 905.52, stop-loss 889.18, target 927.79, horizon 1–2 weeks, risk/reward 1:1.4. The stop-loss at 889.18 is set about 0.64 yuan below the 20-day low of 889.82, i.e., below the next technical defense line after S1 (901.83) is lost; the logic is that if the daily close breaks below 901.83 (S1), the bullish structure fails and the price will test the 20-day low of 889.82, so placing the stop below that low avoids being shaken out by spikes at the lower edge of the range while limiting the loss per trade. In risk terms: the stop distance is 16.34 yuan, equal to 1.1× ATR (ATR14 is 14.854), with a contract multiplier of 1,000, the risk per lot is CNY 16,390 (including fees and slippage); a single-trade risk budget of CNY 66,941 corresponds to a reference cap of 4 lots. Position rules: cap at 4 lots, with the first tranche recommended at no more than half the cap (i.e., 2 lots), adding up to 2 more lots only after an effective break of 911.11 (R1) and core CPI landing without exceeding expectations; if the price reaches the 927.79 target, realize part or all of the position per the 1:1.4 risk/reward, and move the stop up on the remainder using 901.83 (S1) as the trailing-stop reference. Event risk control: if the 10-14 20:30 core CPI y/y prints ≥ 2.6% (more than +0.1 percentage point above expectations), cut the position by half regardless of whether the stop is hit, to guard against gaps driven by a synchronized rise in the dollar and UST yields; with GVZ at the 12th percentile over 1 year and IV−RV at +5.3 volatility points, options are paying for event risk, and the post-release IV decline will pressure the cost of long holdings, so adding before the data is inadvisable. Total open interest is at the 95th percentile over the past year, with the top 20 seats net long 164,243 lots (+3,846 lots d/d); with high open interest, volatility amplifies once the direction reverses, so stop-loss execution must be rigid and the stop must not be moved down.
IX. This Week's Data Calendar
Key data this week (Beijing time): 10-13 16:03, China M2 money supply y/y (expected 7.4%, prior 7.5%, threshold ±0.1%) and new loans (expected CNY 1025B, prior CNY 60B); 10-14 09:30, China CPI y/y (expected 1.0%, prior 0.8%, threshold ±0.2%) and PPI y/y (expected 4.4%, prior 3.8%, threshold ±0.6%); 10-14 10:03 and 10:04, China trade balance (expected CNY 790B, prior CNY 809B) and USD-denominated trade balance (expected USD 116.8B, prior USD 119.1B); 10-14 20:30, US core CPI m/m (expected 0.2%, prior 0.3%), core CPI y/y (expected 2.5%, prior 2.4%, threshold ±0.1%), CPI m/m (expected 0.6%, prior 0.4%), and CPI y/y (expected 3.6%, prior 3.4%); 10-15 20:30, US core PPI m/m, core retail sales m/m, PPI m/m, and retail sales m/m; 10-16 11:30, Fed Chair Warsh speaks. Among these, the 10-14 20:30 core CPI y/y is the key event for this report's invalidation condition, with an actual value beyond ±0.1% of expectations treated as a surprise.
This report is generated automatically from public quantitative and macro data for research and market tracking only. It does not constitute any investment advice.