I. Price Action & Technicals
As of the 2026-09-18 close, the SHFE Silver main continuous contract settled at 16,304.00 CNY/kg, up 4.0659% on the day, the largest single-day gain in the past 5 trading sessions; daily volume was 165,934 lots and open interest 215,014 lots, with volume and open interest expanding in tandem, indicating this rally has been accompanied by incremental capital inflows rather than pure short covering. Looking back over the past 5 sessions: September 14 closed at 15,528.00 (-0.7605%), September 15 at 15,452.00 (-1.9045%), September 16 at 15,832.00 (+2.4062%), September 17 at 15,593.00 (-0.3833%), and September 18 reclaimed the losses of the prior three sessions in one move, forming a classic “probe lower—consolidation—volume breakout” structure.
On the moving average system, MA5 on September 18 was 15,741.80 and MA20 was 16,184.45; the close of 16,304.00 was above both MA5 and MA20, the first return above the 20-day moving average since September 14, with short-term moving averages beginning to turn from declining to flat. Medium-term resistance references the 20-day high of 17,388.00, and short-term support references the 20-day low of 15,285.00, a range width of roughly 2,103 CNY, corresponding to an amplitude of about 13.7%, indicating the market is in a broad range rather than a one-way trend. On the pivot system, the September 18 pivot P was 16,188.33, R1 was 16,436.67, and S1 was 16,055.67; the close of 16,304.00 sits between P and R1, in the stronger zone above the pivot; if R1 (16,436.67) is subsequently broken effectively, the upside opens toward the 17,000 round number and the prior high of 17,388; if S1 (16,055.67) is lost and price falls back below P (16,188.33), this breakout carries false-breakout risk.
On oscillators, <data> did not provide specific readings for SHFE Silver RSI, MACD, or ATR; data pending update; for reference, the international contract SI=F has an ATR of 1.6376 USD, 20-day volatility Vol20 of 34.0753%, and VaR95 of -6.2818%, indicating silver's intraday volatility is significantly higher than gold's and position management should treat it as a high-volatility instrument. On the weekly and monthly timeframes, <data> did not provide the corresponding period close series; data pending update. Overall assessment: the daily timeframe has shifted from weak to neutral-to-strong, but a trend reversal is not yet confirmed; the key validation levels are 16,436.67 (R1) and 16,184.45 (MA20).
II. Fundamental Drivers
On the macro rates side, the 2026-09-19 US 10-year Treasury yield (^TNX) was 4.9980, up 1.03% on the day, approaching the 5% psychological threshold. High real rates weigh on the valuation of non-yielding silver, and this is the largest macro headwind facing the current rebound. On the dollar side, the 2026-09-20 US Dollar Index (DX-Y.NYB) was 100.2540, up a marginal 0.03% on the day, oscillating above 100; the dollar has not shown a trend weakening, meaning USD-denominated silver lacks additional FX tailwind. If Treasury yields subsequently fall back below 5% and the dollar breaks below 100, the macro pressure on silver would ease markedly; conversely, if yields break above 5.1%, precious metals as a whole would come under pressure.
On the inventory side, 2026-09-18 SHFE silver warrants were 1,407,381.00 kg, a weekly change of +9,612.00 kg, showing modest inventory build. COMEX registered inventory on 2026-09-17 was 3,026,059.52 kg, with the weekly change data missing; data pending update. The coexistence of SHFE inventory build and the price rebound suggests this rally is driven more by capital and expectations than by physical tightness, and the divergence signal of “price up, inventory up” warrants caution.
On the term structure side, the international contract comparison shows a CONTANGO structure, with M1-M2 at -0.2750 (-0.4111%), roll yield (RY) at -4.9333%, and slope at 0.2948. A positive market structure with near-month discount and far-month premium typically reflects that spot supply is not tight, carrying costs are positive, and it is unfavorable for longs rolling positions. On the ETF and central bank gold-buying dimension, <data> did not provide silver ETF holdings or central bank silver purchase data; data pending update; on the geopolitical dimension, <data> the past 48h domestic-related headlines are N/A, no usable sentiment input; data pending update. Overall, fundamentals present a neutral-to-bearish combination of “macro pressure + modest inventory build + positive market,” and the sustainability of the price rebound requires further confirmation from the capital side.
III. Positioning & Capital Flows
The international COT positioning (comparison basis) shows that on 2026-09-15 total open interest OI was 103,745 lots, longs L were 20,205 lots, shorts S were 7,081 lots, net long net was 13,124 lots, with a weekly change Δ of -1,262 lots, i.e., net longs decreased versus the prior week. Looking back four weeks: August 25 net was 14,073 (Δ+2,378), September 1 net was 12,598 (Δ-1,475), September 8 net was 14,386 (Δ+1,788), September 15 net was 13,124 (Δ-1,262); net longs oscillated repeatedly in the 12,500–14,400 range, lacking trend accumulation, with the latest week turning to a position cut.
On crowding, 2026-09-15 net long share netPct was 12.6502%, crowding crowd was 55.04, CTA positioning 62.00, and hedge share hedge 25.8114%. Versus September 8's netPct of 13.9332% and crowd of 58.85, both net long share and crowding fell back, indicating long crowding has eased marginally and has not reached extreme crowding thresholds, but incremental long momentum is also lacking. CTA positioning held unchanged at 62.00 for four consecutive weeks, indicating trend-following capital is in a wait-and-see state and has not made directional additions to the current rebound.
On the domestic side, SHFE Silver main contract open interest rose from 199,708 lots on September 14 to 215,014 lots on September 18, a cumulative increase of about 15,306 lots over 5 days, combined with the September 18 volume expansion to 165,934 lots, domestic capital participation has clearly increased. On the options and volatility dimension, the CBOE Silver Volatility Index ^VXSLV on 2026-09-19 was 40.9800, -5.09% on the day; volatility has retreated from highs but remains at an absolute high above 40; VIX was 14.8100, -4.08%, with overall risk appetite stable. If VXSLV continues to fall below 35, it favors price trading time for space; if it surges back above 45, beware of large single-day drawdowns.
IV. Cross-Asset Relative Value
The gold-silver ratio (GC_SI_RATIO) last printed 66.4800, with a 1-year percentile P1Y of 51.19% and a 3-year percentile P3Y of 17.06%. The ratio sits at the mid-point of the 1-year range and at a low level in the 3-year range, meaning that from a three-year perspective silver remains relatively cheap versus gold; if the precious metals complex strengthens as a whole, silver has the relative-value basis for catch-up outperformance; however, the 1-year percentile of only 51.19% indicates the short-term ratio has returned to neutral, and further downside (i.e., silver relatively stronger) requires incremental drivers.
On the oil-gold ratio and copper-gold ratio, <data> did not provide the corresponding price series for crude oil, copper, and gold; data pending update, and this section cannot provide quantitative percentiles for now. Alternative reference information available: the combination of a US Dollar Index at 100.2540 and a 10-year Treasury at 4.9980 typically corresponds to a pricing environment with weak industrial attributes and dominant financial attributes, which is unfavorable for silver's industrial demand expectations but provides some support for its monetary attribute.
On cross-market spreads, the international contract SI=F closed at 66.8250 USD, and the domestic main contract closed at 16,304.00 CNY/kg; the two moved in the same direction (September 18 international +3.53%, domestic +4.07%), with the domestic gain slightly larger than the international, indicating stronger domestic sentiment. The combination of a CONTANGO term structure and RY of -4.9333% means calendar spread arbitrage (buy near, sell far) does not have an advantage in carrying costs, and reverse arbitrage must bear negative roll yield. Overall relative-value conclusion: the low 3-year percentile of the gold-silver ratio is the core relative-value argument for medium-term bullish silver, but short-term cross-validation from oil and copper ratios is lacking; data pending update.
V. Sentiment
<data> the past 48h domestic-related headlines (market=CN, asset=SILVER) are N/A, with no usable news sample; the sentiment score and headline bias cannot be quantified; data pending update. Observable alternative sentiment indicators: the CBOE Silver Volatility Index ^VXSLV at 40.9800 (-5.09%), with falling volatility typically corresponding to easing panic; VIX at 14.8100 (-4.08%) indicates stable macro risk appetite. The domestic market's September 18 volume-driven gain of 4.0659% with rising open interest, the price-volume behavior itself reflects a warming of short-term sentiment. If domestic silver-related headlines subsequently turn concentratedly bullish, beware of reversal volatility after sentiment overheating; if headlines remain absent, price is more likely to be driven by the international market and macro data.
VI. Historical Statistical Patterns
<data>No seasonal return series for SHFE Silver over the same period (late September) of the past 10 years is provided, nor are monthly or quarterly seasonality statistics available; historical statistical pattern data is pending update. The limited statistics available for reference are: international market SI=F 5-day change +5.21, 20-day change -3.80, indicating the past week was stronger than the past month, consistent with a “short-term rebound, medium-term still weak” pattern; DD52w at 51.4268% and DD20d at 8.9684%, indicating the current price still has roughly 51% drawdown room from the 52-week high and about 8.97% from the 20-day high, with medium-term recovery potential exceeding short-term. Sharpe30 at 2.1103, with risk-adjusted returns over the past 30 days positive, but this metric is based on the international market reference basis and cannot be directly extrapolated to SHFE Silver. Until seasonal data is completed, it is not advisable to use historical same-period patterns as a basis for positioning.
VII. Long-Short Game Deduction
Bullish arguments: (1) On September 18, SHFE Silver closed at 16304.00, up +4.0659% on the day, with volume of 165934 lots and open interest building to 215014 lots, showing volume-price-open interest resonance; (2) The closing price of 16304.00 reclaimed above MA20 (16184.45), improving the short-term moving average structure; (3) The gold-silver ratio at 66.48 sits at the 17.06% low percentile over 3 years, with silver relatively undervalued versus gold; (4) International market SI=F cumulative 5-day gain of +5.21%, with overseas markets strengthening in tandem to form resonance; (5) COT net long share at 12.6502%, with crowding at 55.04 down from the prior week, releasing long crowding risk.
Bearish arguments: (1) The 10-year U.S. Treasury yield at 4.9980 approaching 5%, with high real rates suppressing non-yielding assets; (2) The dollar index at 100.2540 holding above 100, offering no support from the FX side; (3) SHFE inventory at 1407381 kg, up 9612 kg week-on-week, a divergence of rising prices alongside rising inventories; (4) Term structure in CONTANGO, RY -4.9333%, with roll costs unfavorable to longs; (5) COT net long at 13124 lots, Δ-1262 lots, marginal position reduction in the latest week; (6) Price still about 6.2% away from the 20-day high of 17388.00, with the medium-term trend not yet reversed.
Short-to-medium-term balance of forces: Short-term (1-5 days) bulls have the upper hand, as the volume breakout and moving average reclaim create momentum; medium-term (1-3 months) bearish factors are more concentrated, with macro rates, inventories and the positive-market structure constituting persistent suppression. If 16436.67 (R1) is effectively broken and VXSLV falls below 35, the short-term bullish case can extend into the medium term; if 16055.67 (S1) is lost, this rebound should be characterized as a technical repair.
VIII. Trading Strategies & Risk Management
Strategy One (short-term trend-following, LONG): Entry reference zone 16304.00-16436.67 (closing price to R1), stop-loss set below 16055.67 (S1), target the 17000 round number, horizon 1-5 trading days, position size recommended no more than 10% of total capital, conviction 6. Trigger condition is a daily close holding above MA20 (16184.45) with volume no less than 70% of September 18's volume.
Strategy Two (range fade, SHORT): If price rallies into the 17000-17388.00 (20-day high) zone and shows volume expansion with stalling gains, a light short can be attempted, entry reference near 17000, stop-loss set above 17388.00, target back to 16184.45 (MA20), horizon 1-10 trading days, position size no more than 8% of total capital, conviction 5.
Risk management points: Silver Vol20 at 34.0753% and VaR95 at -6.2818%, with significant single-day drawdown risk; single-trade risk exposure for any strategy is recommended to be kept within 2% of total capital; if the dollar index breaks above 101 or the 10-year U.S. Treasury breaks above 5.1%, long strategies should proactively reduce positions; if the weekly increase in SHFE inventories continues to expand, long conviction should be lowered. All levels are based on the <data> basis, and live execution requires combining with real-time order flow.
IX. This Week's Data Calendar
Domestic financial calendar for the next 7 days (all MEDIUM level): 2026-09-23 03:15 France manufacturing and services Flash PMI; 03:30 Germany manufacturing and services Flash PMI; 04:30 UK manufacturing and services Flash PMI. The above data mainly affect the euro, sterling and the dollar index, and then transmit to silver via the dollar channel. Domestically, <data> provides no other events this week; data pending update. Key focus on the reaction of the dollar index at 100.2540 and the 10-year U.S. Treasury at 4.9980 following the data releases.
This report is generated automatically from public quantitative and macro data for research and market tracking only. It does not constitute any investment advice.