⚠ Accuracy & Interpretation Notes
- Period-specific average: The 10-year mean ratio of ~81 reflects the 2016–2026 window, which was an unusually elevated period. Longer-term historical averages (20yr+, since 1970s) typically sit in the 50–70 range. Context matters when framing what is "normal."
- Closure rate: All 17 episodes in this dataset closed by the end of the observation window — but this is partly a function of dataset length and threshold definition. Mean reversion is a tendency, not a guarantee; ratios can remain elevated for extended periods (e.g., 2020 peaked at 125.89 and took 4+ months to reverse).
- Rarity claims: The current low-ratio + large price drawdown combination is unusual within this 10-year window, but does not constitute a never-before-seen event in the broader multi-decade history of precious metals.
- No backtested strategy: This report documents historical ratio patterns only. No trading strategy has been simulated or validated. Actual trading involves transaction costs, slippage, and the risk of prolonged adverse ratio moves not captured here.
- Silver drawdown: The ~39% decline from the Jan/Feb 2026 high ($115.08) to current ($70.32) is a post-rally correction within an ongoing structural bull market, not necessarily a return to prior equilibrium levels.
2016–2026 Mean Ratio
80.9
Elevated vs. 20yr avg (~65)
All-Time Peak (10yr)
125.9:1
2020-03-18 (COVID panic)
Current Ratio
64.4
2026-03-30 | -1.79σ vs 10yr avg
Gold Latest
$4,526
10yr range: $1,127–$5,318
Silver Latest
$70.32
Off 39% from Jan 2026 high
Elevated Episodes
17
All closed within dataset
Gold-Silver Ratio — Monthly Average (2016–2026)
1Summary Statistics
Descriptive statistics for the gold/silver price ratio — 2,553 trading days from Feb 2016 to Mar 2026.
Current (2026-03-30)
64.36
The mean of 80.9 is specific to the 2016–2026 window — a period of structurally elevated ratios driven by the 2020 COVID spike and sustained industrial-metal headwinds for silver. The broader 20-year or post-1970s average is approximately 60–70. Claims about what is "historically normal" should specify the time horizon.
2Significant Divergence Episodes
Episodes where the ratio exceeded the 75th percentile (86.2) for ≥5 consecutive trading days. Sorted by peak ratio.
Largest Gap — COVID-19 Panic (2020-03-18): 125.89:1
Gold held at $1,477 while silver collapsed to $11.73 — dollar spread of $1,465.57. This exceeded the prior modern-era high of ~97 (1991 Gulf War). Silver was sold as an industrial metal in a liquidity panic; gold held as safe-haven. The gap took 125 calendar days to fully close — demonstrating that extreme divergences can persist for months before reversing.
Second-Largest — Tariff Shock (2025-04-21): 104.82:1
Gold surged to $3,406 on safe-haven demand while silver lagged on industrial demand concerns from potential tariff-induced slowdowns. Closed in 93 calendar days by July 2025.
| Episode Start | Peak Date | Peak Ratio | Entry |
Gold @ Peak | Silver @ Peak | Closed Date | Cal Days | Status |
| 2020-01-09 | 2020-03-18 |
125.89 | 86.88 |
$1,477.30 | $11.73 |
2020-07-21 | 125d |
CLOSED |
| 2024-12-13 | 2025-04-21 |
104.82 | 86.64 |
$3,406.20 | $32.50 |
2025-07-23 | 93d |
CLOSED |
| 2022-06-28 | 2022-09-01 |
96.67 | 87.39 |
$1,696.60 | $17.55 |
2022-09-14 | 13d |
CLOSED |
| 2019-05-06 | 2019-07-05 |
93.63 | 86.27 |
$1,396.70 | $14.92 |
2019-07-24 | 19d |
CLOSED |
| 2023-02-24 | 2023-03-10 |
91.39 | 86.94 |
$1,862.00 | $20.38 |
2023-03-14 | 4d |
CLOSED |
| 2023-12-29 | 2024-02-07 |
91.35 | 86.46 |
$2,035.20 | $22.28 |
2024-02-16 | 9d |
CLOSED |
| 2022-10-11 | 2022-10-14 |
91.13 | 86.41 |
$1,641.70 | $18.01 |
2022-10-24 | 10d |
CLOSED |
| 2024-02-20 | 2024-02-28 |
90.72 | 87.75 |
$2,033.00 | $22.41 |
2024-03-13 | 14d |
CLOSED |
| 2019-07-25 | 2019-08-06 |
89.75 | 86.54 |
$1,472.40 | $16.41 |
2019-08-27 | 21d |
CLOSED |
| 2024-08-30 | 2024-09-06 |
89.67 | 86.80 |
$2,493.50 | $27.81 |
2024-09-12 | 6d |
CLOSED |
| 2024-03-19 | 2024-04-01 |
89.62 | 86.38 |
$2,236.50 | $24.95 |
2024-04-03 | 2d |
CLOSED |
| 2024-08-05 | 2024-08-14 |
89.51 | 88.70 |
$2,439.40 | $27.25 |
2024-08-15 | 1d |
CLOSED |
| 2022-09-23 | 2022-09-29 |
89.11 | 87.33 |
$1,658.50 | $18.61 |
2022-10-03 | 4d |
CLOSED |
| 2019-11-29 | 2019-12-06 |
88.53 | 86.37 |
$1,459.10 | $16.48 |
2019-12-20 | 14d |
CLOSED |
| 2025-09-15 | 2025-09-17 |
88.25 | 86.61 |
$3,681.80 | $41.72 |
2025-09-19 | 2d |
CLOSED |
| 2024-11-21 | 2024-11-27 |
87.67 | 86.44 |
$2,639.90 | $30.11 |
2024-12-03 | 6d |
CLOSED |
| 2019-11-07 | 2019-11-12 |
87.07 | 86.28 |
$1,452.10 | $16.68 |
2019-11-19 | 7d |
CLOSED |
3Gap Closure Analysis
Time elapsed from each episode's peak ratio back to its entry-level ratio. "Closure" means the ratio returned to where the elevated episode began — not to any broader historical mean.
Closed Within Dataset
17 / 17
Median Closure Time
9 days
All 17 episodes closed within this dataset's observation window. This does not mean every future episode will close — or close quickly. The 100% figure is a property of this specific 10-year sample and threshold definition. The two macro-driven events (2020, 2025) took 93–125 days, demonstrating that "snap back" is not always fast or guaranteed. Markets can remain dislocated longer than expected.
Fastest Closure: 1 Day
Peak 2024-08-14, ratio 89.5. Reversed the following session — consistent with a positioning spike rather than a structural move. Not representative of typical behavior.
Slowest Closure: 125 Days
COVID peak 2020-03-18, ratio 125.9. Required the full industrial-demand reopening trade. Silver did not meaningfully recover until July 2020 — four months after the peak divergence.
4Closure Time Distribution
Duration buckets across all 17 closed episodes in this dataset.
| Time to Close | Count | Share | Distribution |
|---|
| ≤7 days | 8 |
47.1% |
|
| 8–30 days | 7 |
41.2% |
|
| 31–90 days | 0 |
0.0% |
|
| 91–180 days | 2 |
11.8% |
|
| >180 days | 0 |
0.0% |
|
Bimodal Pattern — Fast or Very Slow, Rarely In-Between
88% of episodes in this dataset closed within 30 days; the remaining 12% took 90–125 days. There are zero episodes in the 31–90 day range — suggesting that either the market corrects quickly (routine noise) or the divergence is macro-driven and takes a full cycle shift to unwind. This distribution is based on 17 events; small sample sizes can produce misleading pattern statistics.
5Seasonal & Temporal Patterns
Average gold/silver ratio by calendar month (2016–2026):
| Month | Avg Ratio | Visual |
|---|
| Jan | 78.11 |
|
| Feb | 79.31 |
|
| Mar | 81.32 |
|
| Apr | 83.82 |
|
| May | 83.43 |
|
| Jun | 82.09 |
|
| Jul | 81.53 |
|
| Aug | 80.64 |
|
| Sep | 80.71 |
|
| Oct | 81.01 |
|
| Nov | 80.36 |
|
| Dec | 78.82 |
|
April–May Average Premium
Apr (~83.8) and May (~83.4) show the highest monthly averages — about 5 points above the Dec–Feb lows. Likely reflects gold's seasonal demand window and silver's industrial softness in early spring. These are averages across 10 years; individual years show significant variation.
Dec–Feb Compression
December (~78.8) and January (~78.1) show the lowest average ratios — silver tends to benefit from industrial/tech demand planning and year-end risk-on positioning.
September Peak Clustering
4 of 17 divergence peaks occurred in September — more than any other single month in this dataset. Late-summer liquidity and Q4 positioning may be recurring catalysts, though 17 events is too small a sample to treat this as a reliable seasonal signal.
6Annual Ratio Snapshot
Year-by-year min, max, and average ratio, plus year-end prices.
| Year | Min Ratio | Max Ratio | Avg Ratio |
YE Gold | YE Silver | YE Ratio |
| 2016 | 65.64 |
83.51 | 72.96 |
$1,150.00 | $15.94 |
72.16 |
| 2017 | 67.72 |
79.58 | 73.98 |
$1,306.30 | $17.06 |
76.57 |
| 2018 | 75.70 |
86.58 | 81.16 |
$1,278.30 | $15.43 |
82.83 |
| 2019 | 79.95 |
93.63 | 86.27 |
$1,519.50 | $17.83 |
85.23 |
| 2020 | 69.09 |
125.89 | 88.70 |
$1,893.10 | $26.33 |
71.89 |
| 2021 | 63.30 |
81.95 | 71.80 |
$1,827.50 | $23.33 |
78.34 |
| 2022 | 74.56 |
96.67 | 83.27 |
$1,819.70 | $23.86 |
76.26 |
| 2023 | 76.47 |
91.39 | 83.18 |
$2,062.40 | $23.85 |
86.46 |
| 2024 | 72.67 |
91.35 | 84.77 |
$2,629.20 | $28.94 |
90.85 |
| 2025 | 56.48 |
104.82 | 87.82 |
$4,325.60 | $70.13 |
61.68 |
| 2026 | 44.14 |
66.48 | 59.02 |
$4,526.00 | $70.32 |
64.36 |
Secular Expansion (2016–2024) then Sharp Contraction (2025–2026)
The annual average ratio rose steadily from ~73 (2016–17) to ~88 (2024) as silver underperformed. That trend reversed sharply: 2025 averaged ~87.8 but ended at 61.7, and 2026 year-to-date averages ~59 with a current level of 64.4. The Jan 2026 low of 44.1 was the lowest in this dataset — but ratios in the 40s have appeared in longer historical records (e.g., 2011 silver bull market reached ~32, and 2020 briefly touched ~63 during recovery).
7Key Findings
1. COVID-19 Created the Largest Divergence in This Dataset
The 125.89 ratio on 2020-03-18 was the highest recorded in this 10-year window. Silver traded at $11.73 while gold held at $1,477. The $1,465.57 dollar spread was the widest in modern history for this dataset. The gap took 125 days to close — not the rapid snap-back seen in smaller episodes.
2. The 2025 Tariff Shock Was the Second Structural Divergence
The April 2025 spike to 104.82 was gold-price-led (surging above $3,400) rather than silver-crash-driven. Silver was not at historically cheap absolute prices — the high ratio reflected gold's extreme premium from geopolitical safe-haven demand. Closed in 93 days.
3. Current Ratio (64.4) Is a 10-Year Low — Within This Elevated Window
At -1.8σ below the 2016–2026 mean, the ratio is the lowest recorded in this dataset. However, the 2016–2026 mean itself (~81) was elevated by the 2020 spike and sustained industrial headwinds. Compared to longer-term history (20yr avg ~65), the current ratio of ~64 is below average but not extreme. The silver drawdown of ~39% from the Jan 2026 high ($115.08 → $70.32) is a confirmed, significant correction within an ongoing bull market.
4. Mean Reversion Is a Historical Tendency, Not a Guarantee
High-ratio episodes in this dataset showed strong mean-reversion tendencies — 88% resolved within 30 days, and all 17 eventually closed. However, "mean reversion" does not specify timing, direction of the move (gold down vs. silver up), or magnitude. The ratio can remain dislocated for months (as in 2020), and the relevant "mean" itself is period-dependent. These patterns are informative for context but should not be interpreted as predictive signals.