Weekly Gold and Silver Update 25 July 2006

-

 EXECUTIVE SUMMARY

This week delivered a compelling but contradictory picture. Gold rose 1.8% to $4,055.70 and silver surged 4.6% to $58.49 — silver's outperformance compressing the Gold/Silver Ratio from 71.3 to 69.3, a notably bullish signal for silver relative to gold. The GBDT forecasting model responded with sharply higher median forecasts across all horizons, and gold's 60-day probability of gain jumped to 79.7%. Yet under the surface, the macro environment deteriorated on every front.

The DXY-Metals Nexus Monitor composite scores deepened materially — Gold to −0.87 and Silver to −0.86, from −0.73/−0.72 last week — as all five signal modules held at Strong Reduce. Real TIPS yields continued to rise: the 10-year reached 2.43% (+8bp) and the 5-year climbed to 2.17% (+13bp), both firmly in Strong Headwind territory. The week's most significant signal shift: Gold's 60-day Skew Score flipped from Bullish Lean (+0.09) to Bearish Lean (−0.12), revealing that even as the median forecast rose sharply, the model sees materially more downside risk than upside over the two-month horizon. Silver's skew picture is more resilient, maintaining Bullish Lean at the 20-day horizon.

Note on Real Yields: Real Yields= Nominal Interest Rate minus Rate of Inflation. Meaning even if interest rates rise, but rate of inflation is higher than interest rates, then the real yield is negative-which meansthat investors might as well hold non-yield assets like gold and silver. 

Note on Gold-Silver/USD Correlation;. Normally when USD (as measured by DXY Index ) goes up, Gold-Silver goes down. But Gold-Silver positive correlation can turn negative in periods of high geopolitical tensions, and also when investors expect fiat currencies to be debased by high inflation. 

It's too much of a hassle for me to post the full analysis here. So I only show the key charts  and tables. If you need text and commentary to understand or gain more insights,  email me (tiankhean@gmail.com)  and I will send to you the full Report in .docx format. Always glad to meet more people who are into AI and statistical modeling of financial markets.

Market Snapshot



Monte Carlo Simulation Probabilities




DXY Monitor

Real Yields Monitor


Pinball Loss by Quantile


Skew Score



Comments

Post a Comment

Popular posts from this blog

Spot Silver Forecast for 27 Oct to 21 Nov 2025

-
Image
 A big Thank You to EODHD.com for the data without which this post could not have been written. The Close of Silver (XAGUSD) as of Friday 24 Oct 2025 is 48.62.The 21 business days ahead forecast for spot investment grade Silver bullion is summarized thus:  10% chance of going below 48.32 50% chance of being 54.33 10% chance of being above 60.80 Methodology We prefer to apply probabilistic quantile forecasting due to the highly unstable situation as the MAGA Man keeps changing his policies almost daily, and even just the words he uses generates  asymetric shocks to the model that renders straight forecasts useless.  Model input variables:  10yr Treasury yield US Dollar Index Spot Gold price S&P500 US Unemployment Rate Methodology Methodology summary (plain-language bullets) Data we used Daily end-of-day prices from EODHD for Silver (XAGUSD) and Gold (XAUUSD). Macro/market context series (Dollar Index, 10-year Treasury yield, S&P 500, unemployment). We rei...
Read more

Markov Regime Switching Model for Risk‑On/Risk‑Off Dashboards of Stock Indices

-
Image
Knowing the market’s current risk state is foundational for portfolio decisions. Basic Markov chain models are a good starting point, but they fall short on three realities of financial data. Using as case study China's CSI300 Index, here’s how our model resolves each one and delivers a robust, reproducible Risk‑On/Risk‑Off signal. 1) Beyond “memoryless”: markets have persistence and aging Problem: Standard Markov chains assume memorylessness—the next state depends only on the current state, not on how long we’ve been there. Markets don’t behave that way; regimes “age” and persistence changes with time. Our solution: We use a hidden semi‑Markov model (HSMM) with explicit duration modeling. Each regime has a duration distribution (e.g., Negative Binomial), and transitions are allowed to depend on time spent in the state. This breaks the memoryless assumption, curbs flicker, and captures realistic regime persistence. 2) Heavy‑tailed, skewed returns are the norm—not Gaussian Problem: ...
Read more