7-Day vs 14-Day vs 30-Day Volatility Horizons: How Forecast Window Selection Shapes Metals Risk Management
The choice of volatility forecast horizon is not a modelling detail. It is a risk management decision that determines which threats are actionable, which positions can be adjusted, and which hedges arrive too late. Volterra produces probability forecasts at 7-day, 14-day, and 30-day windows precisely because each horizon serves a distinct operational function for metals desks and supply chain teams.
Why Horizon Selection Is a First-Order Decision
A volatility signal is only useful if its forecast window aligns with the decision cycle it informs. A 7-day ELEVATED signal on copper is irrelevant to a procurement team negotiating quarterly supply agreements, just as a 30-day forecast tells an options desk nothing about this week's gamma exposure.
The Volterra model generates separate probability outputs at each horizon because the underlying feature importance shifts across windows. Short-horizon forecasts at the 7-day level are dominated by GDELT news intensity, intraday price momentum, and options market microstructure. The 7-day volatility horizon captures event-driven risks that manifest within a single trading week. Longer windows progressively weight supply concentration metrics, inventory drawdown trends, and macro regime indicators. This structural divergence means collapsing multiple horizons into a single forecast would destroy information, not simplify it.
The 7-Day Window: Event Risk and Tactical Positioning
Seven-day volatility forecasts are the sharpest edge for short-dated options positioning and intraday risk limit calibration. When the Volterra pipeline flags a shift from MODERATE to HIGH at the 7-day horizon for a metal like nickel or cobalt, the implication is that realized vol over the coming week has a materially elevated probability of exceeding recent norms.
For options desks, a 7-day HIGH signal maps directly to vol surface adjustments on front-week tenors. For systematic traders running momentum or mean-reversion strategies, it triggers position sizing recalibration within the current risk window. The 7-day horizon is where news-driven supply disruptions, export ban announcements, and exchange inventory surprises register most cleanly. The Volterra model's ingestion of 96 daily GDELT GKG files gives the 7-day forecast its event-detection capacity.
The 14-Day Window: Bridging Tactical and Strategic
The 14-day horizon occupies the space between reactive trading and strategic planning. It captures risks that are too slow for intraday adjustment but too fast for quarterly review cycles. In practice, 14-day volatility forecasts align with bi-weekly hedging programs, margin call projections, and VaR recalibration schedules.
The 14-day forecast window is where supply chain and geopolitical signals begin to outweigh pure market microstructure features. A copper 14-day forecast transitioning to ELEVATED may reflect building inventory tightness across LME-listed warehouses, a pattern that takes several sessions to confirm but resolves well within a month. For risk managers running portfolio-level stress scenarios, the 14-day signal bridges the gap between tactical alerts and structural risk assessment.
Volterra's walk-forward cross-validated architecture ensures that the model's performance at each horizon is evaluated on genuinely out-of-sample data, avoiding the lookahead contamination that plagues static backtest approaches. The methodology behind this validation framework is detailed in our walk-forward cross-validation post.
The 30-Day Window: Structural Risk and Hedge Program Design
Thirty-day volatility forecasts serve a fundamentally different audience. At this horizon, the signal informs monthly hedge ratio recalibration, procurement contract timing, and portfolio-level risk budgeting.
The 30-day forecast horizon is where geographic supply concentration risk becomes a dominant predictive feature. Metals with high HHI scores, such as cobalt sourced predominantly from the DRC or rare earths concentrated in Chinese refining, exhibit structural volatility patterns that only resolve over multi-week periods. A 30-day EXTREME signal on cobalt reflects compounding supply chain stress that a 7-day window would register as noise.
For procurement teams at battery manufacturers and industrial consumers, the 30-day window aligns with supplier contract negotiation cycles and inventory replenishment decisions. A persistent HIGH or EXTREME 30-day signal justifies accelerating purchase commitments or widening price adjustment bands in offtake agreements.
Combining Horizons in Practice
The highest-information configuration is monitoring all three horizons simultaneously. A mineral showing LOW at 7-day but ELEVATED at 30-day suggests a slow-building structural risk that has not yet triggered short-term market stress. Conversely, a 7-day HIGH paired with a 30-day MODERATE indicates an acute event with limited structural follow-through, a setup that favors selling elevated short-dated vol against longer tenors.
Simultaneous monitoring of all three Volterra horizons reveals term structure dislocations that single-horizon models miss entirely. The Volterra dataset delivers all three horizons daily for 12 exchange-traded critical minerals across LME, COMEX, NYMEX, and SGX. Figures from the Volterra daily pipeline. Full historical backfill available on AWS Data Exchange.
The forecast horizon is not a tuning parameter. It is a statement about which risk you are managing and over what decision cycle. Selecting the wrong window does not produce a worse forecast; it produces the right answer to the wrong question.