7-Day vs 14-Day vs 30-Day Volatility Horizons: How Forecast Window Selection Shapes Metals Risk Management

The forecast horizon you select for a volatility signal is not a minor parameter choice. It determines the risk you can identify, the actions you can take, and the decay profile of the information embedded in the prediction. A 7-day volatility probability and a 30-day volatility probability are answering fundamentally different questions, and treating them interchangeably introduces misalignment between signal and execution.

The Volterra model produces probability forecasts at all three horizons: 7-day, 14-day, and 30-day. Each serves a distinct function in metals risk management. Understanding the structural differences between them is a prerequisite for extracting value from any multi-horizon signal set.

7-Day Horizon: Tactical Positioning and Near-Term Event Response

The 7-day window captures fast-moving, event-driven volatility. News flow, inventory drawdowns, exchange notices, and geopolitical disruptions all register most sharply at this horizon. 7-day volatility forecasts are most useful for options desks adjusting gamma exposure and for systematic traders managing short-dated positions.

At the 7-day horizon, the signal-to-noise ratio from alternative data sources peaks. GDELT-derived news intensity, for example, decays rapidly; a spike in coverage of Indonesian nickel export policy carries predictive weight for the next five trading days but attenuates well before the 30-day mark. The Volterra pipeline processes 96 GDELT GKG files daily, and much of that information's predictive contribution concentrates in the 7-day window. For a deeper look at how news data enters the model, see our analysis of how GDELT news flow becomes a mineral volatility signal.

The practical implication: 7-day forecasts should drive decisions with short execution cycles. Adjusting short-dated vol surface nodes, sizing weekly option positions, or triggering intraday hedging overlays all map naturally to this horizon.

14-Day Horizon: The Bridge Between Tactical and Structural

The 14-day window occupies a space where both event-driven and structural factors contribute. It captures the persistence of supply disruptions, the follow-through from regulatory announcements, and the propagation of logistics bottlenecks into physical premiums. 14-day volatility forecasts balance news-driven signal strength with enough forward horizon to reflect slower-moving supply chain dynamics.

For risk managers running weekly VaR recalibrations, the 14-day horizon provides a natural fit. 14-day volatility forecasts align well with the rebalancing cadence of most institutional hedging programs. It is long enough to capture second-order effects, such as how a port closure in a concentrated supply corridor propagates into futures term structure, but short enough that the forecast retains actionable precision.

The Volterra model's walk-forward cross-validation methodology, detailed in our post on walk-forward validation in commodity ML, is calibrated separately at each horizon. This means the 14-day model is not simply a blurred version of the 7-day model; it is trained on its own target distribution with its own optimal feature weighting.

30-Day Horizon: Strategic Risk Budgeting and Procurement Exposure

The 30-day window addresses a different audience and a different decision cadence. At this horizon, event-driven news signal has largely decayed. What remains is the structural backdrop: geographic supply concentration, inventory-to-consumption ratios, term structure shape, and macroeconomic context. 30-day volatility forecasts are most relevant for strategic risk budgeting, procurement hedging, and portfolio-level allocation decisions.

30-day volatility forecasts reflect structural supply concentration risk more than shorter horizons. A commodity with high Herfindahl-Hirschman Index scores, such as cobalt with its heavy concentration in the DRC, will exhibit persistent elevated risk at the 30-day level even when 7-day signals are quiet. The structural factors that drive 30-day predictions move slowly but carry larger magnitude when they do shift.

For procurement teams and treasury desks managing monthly or quarterly hedging mandates, the 30-day horizon is the primary input. It governs the sizing of forward hedges, the strike selection on protective puts, and the risk budget allocated to specific mineral exposures.

Using Multiple Horizons Together

The real value of a multi-horizon signal set lies in divergence. When a mineral shows LOW at 7-day but ELEVATED at 30-day, that pattern suggests building structural risk without a near-term catalyst. The reverse pattern, ELEVATED at 7-day and LOW at 30-day, typically indicates an event-driven spike that the model expects to dissipate.

The Volterra dataset delivers all three horizons simultaneously across 12 exchange-traded critical minerals, enabling desks to construct horizon-weighted risk composites tailored to their specific execution cadence. Options desks may weight the 7-day signal at 60% and the 30-day at 10%, while a procurement team inverts that weighting entirely. For a full view of how these volatility probability signals integrate into vol surface positioning, we have published a dedicated framework.

Figures from the Volterra daily pipeline. Full historical backfill available on AWS Data Exchange. The complete multi-horizon dataset covers all 12 minerals across LME, COMEX, NYMEX, and SGX, with five risk levels per horizon per mineral delivered daily.