I. The Instinct to Aggregate
In financial data science, aggregation is often treated as a virtue. If a metric can be computed for every listed contract, the natural instinct is to average them and derive a “total market” score.
In Week 6, we deliberately resisted this impulse. The Numerical Stability Index (NSI) was defined over a disciplined 15–45 Days to Expiration (DTE) lens. In Week 7, we showed that the ultra-short boundary (0DTE) is structurally hostile by geometric necessity and cannot serve as a macro baseline.
We now formalize the mathematical hazard underlying full-surface aggregation.
Without boundary discipline, aggregation induces Expiry Dilution. Rather than improving accuracy, full-surface averaging produces a scalar that does not faithfully represent computational topology.
II. Long-End Asymptotics
Recall the Stability Signal:
\[ s = \frac{\text{Pricing Error} \cdot \text{Vomma}}{\text{Vega}^2}. \]
To understand why long-dated contracts distort aggregation, we examine the asymptotic behavior of this expression as time to expiration \( T \) grows large.
In classical option pricing frameworks:
- Vega scales proportionally with \( \sqrt{T} \).
- Therefore, the denominator \( \text{Vega}^2 \) scales proportionally with \( T \).
- Vomma, while model-dependent, does not grow proportionally with \( T \); its magnitude remains bounded relative to the squared slope.
Substituting these proportional relationships into the Stability Signal:
\[ s \propto \frac{\text{Error} \cdot \text{Bounded Growth}}{T}. \]
Figure 14. Asymptotic scaling of stability components. - As time to expiration increases, Vega² grows proportionally with ( T ), while curvature terms remain bounded. The Stability Signal therefore decays asymptotically. Near expiration, slope collapse amplifies curvature interaction.
As \( T \to \infty \), the denominator dominates.
The Stability Signal is forced toward zero.
Long-dated contracts therefore occupy extremely wide monotonic basins. Their local geometry is insulated from curvature compression by the sheer magnitude of Vega².
This is not a heuristic claim. It is an asymptotic property of the model.
III. The Mechanism of Expiry Dilution
An options chain typically contains a large population of long-dated contracts.
Because their Stability Signal is asymptotically suppressed, they consistently reside in the Stable regime, even during severe macroeconomic shocks.
At the same time, the active, gamma-dense region of the surface (15–45 DTE) can experience sharp regime compression:
- Pricing error interacts violently with curvature.
- Contracts migrate into Oscillatory and Divergent regimes.
- Path Stress increases materially.
If these two regions are averaged together without distinction, the geometric inertia of the long-dated contracts overwhelms the curvature compression of the active region.
Mathematically:
- The majority of contracts contribute near-zero \( s \).
- A minority contribute large \( s \).
- The aggregate mean converges toward the inert asymptotic region.
The resulting metric suppresses the very instability it is intended to measure.
This is Expiry Dilution.
IV. Averaging Incompatible Regimes
The listed options surface is not geometrically homogeneous. It contains structurally incompatible regions:
- 0DTE: A boundary condition where Vega collapses and curvature dominates.
- 15–45 DTE: The active institutional zone, sensitive to macro compression.
- Long-Dated LEAPS: An asymptotic regime where Vega² suppresses curvature.
Averaging across these regimes produces a scalar without consistent geometric interpretation.
The result is not a measure of structural integrity.
It is a weighted mixture of incompatible asymptotics.
Without boundary discipline, the metric loses meaning.
V. Empirical Demonstration
The effect of Expiry Dilution is observable.
When the full surface NSI is computed during the Q1 2020 regime inversion and compared against the disciplined 15–45 DTE benchmark, the difference is clear.
Figure 15. Expiry Dilution Under Full-Surface Aggregation. - The 15–45 DTE benchmark captures structural compression during Q1 2020, while full-surface aggregation exhibits reduced sensitivity due to long-dated asymptotic inertia. The disparity reflects geometric dilution rather than economic divergence.
During March 2020:
- The 15–45 DTE NSI declined materially, reflecting regime compression.
- The full-surface metric exhibited significantly reduced amplitude.
The geometric inertia of long-dated contracts suppressed the signal.
The full-surface metric behaves as a structural low-pass filter, masking curvature turbulence in the active region.
VI. Measurement Requires Boundary
Stability measurement must align with geometric domain.
To track computational topology accurately, one must define:
- The region of interest,
- The asymptotic regime being measured,
- The structural interactions under consideration.
The NSI benchmark enforces this discipline by restricting aggregation to the 15–45 DTE lens.
Measurement without boundary produces incoherent metrics.
Over the past eight essays, we have built a complete diagnostic framework:
- Latent vs observed stability,
- Stability Signal derivation,
- Finite-precision distortions,
- Path Stress accumulation,
- Surface-wide structural compression,
- Expiry Dilution asymptotics.
Diagnosis is now complete.
But diagnosis alone does not govern computation.
If a solver can identify that it is entering a Divergent or Oscillatory regime before iteration begins, how should it respond?
Measurement without execution is passive.
Governance without measurement is blind.
In the next essay, we transition from diagnosis to execution: Governance vs. Heuristic Damping.