Blackwell–Girshick Equation¶
A two-term identity separating the variance of an independent-count random sum into mark-size and count-uncertainty contributions.
Core Idea¶
For a nonnegative integer random count \(N\) independent of iid marks \(X_i\), let \(S=\sum_{i=1}^{N}X_i\) and set \(S=0\) when \(N=0\). If \(N\) and \(X_i\) have finite second moments, the Blackwell–Girshick equation is
Conditioning on \(N\) separates variability in individual mark sizes from variability in the number of marks. This is exact within the stated model, not an approximation and not a tail-probability formula. A Poisson count is one special case, not a requirement.[ref-1902ecbf8aa9][ref-8fd359d4b000]
Scope of Application¶
Daniel's health-insurance example has 200 expected claims over ten days and independent exponential claim sizes with mean 500 dollars. The mark-size and count terms each equal $50$ million dollars squared, yielding aggregate variance $100$ million dollars squared. Vatn applies the same roles to cumulative equipment-failure costs: \(N(t)\) counts failures, \(V_i\) is cost per failure, and the total cost has variance \(\mathbb E[N(t)]\operatorname{Var}(V)+\operatorname{Var}(N(t))(\mathbb E[V])^2\). Both are model-based applications, not claims of universal independence in real claims or failures.[ref-b7aa87e4f48a][ref-057a38e4160f]
Correlated marks, count-dependent mark sizes, or missing second moments can invalidate the displayed equality. In particular, the broader random-sum formula in Cohen's paper has extra structure and should not be silently substituted for the classical two-term equation.[^ref-1902ecbf8aa9]
Clarity¶
The equation says that an aggregate can vary because events differ in size, because the number of events differs, or both. The two terms have the same squared units as \(\operatorname{Var}(S)\). Wald's companion equation \(\mathbb E[S]=\mathbb E[N]\mathbb E[X_1]\) concerns only the mean; neither equation alone determines extreme-event probabilities.[^ref-8fd359d4b000]
Manages Complexity¶
Instead of constructing the full compound distribution, one can calculate variance from the count's mean and variance and the mark's mean and variance. This compression supports comparisons between frequency and severity contributions. It deliberately discards distributional shape, so tail or quantile questions need additional assumptions and analysis.[^ref-b7aa87e4f48a]
Abstract Reasoning¶
Given \(N=n\), the sum has mean \(n\mu\) and variance \(n\sigma^2\) for mark mean \(\mu\) and variance \(\sigma^2\). Averaging conditional variance gives \(\mathbb E[N]\sigma^2\); varying the conditional mean gives \(\operatorname{Var}(N)\mu^2\). Their sum is the formula. If the conditional mark mean or covariance changes with \(n\), redo the conditional calculation rather than reusing this right-hand side.[ref-1902ecbf8aa9][ref-8fd359d4b000]
Knowledge Transfer¶
The insurance and equipment examples share a literal count–mark–sum mapping despite differing carriers. The proposed DAG parent is Decomposition, because the formula exactly reconstructs variance from two distinguished contributions; its strict probability-theory accent is the independent random-count model. Before applying it in a new field, verify iid marks, independence from the count and finite moments. The 1947 Blackwell–Girshick attribution is reported by Cohen; its original theorem text was not directly checked in this package.[ref-1902ecbf8aa9][ref-b7aa87e4f48a][^ref-057a38e4160f]
[^ref-1902ecbf8aa9]: Joel E. Cohen, “Sum of a Random Number of Correlated Random Variables that Depend on the Number of Summands”, The American Statistician 73(1) (2019), §4.1 Eq. (13). [^ref-b7aa87e4f48a]: James W. Daniel, Poisson Processes actuarial notes, Fact 1.11 and Example 1.12. [^ref-057a38e4160f]: Jørn Vatn, Counting Processes, NTNU notes, “Compound HPPs.” [^ref-8fd359d4b000]: Rick Durrett, Probability: Theory and Examples, §2.3 excerpt, random-sum variance theorem and proof.
Relationships to Other Abstractions¶
Current abstraction Blackwell–Girshick Equation Domain-specific
Parents (1) — more general patterns this builds on
-
Blackwell–Girshick Equation is a kind of Decomposition Prime
A random-sum variance is exactly decomposed into mark and count components.
Hierarchy path (1) — routes to 1 parentless root
- Blackwell–Girshick Equation → Decomposition
Neighborhood in Abstraction Space¶
Blackwell–Girshick Equation sits in a sparse region of the domain-specific corpus (80th percentile for distinctiveness): few abstractions share its structure, so a faithful description tends to retrieve it precisely.
Family — Foundations of Probability & Inference (29 abstractions)
Nearest neighbors
- Kolmogorov's Three-Series Theorem — 0.84
- Yule–Simon Distribution — 0.83
- Average Order of an Arithmetic Function — 0.82
- Large Set (Combinatorics) — 0.82
- Random Variable — 0.82
Computed from structural-signature embeddings · 2026-10-08