Two-Moment Decision Model¶
A decision model that ranks uncertain alternatives using two moments, usually mean and variance, of their outcome distributions.
Core Idea¶
Two-moment models compress each uncertain alternative to expected outcome and dispersion. The choice rule then trades the first moment against variance or standard deviation, as in mean–variance portfolio analysis.
Compression is an assumption, not a theorem of all decisions. Distributions with the same first two moments can differ in skewness or tails, and the moments do not rank alternatives without a stated preference or dominance rule.
Scope of Application¶
- Finance. Constructs mean–variance portfolios.
- Economics. Models choices under summarized uncertainty.
- Operations research. Builds tractable risk–reward optimization.
- Decision analysis. Tests sufficiency of distributional summaries.
Clarity¶
State alternatives, outcome horizon, probability model, exact two moments, units, covariance treatment, and preference rule. Test whether tail or asymmetry invalidates the compression. Inclusion test: Include decision rules whose sufficient inputs are two declared moments of each uncertain consequence. Exclusion test: Exclude full-distribution expected utility, mean-only ranking, robust worst-case choice, and labels that report two statistics without using them to decide. Nearest boundary: A mean–CVaR model also uses two summaries but is not a two-moment model because tail risk is not a distributional moment. Exit condition: The class ends when a third moment, full density, or non-moment tail measure is essential to ranking. Common misclassifications: It is not full-distribution expected utility. It is not mean-only optimization. It is not any model with two inputs. It is not automatically sensitive to tail risk. Nearest named distinctions: Mean–CVaR model: Uses a tail risk measure, not two moments. Expected utility: Integrates utility over the full distribution. Descriptive statistics: Does not become a decision model until linked to choice. Two-criterion optimization: Criteria need not be moments of uncertainty.
Manages Complexity¶
For two-moment decision model, separating Alternatives from Outcome variable exposes the first dependency. Relating Expected value to Distributional limits then prevents the observed two-moment decision model outcome from replacing its defining mechanism.
Abstract Reasoning¶
- For two-moment decision model, fix Alternatives and its units or identity.
- Establish how Outcome variable functions inside two-moment decision model from cited evidence.
- Test Expected value directly instead of inferring two-moment decision model from resemblance.
- Map Variance or deviation to the defining two-moment decision model relation.
- Use Distributional limits to challenge the closest alternative to two-moment decision model.
- Report the two-moment decision model boundary, uncertainty, and surviving conclusion.
Knowledge Transfer¶
Mean–dispersion reasoning transfers when preferences and distribution families make the first two moments decision-sufficient. A ranking cannot transfer to heavy-tailed or asymmetric settings without checking omitted risks.
Relationships to Other Abstractions¶
Current abstraction Two-Moment Decision Model Domain-specific
Parents (1) — more general patterns this builds on
-
Two-Moment Decision Model presupposes Risk–Return Tradeoff Prime
Two-Moment Decision Model presupposes Risk–Return Tradeoff because the model jointly ranks uncertain alternatives by expected outcome and dispersion.
Hierarchy paths (4) — routes to 4 parentless roots
- Two-Moment Decision Model → Risk–Return Tradeoff → Trade-offs → Constraint
- Two-Moment Decision Model → Risk–Return Tradeoff → Risk → Uncertainty
- Two-Moment Decision Model → Risk–Return Tradeoff → Risk → Probability → Measure → Set and Membership
- Two-Moment Decision Model → Risk–Return Tradeoff → Risk → Probability → Measure → Aggregation → Micro Macro Linkage
Neighborhood in Abstraction Space¶
Two-Moment Decision Model sits in a crowded region of the domain-specific corpus (17th percentile for distinctiveness): several abstractions share nearly its structure, so a description that fits it tends to fit its neighbors too.
Family — Decision & System Modeling Frameworks (30 abstractions)
Nearest neighbors
- Chance-Constrained Programming — 0.92
- Probability matching — 0.91
- Non-Consequential Reasoning — 0.91
- Silverman's game — 0.91
- ÉLECTRE — 0.90
Computed from structural-signature embeddings · 2026-10-08