Expected Value Calculation¶
Method — instantiates Probabilistic Risk Weighting
Multiplies or otherwise combines probability and consequence on a common scale to rank options by expected gain, loss, or exposure.
Expected Value Calculation is the pure combining rule of risk weighting. It puts probability and consequence onto one common scale and multiplies them, collapsing a spread of uncertain outcomes into a single number — the expected gain, loss, or exposure — so that options can be ranked and the best one chosen. Its distinctive move is commensuration: every outcome is expressed in the same unit (usually money or a utility score) and reduced to one figure per option, which is what makes the choice legible and rankable. That is also its notorious blind spot. Because expected value is risk-neutral by construction, it weights a certain small loss and a tiny chance of ruin identically whenever their expected values happen to match — the reason it must hand catastrophic tails to a different tool.
Example¶
A manufacturer has discovered an intermittent defect in a finished product batch and must choose between two options: ship now, or hold the batch for a rework. Shipping carries a probability of a field failure — illustratively 3% per unit — each failure costing an estimated $2,000 in recall, warranty, and handling. Reworking costs a certain amount per unit up front, with no failure risk.
The calculation puts both options on the common scale of dollars per unit. Ship: 3% × $2,000 = an expected $60 of failure cost per unit. Rework: a certain, known cost per unit. Expected Value Calculation ranks the two by that single figure and points to the cheaper expected path — ship if the expected failure cost is below the rework cost, rework if not. The number is defensible and auditable, and it turns an argument about intuition into an argument about two inputs. But note the boundary: if a field failure could injure someone — a catastrophic, irreversible outcome — the expected-cost ranking is precisely the wrong instrument, because it would trade a small chance of harm against dollars as if they were the same currency.
How it works¶
- Put everything on one scale. Translate each outcome's consequence into a common unit — money, utility, quality-adjusted life years — whatever the decision runs on.
- Weight by probability. Multiply each outcome's value by its probability of occurring.
- Sum per option. Add the probability-weighted values to get one expected figure for each option under consideration.
- Rank and choose. Order the options by expected value and select the highest expected gain or lowest expected loss.
Tuning parameters¶
- Common scale / unit — what everything is converted into. The choice quietly reveals or hides incommensurable harms; a dollar scale silently prices safety and dignity.
- Utility curvature — linear (strictly risk-neutral) versus a concave utility that penalizes large losses more than proportionally. Adding curvature is the standard partial fix for expected value's indifference to variance.
- Outcome resolution — a few lumped outcomes versus a fine distribution; finer resolution is more faithful but more expensive to build and to defend.
- Probability source — where the probabilities come from (a forecast, a base rate). The calculation is only as trustworthy as the inputs it multiplies.
When it helps, and when it misleads¶
For repeated, comparable, survivable decisions it is the correct and efficient rule. It makes trade-offs explicit, rankable, and auditable, and across many independent bets it maximizes the long-run total — exactly what you want when no single outcome can sink you.
It misleads when the downside is not survivable, because it is silent on variance and ruin. Averaging a catastrophic, irreversible, or non-compensable outcome into an expected figure treats a 1% chance of disaster as interchangeable with a certain 1% loss; the St. Petersburg paradox is the classic demonstration that raw expected value, taken alone, can prescribe choices no sane actor would make.[n1] It also invites false precision and consequence-scale collapse — forcing safety, reputation, and money onto one axis so the incommensurable disappears. The classic misuse is running it on a bet-the-company or safety-critical decision where a single bad draw is unrecoverable. The guarding discipline is to reserve it for comparable, survivable risks and, when the downside turns catastrophic or irreversible, hand off to a tail-sensitive rule rather than trusting the average.
How it implements the components¶
consequence_estimate— its first step is to express each outcome's consequence on a common value scale; the valuation is part of the calculation, not an input it merely borrows.risk_weighting_rule— the probability-times-consequence combination is the weighting rule, collapsing likelihood and impact into one expected figure per option.action_priority_mapping— ranking options by expected value maps directly to the choice: pick the highest expected gain or the lowest expected loss.
It combines and ranks but does not itself produce the probabilities it multiplies (probability_estimate) — those come from Probabilistic Forecast or the Actuarial Risk Model. And, being risk-neutral, it pointedly does not give catastrophic tails special treatment (tail_risk_exception_rule); that is the Actuarial Risk Model's catastrophe load — its nearest twin among these methods — which loads for the very tail this rule averages away.
Related¶
- Instantiates: Probabilistic Risk Weighting — it is the combine-and-rank core that turns weighted risks into a choice.
- Consumes: Probabilistic Forecast supplies the probabilities it multiplies.
- Sibling mechanisms: Actuarial Risk Model · Bayesian Risk Update · Probabilistic Forecast · Risk Matrix · Risk Scoring Model · Scenario Probability Table · Risk Register · Probabilistic Safety Analysis · Decision Tree
Editorial Notes¶
Form Classification¶
Form family: Analysis, Modeling & Optimization
Rationale: Expected Value Calculation operates as a computation, comparison, model, or analytic representation used to infer, estimate, or choose because it multiplies or otherwise combines probability and consequence on a common scale to rank options by expected gain, loss, or exposure.
Independent corroboration: The frozen evidence defines Expected Value Calculation as 'Multiplies or otherwise combines probability and consequence on a common scale to rank options by expected gain, loss, or exposure', so its operative form is Analysis, Modeling & Optimization.
Review outcome: Independent reviewer agreement; high confidence.
Origin Attribution¶
Primary origin: Mathematics
Origin pattern: Single lineage
Present-day reach: Universal
Rationale: Probability theory formalized mathematical expectation as the sum of possible values weighted by their probabilities; economics later made it central to appraisal and choice.
Related originating lineages:
- Economics & Finance — Decision and utility theory materially developed its use for ranking risky choices. Probability-weighted valuation of consequences is foundational to economic decision theory and finance.
Review resolution: HM Treasury's formal appraisal guidance states the probability-times-value calculation explicitly. Its mathematical structure is primary, with economics a historically important application lineage.
Review outcome: Researched adjudication after independent review; high confidence.
Sources consulted:
Notes¶
[n1] The St. Petersburg paradox — a gamble with an infinite expected payoff that almost no one will pay much to enter — is the classic demonstration that expected-value maximization alone can misdescribe rational choice. Daniel Bernoulli's resolution, diminishing marginal utility, is the ancestor of the utility-curvature and tail-exception fixes that expected value needs but does not supply on its own. ↩