Mission value and task expected value

Optimitron shows two different quantities. They answer different questions.

Mission outcome: value if achieved

A mission describes an outcome, such as ending war or ending disease. The mission value asks what that outcome would be worth under one stated scenario.

Mission value is not a forecast. We do not estimate the mission's probability.

The task tree labels comparable mission scenarios Value if achieved. It does not use the quantity as task expected value or task priority.

The tree leaves a mission value empty when the available scenario measures a different quantity. The methodology can still show that scenario as context.

Mission comparison scenario

End War, End Disease, and End Poverty use this calculation:

Value if achieved = current annual outcome value × 73.4 years

The extinction reference scenario values today's population once. It does not multiply current lives by the comparison period. The tree does not compare it with the three annual-flow scenarios.

These are chosen scenario assumptions. They are not facts about every task. Changing an assumption changes the mission value.

The 73.4-year period does not become a default task duration. Each task keeps the duration supported by its own evidence.

Task scenario: probability-weighted expected value

A task estimate describes one stated success scenario. Current task-tree estimates use this calculation:

Task scenario expected value = stated outcome value if the task succeeds × P(success | task funded or done)

The probability is the stated chance of success if the task gets funded or completed.

Current task-tree estimates do not subtract the chance of success without the task. They do not estimate the value caused by the task alone.

Several alternative mechanisms can target the same outcome. Their estimates overlap. Do not add them.

When a without-task baseline exists

A sourced without-task baseline can support a marginal decision estimate. This estimate measures the value caused by the task alone:

Marginal decision estimate = stated outcome value × [P(success | task funded or done) − P(success | no task funding or work)]

Use this formula only when sources support both probabilities. Call the result marginal. Current task-tree estimates do not use this formula.

The task record stores the probability-weighted scenario value. The queue uses it with task effort and cost.

Use the task's actual effect duration. A one-time benefit stays one-time. A five-year estimate stays five-year unless new analysis recalculates it.

Recalculate the value before you change its duration or assumptions.

Do not add task estimates into a mission

Adding overlapping task estimates can count the same outcome more than once. A subtask must not change the mission's value if achieved.

Show mission outcomes and task scenario estimates separately.

Worked root scenario: optimizing Earth

The task tree shows Value if achieved $7.4 quadrillion for Optimize Earth. It does not add the mission values below it. Disease productivity losses and medical costs overlap the root opportunity-cost estimate.

This is not expected value. Expected value would also require a defensible probability that humanity achieves the objective. The current model does not invent that probability.

Worked mission scenario: ending war

The task tree shows Value if achieved $834T. It does not claim that ending war has a known probability.

This scenario uses the whole annual cost of war. The 1% Treaty addresses a smaller outcome and needs its own task scenario estimate.

Worked mission scenario: ending disease

This scenario covers economic drag. It does not price pain or suffering.

Worked mission scenario: ending poverty

This population-equivalent scenario values the income gap at the global median. It is not a forecast of income distribution or mission success.

Reference scenario: preventing extinction

The task tree does not display this number as a mission value. It is not comparable with the three annual-flow scenarios.

The scenario excludes every future generation. It does not multiply current lives by the 73.4-year comparison period. It does not estimate extinction probability. The value of a statistical life represents collective willingness to pay for small mortality-risk reductions. It is not a price assigned to a certain death.

Funding context: minimizing animal suffering

No accepted dollar conversion covers all species and harmful experiences. The task tree therefore leaves this mission value empty.

This amount covers farmed-animal advocacy. It excludes companion-animal markets, wild-animal welfare, and most direct animal care. It does not include pet food or veterinary spending. It is an input cost, not the value of minimizing animal suffering. The tree does not display it as a mission value or use it for ranking.

Rethink Priorities estimates a median welfare range of 0.332 for chickens, compared with 1.0 for humans. That estimate concerns possible experience intensity. It does not mean one chicken life equals 0.332 human lives. A dollar valuation would also require the affected populations, duration, intensity, intervention effect, and moral weights. No accepted model currently supplies all of those inputs.

Write a task estimate

1. State the outcome and the task scenario.
2. Estimate P(success | task funded or done).
3. Use the effect duration supported by the evidence.
4. Calculate the task scenario expected value.
5. Record the formula, assumptions, and sources with the estimate.

If a sourced without-task baseline exists, record both probabilities. Label the result as a marginal decision estimate.

Leave the value empty when the evidence cannot support an estimate. A missing estimate is clearer than a guess that enters the ranking.