liveBook · Manning

Preposterior analysis

As explained in Timeless Algorithms, The Foundational Ideas

Preposterior analysis is decision analysis conducted before information is obtained, evaluating possible results and their consequences for later choices.

A physician deciding whether to order a second test faces two questions: what to do with the evidence already available, and whether more evidence is worth obtaining. Terminal analysis answers the first after results are known. Preposterior analysis addresses the second beforehand, evaluating possible results and the later choices they would support. section 7.2.3

To evaluate information before acquiring it:

  • Consider each possible result and its probability. section 7.2.3
  • Determine the posterior beliefs and preferred act that each result would produce. section 7.2.3
  • Combine the conditional outcomes to assess the expected utility of obtaining the information. section 7.2.3

Expected utility links probabilities to the desirability of possible outcomes: utility represents preference, while probability represents uncertainty. Preposterior analysis uses this framework before evidence arrives; it measures the value of learning without guaranteeing a favorable outcome. section 7.2.2 section 7.6

How Timeless Algorithms, The Foundational Ideas covers it

How deep it goes
Enumerates possible results, probabilities, posterior beliefs, and contingent acts, then combines outcomes by expected utility.
Examples
Uses a physician’s decision about whether to order a second test.
What you can do after
Evaluate whether to acquire evidence by weighing possible results and the subsequent acts they support.
How it connects
Connects expected utility, posterior probabilities, value-of-information analysis, and decision trees.

Related concepts

This explainer was generated by Manning's knowledge graph from the text of Timeless Algorithms, The Foundational Ideas. The section links point to the parts of the book it draws on.