Valuation Tasks

A basic, yet highly useful, measure of a task is the probability of success. This value allows for the construction of a binary tree that calculates the task's expected value (probability of success * value of gain - (1 - probability of success) * value of loss). These values can be compared to thresholds for reward vs. risk to make decisions for proceeding with or abandoning the task (where success means not losing money).

Valuation assumes that the question and associated answers critical to understanding the task have been reduced to a manageable size (less than 9). As part of the fusion process prediction markets can be used to distill group opinion into a relevance ranking of these questions; the highest scoring answers composing the infosphere. Relevance can be defined as the number of possible solutions that would use the answer.

Using Bayesian inference, the probability of task success can be found by applying the team's judgement to the infosphere. Because the infosphere allows for understanding the entire task at one time, the team is able to first arrive at a subjective probability based on its collective experience. This number is then combined with an objective score, the result of testing, simulation, or a utility function, to deliver the probability of success.

Voting procedures that partially replicate prediction markets can augment the information from the statistic. Anonymous voting will improve the accuracy of the result by eliminating "noise". Visible voting of ranges can illuminate outliers that may be highly valuable to "sensemaking".

The probability of success can extend analysis when input into models such as real options, capital allocation, or information flow. Such metrics can include the price of risk, risk-adjusted returns, the value of time, economic value added, and potential non-turbulent changes in workflow velocity.