Probabilities for a Chance Event with Multiple Dependencies (Command Program)
The number of probabilities displayed is the product of the number of states of all generic events (events whose states are not specified) contained in a command, including the states of generic influencing decisions or controlled events. If you specify the state for every event that influences the probability distribution to be displayed, the Session Log displays exactly one probability.
To display the conditional probabilities of a chance event that is conditioned on more than one event, the same procedures apply as for a single conditioning event, although multiple conditioning events should be separated by commas.
For example, if you examine the program defined in Displaying Probability Distributions, you can see that the event Success directly depends on the events OperatingCosts and MarketShare. Because OperatingCosts and MarketShare depend on the decision Modernize and on the event Competition, Success indirectly depends on those two events as well.
Success.Yes | MarketShare.Gain
The two probabilities displayed by the Session Log correspond to the two alternatives of the decision Modernize. Marginal computations were performed for the chance events OperatingCosts and Competition.
Success ~~~Modernize.Yes : 0.62 ~~~Modernize.No : 0.66
With the Command | Formulas turned on, it becomes clear how DPL calculates these probabilities.
~~~Modernize.Yes (OperatingCosts.Sink | Modernize.Yes) * (Success.Yes | OperatingCosts.Sink,MarketShare.Gain) + (OperatingCosts.Rise | Modernize.Yes) * (Success.Yes | OperatingCosts.Rise,MarketShare.Gain) : 0.62 ~~~Modernize.No (OperatingCosts.Sink | Modernize.No) * (Success.Yes | OperatingCosts.Sink,MarketShare.Gain) + (OperatingCosts.Rise | Modernize.No) * (Success.Yes | OperatingCosts.Rise,MarketShare.Gain) : 0.66
Versions: DPL Professional, DPL Enterprise, DPL Portfolio