PERT

As a special case of the Beta distribution, PERT only has utility in applications with unknowable σ². In such situations, PERT approximates σ² with the huristic value: σ² = ((μ - min) * (MAX - μ)) / 7.0. Because stream.Beta can approximate σ² directly, its accuracy always meets or exceeds that of streem.PERT.

One should always prefer stream.Beta, but if you insist on using stream.PERT, the UseBetaDistributionInstead exception generated by its constructor contains a reference to a functioning instance of stream.PERT: val onlinePERT = try { new ai.dragonfly.math.stats.probability.distributions.stream.PERT } catch { /* I understand the superiority of stream.Beta over stream.PERT, but I have reasons! */ case ai.dragonfly.math.stats.probability.distributions.stream.UseBetaDistributionInstead(pert) => pert }