3 Sigma
The 3 Sigma rule stretches the net wider still, telling you that about 99.73% of your data falls within three standard deviations of the mean. That means almost every single outcome β all but about a quarter of one percent β lands inside this range. Anything outside it is genuinely rare and worth investigating.
As before, the "mean" is your average and the "standard deviation" (Ο) shows how much your data spreads out. To find the 3Ο boundaries, subtract three standard deviations from the mean for the low end and add three for the high end. The 99.73% is a fixed characteristic of the normal (bell) curve, not something you calculate each time.
This idea is the foundation of "control charts" in quality management β tools that plot process results over time with upper and lower limits set at Β±3Ο. If a data point falls outside those limits, it signals that something unusual (a "special cause") is affecting the process and needs a closer look. For a project manager, 3Ο is the standard boundary that separates normal, expected variation from a true warning sign.
Think of it like guardrails on a mountain road. They're set far enough out that normal driving never touches them. But if a car actually hits a guardrail, you know immediately that something went seriously wrong β just as a data point beyond 3Ο tells you a process has broken from its normal behavior.
With the same task averaging 10 days and a standard deviation of 2 days: three standard deviations below is 10 β (3 Γ 2) = 4 days, and three above is 10 + (3 Γ 2) = 16 days. The 3 Sigma rule tells you 99.73% of tasks will finish between 4 and 16 days. So if a task ever took, say, 20 days, that would fall well outside the 3Ο range β a strong signal that something abnormal happened and you should find out why.
Every PMP formula explained free β plus worked examples and practice in PMP Math, and full timed mocks in the simulator.