Bypass Risk Score per Project(3.1.5)

Each gauge compresses two measures of a single project's status transitions into one number between 0 and 100, over the trailing 90 days. The first is the bypass count: transitions that went from a not-started status directly into a done one, skipping everything in between. The second is the reopen count: transitions that moved an issue back out of a done status. Both are counted against the same population, the project's total status transitions in the window, so the score is the share of that project's recent movement that took one of those two shapes.

The window is fixed at 90 days and does not follow the page's date filter. That is deliberate, so that two people looking at the same project on the same day read the same number. The window each gauge was computed over is printed on the panel.

What you can conclude

• A score below 40 means most of the project's recent transitions followed the intermediate steps of its workflow.
• A score of 40 to 70 means a noticeable share of recent transitions skipped straight to done or came back out of it, which is the range where it is worth asking whether the workflow matches how the team actually works.
• A score of 70 and above means most recent transitions took one of those two shapes. That is the point at which the difference between the designed workflow and the practiced one is large enough that a reviewer would ask about it.
• Comparing scores across the portfolio shows where the two patterns concentrate. Read the gauges against each other, not in isolation.

How this chart works

Grid of semicircular gauges, one per project, each showing the score from 0 to 100. Bypasses carry slightly more weight than reopens in the blend. Color zones: green below 40, amber 40 to 70, red 70 and above.

A project appears here only once it has at least one bypass in the window, so a project with none is absent rather than shown at zero. A project with fewer than 10 transitions in the window is shown without a score: its gauge is drawn empty and reads n/a, labeled Not scored, rather than showing a low number. At that volume a single transition would move the score by more than 10 points, which is a third of the narrowest band, so the number would be an artifact of thin data rather than a measurement.