A changed estimate for one task is not automatically permission to rewrite the reference plan for everything else. State which task and baseline are authorised to change before opening the update dialog.

Use a separate Project desktop for Windows practice file. Keep the original reference evidence. Baselines store comparison fields including dates, duration, work and cost. In Gantt Chart, select the intended tasks, choose Project, Set Baseline, Set Baseline, select the target baseline and use Selected tasks rather than Entire Project for a scoped update.

Make the unchanged tasks part of the check

Create three fictional subtasks A, B and C under a summary. Suppose their baseline work values are 2, 3 and 5 hours. In the trial, only B's accepted estimate changes to 4 hours. Record all three original values, not just B's, before updating the chosen baseline for B.

Under Roll up baselines, To all summary tasks updates the corresponding summaries from selected tasks. From subtasks into selected summary task(s) addresses selected summaries, including saved baselines for added tasks and subtask deletions. Microsoft says to select both options when both subtasks and their summary tasks are selected. Choose according to the intended scope rather than dismissing these controls as cosmetic.

For the simple work-total example, A and C should retain 2 and 5, B should become 4, and a correctly updated matching summary should total 11. A summary still showing 10 does not invalidate the individual B value; it identifies a roll-up result that needs explanation. Do not add durations or dates as if they were additive work quantities.

Check which comparison field the report uses

Before confirming an overwrite prompt, verify the baseline slot and selection again. Saving a different baseline preserves another comparison set, but Microsoft's standard variance calculations use the original Baseline fields rather than Baseline1–10. Inspect the fields behind the report instead of assuming a newly saved slot automatically changes its variance.

Compare the accepted reference values with the recorded approval scope. If unrelated A or C changed, investigate before issuing the file. If the parent summary differs, review the selected roll-up options and the subtasks actually included.

Keep the reason, authorisation and baseline identity with the handover. The purpose is a traceable reference revision: a reader should be able to identify what changed, what stayed fixed and which comparison the published report is using.

Sources: Microsoft: baseline updates.