The cost of an item on a roadmap is not the time it takes to build. It is the attention it removes from everything else on the list, and attention does not divide cleanly. Two projects at half attention are reliably worse than one at full attention and one not started, because the half-attended project accrues decisions nobody has the context to make well.
This is why adding a priority feels free and is not. Nothing is removed, the quarter still looks plausible on a slide, and the degradation shows up distributed across every other item as slightly worse judgment.
The test I have come to trust: for each item, name the person who will lose sleep over it. If two items have the same name, one of them is not happening — you just have not chosen which yet. Choosing is strictly better than letting the quarter choose for you, and it is almost always the same answer you would have reached in week nine, minus nine weeks.
A roadmap with four items and one visible cut is more honest, and more persuasive to the people who have to execute it, than one with nine items and a hopeful tone. The cut is the part that signals you understand the constraint.