A single confident date is the least honest thing an engineer can say. Software estimates are uncertain, and pretending otherwise just moves the disappointment to the end.

We estimate in ranges tied to what we know. Early on, when the unknowns are large, the range is wide. As discovery answers questions, the range narrows, and we say out loud what would move it in either direction.

This is not hedging. A range with its assumptions written down is far more useful than a date with none, because you can see what to decide to make it smaller.

Clients who have been burned by confident dates tend to relax when they see the reasoning. A number you can interrogate beats a number you have to take on faith.