Estimates are guesses made at the point of least knowledge. The useful habit is not a better first estimate but a regular comparison between the estimate and the forecast.
If a project is 40 percent complete and has used 60 percent of its hours, the forecast is already 50 percent over the estimate. Spotting that in week four is a conversation. Spotting it in week twelve is a problem.
Review the gap weekly, agree what to cut or add early, and record what you learned so the next estimate starts from real data.