We study the revision properties of the Bank of Canada’s staff output gap estimates since the mid-1980s. Our results suggest that the average staff output gap revision has decreased significantly over the past 15 years, in line with recent evidence for the U.S. Alternatively, revisions from purely statistical methods to estimate the gap have not experienced the same drop in magnitude. We then examine the usefulness of real-time gap estimates for forecasting inflation and find no deterioration in forecast performance when inflation projections are conditioned on real time rather than on final estimates of the gap.