The past X weeks of point-estimates is what you use to forecast which things fit in the next Y weeks, and you can't have both stability and forecast accuracy. Any attempt to permanently "peg" a point to a certain number of man-hours is going to interfere with that accuracy.
In the end, the work takes the time it takes, and nobody knows how long that will be ahead of time. Fiddling around with estimates helps with ranking but not prediction.
Over time the estimates should be trending closer to outcome, as the process improves in breaking down and specifying the details that impact prediction & work, and the statistical gap from previous estimates gets baked into future estimates. The process, capabilities, ability to identify diverging factors, and correction of initial estimates should all be maturing concurrently.
The entire point of using fuzzy numbers is to enable fuzzy yet usable predictions. Similar work in a similar situation, armed with specific statistics and outcome, should be highly predictable at the team and individual level over time.
