There are far better ways to produce a forecast than using story points that require very little of the team’s time, are far more accurate, and can be done on a continuous basis or even semi real-time.

Numbers, analytics, and models are useful tools when used appropriately. To say they should be avoided is like saying we should avoid using cars because they could be misused.

Misuse of those tools is a people / culture problem rather than a process or technique problem. If your boss is willing to misuse a particular tool, merely taking that tool away won’t resolve the problem.

If forecasts can be done quickly, cheaply, and provide value then why shouldn’t we do them? The real issue is managing expectations. Every forecast by definition is wrong, but that doesn’t mean it’s of no value. Educating the consumers of the forecast as to what it is telling them and what it isn’t is a better use of time than trying to dodge forecasts.