Thanks for the article, Kiryl. I really liked the approach proposed by in Allan Kelly’s “Xanpan” where he says that in his team bugs were usually estimated after dealing with by the person who worked on it. They did it that way due to a higher level of uncertainty than with usual tasks.
I’ve tried that approach, as well as estimating a “Done” bug together as a team, and our velocity got more accurate. However, if a team doesn’t count their velocity, it makes no sense to perform estimations of bugs — absolutely agree.