Interesting perspective Kiryl Baranoshnik. I completely agree with your point about the effect of “absolute estimation”. Further on that, what is dangerous is that when we provide a single value for the estimation, by definition we are saying that we are highly confident in that one value. However, I’d say its more appropriate to have an estimate range as it more accurately reflects the uncertainty and knowledge associated with the estimate.
I’m curious to know how would you define an estimate that is “good enough at a much lower cost”? Also what does this cost comparison look like and include?
Whilst I think relative sizing is important, my understanding of the end goal of estimations is to usually answer “how long will it take to deliver this project to the customer?” and/or “how much will it cost?”. So in the case of the relative sizing, how does it help us answer those questions? Does the effort to translate the “sizes” get included in that time taken to do the estimation in your opinion?