This conversation makes me wonder: has there been any halfway decent quantitative research on estimation methods? I've lived through what seems like the de fac

This conversation makes me wonder: has there been any halfway decent quantitative research on estimation methods? I've lived through what seems like the de facto agile standard: you take some piece of work, and estimate it out according to a point scale, assuming the "average" dev does the work. Curious to know now if it has any effect on project timeliness. Or are these estimates generally "vanity metrics" in practice? Anecdotally, they are generally pretty useless for planning unless they are at a very fine grained level, with low complexity, on an extremely well-understood piece of software.