Ask HN: What is your experience with complexity vs. time based estimations?

Time based estimations are always wrong.

Complexity based estimations are so abstract as to be useless.

I honestly haven't found a good way of estimating, in my many years of working in software teams.

The best seems to be: set a reasonable "deadline" that seems achievable, agreed with the team, and work towards that. At least that gives everyone a goal, a focus, and a sense of "cutting your coat according to your cloth".

Could you elaborate on complexity estimations? Not sure what it means.