The important question that needs to be distinguished is: How much R&D is involved in your project? It doesn't matter if it is a software project or a physica

The important question that needs to be distinguished is: How much R&D is involved in your project? It doesn't matter if it is a software project or a physical project. If the percentage of unknown terrain is high, the project will become complex. When a very progressive and innovative bridge/house needs to be build with new materials and a unknown construction etc., the project will be complex and you can't design it (completely) upfront. For example, if your software project is just a CRUD application (without complex business logic), a qualified and experienced software developer can estimate it and deliver it on time. (There are a lot of frame works and tools already available for this kind of software. So, nobody doesn't need to reinvent the wheel.) If the wheel is not invented yet for your problem, yes, it can take a while and your luck is in the hand of gods. The other issue is that there are a lot of developers that can't get working any shit. But incompetent people are everywhere.