How I see it:
He should have release public (or to backers) version way more early on, to keep up motivation.
This also allows players to give feature requests to currently relevant ingame mechanics, which would be way more on point than generalized thoughts.
I have ton of fun with my world generator for 7 Days to Die.
NitroGen for Alpha 7DtD
This is mainly because people use the program, and give me regular feedback and bug reports.
It also lets me know what part of the program people consider most relevant (number of requests on that element).
And having regular updates also lets me keep a realistic program structure in mind, that does not stray off into over-engineered details. When refactoring I always try to keep it in a running state.
(the time within start and end of refactoring is very risky, as stopping for a longer time somewhere inbetween would leave the project a mess)
And if the motivation fades: well there is always at last the last working version remaining. So the project did not end in a drawer so to speak.
...keeping the project under wraps so many years is the main reason it failed in my opinion...