[The Fifth Horseman]

Developed a simple game on my own once, betatested another.

First, always start by planning out the core elements of the game mechanics and the engine features you will need to support them.
Second, NEVER fix problems / add features with a "whatever works" approach. This eventually leads to compound kludges which may turn out to be showstoppers later on and are even harder/more time-consuming to fix than the original problem.
Third, no engine survives first meeting with the players/testers. The real development only begins in alpha and continues over most of the beta, no matter how much work you put in before.