I have recently written about the importance of being both designer and developer when working as an independent creator and so it is possibly ironic that I just had to perform a fairly heavy strip back and re-write because I lost sight of my job as a designer. The designer needs to protect the heart of the game from the developer, which when you're both can get lost in the mechanics of playtesting and problem solving. This, I think, happened in the most recent iteration of Moonflight's decks.
The essence of Moonflight is that a player scores only their hands at the end of the game. My first instinct was to make this interesting to the players by allowing them to extend their hands with their graveyard, tableaux and marketplace and by increasing their hand limit. This instinct was largely sound but it led to the scoring being based differently for each character, which wasn't totally unsound but I built on it by trying to give each character a range of winning conditions. This was unsound and led to a few design problems, for one the decks scored so differently that players struggled to relate their actions and aims and each player's engine ended up being vastly different. I wrongly believed at that point that the value in such a game for players would come from mastering a range of different tactics and options or adjusting and combining them during play. The value and satisfaction comes from building a single engine, possibly or even preferably exactly the same one as your opponent, just a tiny bit slicker and better than their's. More fundamentally from almost the first moments of design I allowed the decks to score from their extended hands, but rather than allowing the deck unbuilder concept to flourish interestingly, it watered it down and washed away its essential elegance until the game end triggering became an issue not of climactic interaction but of secondary importance.