I remember listening to a commentary track that the director Paul Thomas Anderson did about his film Boogie Nights, in which he explained why a scene was cut from the final film. He was watching a cut of the film, and as one scene was wrapping up, he was giddily anticipating the next scene, which was one of his favorites. But when that first scene had ended, it wasn't followed by the scene he was anticipating - it was followed by another scene that he had seemingly forgot was in the film.
And I'm paraphrasing here - it's been maybe a decade since I've seen the film, and even longer since I've listened to that commentary track - but Mr. Anderson said something along the lines of, that's when he knew that that scene, the one he had forgotten, did not belong in the film. If he, the director of the film, the person who had lived with it for years - as a script, during the shoot, during post-production - forgot that the scene existed, then it wasn't worth keeping in the final cut.
My rulesets tend to be very streamlined. I like to say that I make my rules simpler and simpler until they're simple enough that even I can remember them. Which isn't entirely true - after nearly forty games as a designer, and almost as many that I've had my hand in as a developer, I often need a refresher before playing a game again or answering someone's rules question online - but does get at an essential part of my process which is similar to what Anderson went through with his film; namely, when I find myself "forgetting" a rule, chances are that particular rule doesn't belong there.
Here's an example. In This Guilty Land, one of the things you can do with an Organization Card is form a radical political party. This scores you some points immediately (equal to your Organizational Capacity) and it flips some aligned Compromise markers to your side, and replaces others with opposing-aligned Compromise. In the very first version of the game, this galvanic shift took place only in a number of Regions equal to the value of the card. So, a "2" Org card would see this happening in two Regions, a "3" in three, and so-on.
And when we started testing the game, we played the rule as written. It wasn't too long though before I forgot that bit of nuance, and started playing as if the political party shift took place in all Regions. This made the move much more powerful (especially for Justice, who can use it to greatly increase their end-game scoring). Going back over the rules, I was shocked to discover this oversight. But I also realized that the way we had been playing it felt right, felt organic and natural, in a way that the rule-as-written didn't. So I changed it.
The original rule was sound, and I understood on an intellectual level why I had written it that way. It had a purpose, and was part of a careful calculus. Mathematically, it made perfect sense. But I'm not a mathematical designer. My process is more instinctual than analytical, more about what feels right in my gut than what adds up on paper. If I forget a rule, chances are there's a reason for it; if, in my enthusiasm for one of my designs, I play it a certain way, that's probably the way the game should be.
My gut isn't always right. There are times when I forget a rule and that rule should very much be there. But when that's the case, the thing is dead on the table, a glorious and awful mess, and in the postmortem I go back over the rules and realize at what point I botched it. I put the rule "back" in, and voila, the thing works. Even in those cases, my forgetting is still useful, because then I need to figure out why I forgot the rule. Is it just absent-mindedness on my part? (As Mary can attest, I am more than a little absent-minded.) Or is there something deeper at work? Maybe that piece of the puzzle works, but maybe the game, on a big picture level, doesn't feel right, hence the overlooking.
That's harder to do sometimes, and harder to admit to myself. It's easy to trust your gut when the result is something concrete: cut this rule, make this thing easer, and so on. It's harder when it's more nebulous: it works, mathematically, everything is just-so, and yet something isn't quite there yet.