In that L5R example and similar cases, the question that seems obvious to me is why is the dishonesty necessary?
A simple rule could be agreed upon: A PC cannot die without the player's consent, nor an NPC without the GM's. The reciprocity is not necessary, but I like it. The point is simply for everyone to agree on the game being played. It's not necessary to lie about abiding by the rules!
If we are not willing to accept the outcomes of algorithms, then why pile them up into a complex structure? Why not turn instead to rules directed at producing the range of results we do want? D&D was originally designed to produce results that can easily overturn any preconceived story. The companions who are "supposed" to end up slaying the Dark Lord can instead fall in a fight with his lowly minions in the first chapter!
The more specific rules get "fudged", the less they really are rules. It's a slippery slope from playing a game to sitting down for the GM's story hour. When I go to watch a movie, I don't except footage of the director rolling dice before each scene. That would be just a distraction from the story. Likewise, I don't want to spend hours making irrelevant decisions and dice rolls in an illusion of a game. If it's billed as a game, then I want to play for real.
This is quite another matter from the old-style referee's adjudication. When I DM OD&D, I toss dice in plain sight of the players in every case except when it really would give information they should not have -- but then it remains information for me nonetheless. I don't save any figure arbitrarily, because then its demise would also be by my choice (as I chose not to "fudge" that one time). When I make a ruling, there is likewise no need to keep the reasoning secret other than that the players have not yet discovered the reason for themselves. Barring that, we often consider situations together and come to a consensus on how to treat them. Reasonable people can disagree -- but we can also agree to an extent that might surprise some of the rules-lawyer persuasion.
It is quite possible to have a story game in which the final outcome is not in doubt, in the minds of any of the participants. The meaningful questions to be answered then have to do with the particular route between the story's beginning and end. What happens along the way? Surely there is a field of more and less desirable collateral results, and the game's challenge lies in finding a better rather than worse path through that.