You are not going to get away with disclaiming that somehow the "internal logic of the universe" is responsible for the result when you hand them a bag of poo... You made the table. It is your logic. You own it.
What you're forgetting in your equation is that the players know there’s a 50% chance they’re getting a bag of poo if they attempt something on Valentine's Day. You're describing an event table, and the magic hat in my example works exactly the same way. Both involve a pre-defined set of outcomes and a die roll to determine which one occurs.
But where we’re talking past each other is in how that table, or hat, is introduced. Your table is presented as a direct action by the referee toward the player (i.e., "here’s your Valentine’s gift"). Mine is part of the setting, something the players discover and choose to engage with. So rather than just being handed a result, the players trigger it through their own decisions. That’s a meaningful distinction.
Suppose the magic hat gives a random item based on a 1d4 table with a 50% chance of producing a bag of poo. Now imagine three scenarios:
The player hasn’t investigated what the magic hat does, so they’re unaware of the risk.
The player has been misinformed by another character and chooses to believe them.
The player has learned the hat’s function and chooses to use it anyway.
In all three cases, the consequence flows from the player’s choices. That’s not "whim", that’s interaction with a consistent setting. “Whim” would be me deciding on the spot that it would be funny if the hat had a 50% poo chance, without prior prep or justification. Or, in scenario #2, deciding an NPC lies without any precedent or reason for them to do so.
But if I prepped that hat, placed it in the setting, and the players acquired it through play, then what happens after that is on them. If they’re unhappy because they didn’t investigate or misjudged someone’s advice, that’s a reflection of choices made in play, not arbitrary GM fiat.
Also, keep in mind how I run my Living World sandbox: the world isn’t out to get the players. It doesn’t care about the players. While there are weird magic items, including things like the poo hat, they’re rare, and usually the result of misfired enchantments or eccentric creators. They’re not lurking in every treasure chest. Their existence is part of the world’s internal logic, not a gag at the players’ expense.
You'd have a stronger position if the published game mechanic handed you the answer, and as a GM you are merely the mechanical processor, making no decisions of your own.
My reply to that is simple:
The problem with that assertion is that the GM’s "whim" is present, regardless. I’m pointing out that whether that whim is exerted by choosing to follow the whims of a game designer, at the time of making up a table of outcomes, or in the moment, is not relevant. It is still "whim".
So it isn’t a stronger argument, neither for your premise nor for mine. Plausibility isn’t a binary yes/no switch. My whole process is intricately interwoven with human judgment. If someone doesn’t trust human judgment, then my Living World sandbox won’t work for them. Because yes, the referee’s decisions matter, meaning yes, they’re responsible for the consequences of their game.
If someone runs my sandbox tools and the result feels like a world full of metaphorical bags of poo, that’s on the referee’s. (And to be clear, I’m treating the bag of poo as shorthand for “not fun.” That’s different from a challenging or dangerous world the players knowingly enter.)
But the Living World sandbox doesn’t ask the referee to figure it out magically. I provide concrete processes and methods, during prep and during play, to help ensure that the campaign feels fair, interesting, and engaging. The idea isn’t to eliminate judgment, but to support it, so that players feel challenged in the ways they want to be challenged.
But in the end, with my living world sandbox, it is up to the referee to use it well.