I mean a lot of us are of the opinion GF doesn't provide a great implementation of mons anyways.
As a postscript to the previous post, this is basically how Smogon started. However, the issue with this general argument is that every other mod to the cart mechanics is done with convenience in mind, generally just preventing us from needing to be nerds that can correctly read how many of the 48 pixels are showing in the health bar, or watching minute-long animations and having to keep your eyes glued to the screen to write down everything that's going on, etc. Most of these mods are done with that type of reason in mind, not as a way to edit the metagame. Off the top of my head, the two mods that were metagame-driven are OHKO clause and sleep clause, and despite the similar names these do not operate in the same way. OHKO clause is analogous to Sleep Moves Clause, not Sleep Clause, since it bans the use of OHKO moves entirely, starting from the team validator. Sleep Clause is analogous to some fucked up alternate reality version of OHKO clause where you're allowed to get one Fissure kill for some reason but then that's it. In the context of Smogon history (all the process rules on this site are designed to learn from past mistakes), we've had notoriously bad luck when it comes to nerfing non-pokemon mechanics as opposed to banning them, as can be seen by the million different Baton Pass nerfs we went through or whatever the fuck BW OU was. This is what makes Sleep Clause an archaic mod, it's the last surviving relic of a time where we thought shit like this was a smart move, and nowadays it's looked down upon for similar reasons to why complex bans are looked down upon.
I understand what you are saying and I think this is a thoughtful analysis of the issue and potential impact, so thank you.
BUT, can you please provide a definition of "Cheese" and of "Honest", because it seems like you and others are ascribing some degree of morality to these terms, with one being undesirable. How can an impartial observer measure these qualities? Are they based on some expected value of outcomes? A quality of play? A certain je ne sais quoi lol?
Further, you use the word "decent" a few times to describe both the offensive pokemon that may use these moves, as well as the odds of breaking through a defensive stop. i.e. paralysis odds are "indecent", whereas hypnosis odd are. What does that mean in both of these contexts? A combination of speed and power? A specific likelihood?
I think it is a very prickly pear to be poking what constitutes "valid" or acceptable play. The game is the game; there's no right way to play it. If we want to bend the rules because we don't like a particular outcome, then OK, but let's be honest with ourselves.
"Cheese" and "honest" are just the colloquial terms that are often thrown around when strong players are talking with each other about the meta. I understand what you mean with the morality point, so I'll expand on this a little. A pokemon is normally called "cheese" if it's significantly easier to use than it is for someone to safely respond to it (assuming that the person playing against the cheese is a strong player). Notably the part about being hard for someone to safely respond to it can apply either in the builder or in game - something like Stored Power Latias under screens is considered cheese because its strong matchups are near auto-wins without much effort from the Latias player or ability to stop the Latias from a playing. The usefulness of calling things cheesy is that it effectively signals which threats in the meta
need to be countered more directly if a player is aiming for a high winrate. Since the path from getting a weak matchup vs a cheesy mon to losing is very hard to stop, it's important to make sure that the most uninteractive mons in the metagame from a gameplay perspective are amongst the list of mons that you have good answers to (since you can't have good answers to every threat in mons).
Calling something cheesy is not an argument for banning something in and of itself - it needs to be supplemented with an analysis of what a good player could do in order to secure wins against the cheese strategy. Essentially my broader argument with sleep is that it's a form of cheese that is not friendly to this process - even if I make the conscious decision to try and run the best checks possible for the sleep threat, I can still potentially lose to sleep rolls, opening up the possibility of losing in uninteractive games, in a way that I also cant really address in the builder. It's also worth keeping in mind that one form of counterplay in the builder isnt really adequate as a response to a threat - you need an assortment of available options, and ideally when you add all of them together, you should have a good enough range of options such that you can address the threat on every team you use, while also being able to use a range of different playstyles.
There isn't really a good objective way to analyze something like this, it will always come back to the limits of our creativity, but on a subjective level that's why it helps to talk to other good players - people go away, they look through all the types of options that might help vs the threat in the builder, they make a judgement of which ones they think could be viable, and then they bank that information for later whenever they need to figure out a list of potential directions they could take a teambuild. And oftentimes in this process, someone will miss something on their own, but then they'll be shown an option by a friend who figured out something while they were building.
Expected outcomes are probably the closest way to describe the type of assessment being made here, but again there's way too many factors to get a proper mathematical read on what works and what doesn't - at the end of the day it's all just opinions. Regardless of this, I think it is important for players at a high level to have some sort of handle on how much they're expecting to win in any given setting, and how well their skill matches up against their opponents, because the metagame will literally change based on the answers to these questions. I generally average a 75 percent winrate on the top of ladder, with an 80 percent winrate in tournaments, and as a result of this I need to take a different approach to the metagame than what I would recommend to the average tournaments player who is aspiring for a 60% winrate. Typically from my perspective, I need to be avoiding cheesy mons and I need to be extra careful when using anything that creates predictable interactions, as the chance of getting a play right approaches 50 percent for both players the more that both players understand the full situation. For the purposes of getting 60% winrates though, it's probably more important to focus on using the best mons available, as they will have the most tools to bail you out if you make small misplays.
"Honest" is just the opposite of "Cheese"
"Decent" in the context i'm using here wasn't intended to be a synonym for "honest", I was talking about the other meaning of decent, i.e. not good but also not bad. In pokemon terms this would mean mons around B rank in the VR (my personal version of the VR has Lilligant-H in the B+/A- range so I just called it "decent")
got Spore, we wouldn't be having this conversation. This indicates to me that sleep moves are the problem - I'm hesitant enough to use moves like Focus Blast due to its accuracy; if sleep is so powerful that people are willing to take a near coin flip 60/40% risk to use it, that's a huge issue. Any Pokemon with high speed that gets a reliable sleep move is an issue - this is not a Pokemon specific problem - literally just about every Pokemon becomes a problem with Hypnosis. 
