So.. ya since Smash is uniquely made that way, it's pretty tough to know where to draw the line.
This is vaguely true. For example, we turn items off despite items-on not being broken in any sense. There are ways around this, but without giving it proper thought you realize that such arguments will be arbitrary and thus allow arbitrary bans. For example, you could claim "well, items-off is just
better than items-on." But, even with a majority approval, this is still subjective, and thus would allow for a similar argument for anything: "x being banned is better than x not being banned."
Instead, I think a good point to make is that we want to limit randomness.
We do not wish to ban randomness in general. However, even this is arbitrary, and your mileage may vary as to whether items-on is random enough to warrant a ban. One thing we can agree on, however, is that Peach's down-B is necessarily random, and we will likely never ban that move. With that in mind, we always have some sort of "minimum amount of randomness" with which to compare everything. Most players would agree that items-on impacts results in a far more serious manner than Peach's turnip, and so, if Peach's turnip is your "threshold" for randomness, you would disable items.
What is nice about the above logic is that it prevents unwarranted bans on randomness alone. Things like the cars on Mute City and the lava on Brinstar do not necessarily have a large enough impact to warrant a ban.
However, there are problems with this "Peach's turnip threshold:"
For one, the threshold is arbitrary. Luckily, it's a "canonical" choice (but this doesn't make it any less arbitrary!). However, what's to stop someone from choosing a stronger threshold (e.g. "all random is bad") and banning even more?
For two, the threshold is still scrubby (though this is debatable). If things aren't broken, instead simply impacting results in the short-term, why should we ban them?
Another way to address this issue is to come up with a "skeleton" as to what our starting game should be. However, this is just as arbitrary: what's the difference between saying "items-off is toggled" and "Mute City is banned?" There isn't really any.
Note however that this problem does not apply to things like "stock vs. timer," in which you're making a choice between two
different, equally valid, rulesets. One analogy for this specification with regards to stages would be if versions 1.1 and 1.2 had different stages; you would not call anyone scrubby for picking one over the other. They legitimately are different games.
For the most part, I can't reconcile
Playing to Win with our choice to disable items. To a degree, I accept that this game is not meant for out-of-the-box competition, and so changes can (and should) be made to the starting rules. But this leads to the valid conclusion that we can ban basically whatever we want for virtually any reason: "Mute City is janky. What do you mean I'm a scrub? We banned items for no reason too, didn't we?"
Such is Smash. Too bad Sakurai didn't specify any guidelines when he designed this ****. Though his guidelines probably would have sucked anyway.