Let’s be consistent and start pruning the pure RNG elements from the game. Direct Confusion moves, Quick Claw, Attract, Focus Band, and Metronome. All of these are uncompetitive, and deserved to be removed.
First, let's talk status and probability.
Sleep, at base, has a 100% chance of stopping an opponent for 1 turn or more, a 67% chance of stopping an opponent for 2 turns or more, and a 33% chance of stopping an opponent for 3 turns. This is done in 1 move, has no type immunities, and is non-volatile. All chances adjusted to Hypnosis' accuracy for a wide-spread move, that's a 40% chance of nothing, 60% of 1+ turn, 40.2% chance of 2+ turns, and 19.8% chance of 3 turns.
Paralysis, at base, has a static effect of lowering afflicted target's base speed down to 50%, then has a 25% chance of stopping a target each turn. Has a type immunity (Electric) and is non-volatile. It has a 25% chance to stop an opponent 1 turn, 6.25% chance of stopping an opponent 2 turns in a row, and a 1.56% chance to stop an opponent 3 turns in a row. In reverse, it has a 43.75% chance of stopping an opponent once in 2 turns, and a 57.82% of stopping an opponent once in 3 turns. A base 90% filter should be applied for Paralysis Wave's accuracy as the most generic enabler, but since Paralysis counts each turn unlike how Sleep calculates everything the turn it's used, applying this multiplier is not so simple.
In comparison to those two:
Confusion is a volatile status with a 33% chance of stopping an opponent each turn, making them damage themselves. This works similar to paralysis odds, but without the static effect and with the play-around of switching out due to its volatile nature. The odds are low enough to make it reasonable, but most importantly, it allows play-around by just switching off given its volatile nature. At worst, Confusion inducing moves are comparable to Yawn, but with a much smaller punishment for not switching out.
Attraction has a 50% chance of stopping an opponent each turn, with no further effects. This is also not uncompetitive because even though the effect is devastating, it's a volatile effect which allows the above-mentioned work around. More importantly, it has the additional restricition of requiring the target to be of the opposite gender to work, and hard-fails always on non-gendered mons. This not only allows a play-around in-game, but makes it extremely unstable for non-RNG reasons, making it a non-issue.
(flinch wasn't mentioned, but wanted to add a comment on it, for the record: maths on flinch can also be done, but flinch is only a problem in 3 mons that combo it with Serene Grace, out of them 1 is already banned and 1 is dexited, making it irrelevant for this gen. In a generational vacuum, it's held back by the combination of limited users, requiring the user to be faster, and the need to do nothing but spam the flinching move every turn, making it very restrictive on the user's part in comparison to the above statuses. Flinch chances are, thus, mostly a curiosity and have little argumental value)
Metronome is also a non-issue as it's so ridiculously random and broad, there's no way to use it in any competitive way.
On the other hand, there's Focus Band and Quick Claw.
Focus Band does have an argument to go against it. Focus Band survives outside of sight due to the very low chance of it being relevant (10% in a very specific scenario). In short, the only reason why Focus Band is allowed is because no one wants to touch a ban of something that no one uses, and it's too bad of an item to actually use, even if it is uncompentitive by nature.
Quick Claw has been unironically talked about in the past, and only dodges the banhammer for the same reason as above: no one wants to ban something no one uses even if it's uncompentitive. Popularize another team that abuses it, and we may see it banned. It's the only thing in this massive comment that I would also agree on banning, but no benefit to starting that discussion while it's irrelevant.
King's Rock, btw, belongs with these two as a "uncompetitive item that no one wants to bother banning as long as it remains irrelevant". Except that it drew enough attention to ban it.
Does King's rock stack with Population Bomb on Maushold? Because if it does, then we have a 100% flinch chance attack, and with webs support or a tidy up boost, that might be overpowered.
That calculation is too annoying for me to do right now, but simplified by assuming all hits connect, it has a 65.14% chance to flinch. This becomes much more complicated when you add how accuracy affects this, since Population Bomb also has a 10% chance of missing each hit and missing interrupts the attack altogether.
When you add accuracy, it's actually worse than Skill Link + perfect accuracy 5-hit moves (chance is 41%), and both are worse than Serene Grace abusers. Mathematically speaking, anyway.