You can't prepare for something that is purely RNG and operates independantly regardless of player action. It is noninteractive and removes player autonomy and skill. This is the kind of uncompetitiveness.
But if that was the problem, King's Rock would have been banned way back when tiering policy was first developed. It was always a 10% chance of random BS, which could happen on any pokemon and is impossible to anticipate. It showing up on Cloyster is both more powerful and less random. Suddenly it's a common item pick for a specific pokemon, that you can scout like any other item, and has a chance of flinch that is high enough that you should reasonably take it into your calculations*. That's not RNG, that's a tradeoff the same as picking Fire Blast over Flamethrower is.
And that's when the ban came through. It wasn't banned for being RNG, it was banned for being good and annoying.
And that's a valid reason to ban things, to be fair. I'm just saying that a lot of debaters on every side here are holding up the policy like they're sacred texts, arguing over what the meaning of "is" is, but in practice a lot of people are going to vote based on "the meta will be less shitty if X", and that shouldn't be forgotten.
*41% chance of increasing damage by 100%, repeating. Which makes it better than LO but less reliable
and they shouldn't have to do that shit either. this is actually kind of the worst example you could use because jirachi's continued existence is a massive plot hole in any self-consistent tiering policy
Only if you think flinching in general is a problem**. Jirachi doesn't have an RNG chance of winning. At 60%, Jirachi's flinch chance is more reliable than not. The person who stays in and tries to hit through the flinching is the one who hopes the RNG works in their favor, the person using Jirachi is just abusing a stunlock, a perfectly reasonable thing to do in competitive games.
**And we haven't banned Fake Out, so it's not