IMO, the line that needs to be drawn should be around what RNG causes significant lack of competitiveness even when information is symmetrical between two players. For Scald, Discharge, and that general family of moves, both players are aware of the Pokemon most likely to carry those moves and can account for them accordingly, either in the teambuilder or battle. In this case, those kind of move RNGs are very well on the "acceptable RNG" side of the line. Moves like Spore sit firmly in this camp because information about Spore's users and methods to mitigate its effectiveness is symmetrical to both players. Both players know that either Amoonguss or Breloom (or any mushroom Pokemon) will carry Spore almost 100%, and if either Pokemon is relevant in their respective metagames, team builders can account for some method of mitigation.I think almost everyone who plays these tiers accepts that some forms of RNG just have to go in order for a reasonably competitive metagame to be maintained. That's why I really don't like these "RNG is just part of Pokemon/Pokemon shouldn't be Chess"-type arguments. Yeah, we know that. The discussion isn't "ban all RNG mechanics" vs "allow all RNG mechanics"; a line has to be drawn and the argument should be where we draw it.
King's Rock, Quick Claw, and other "Luck Items" straddle close to that line. The 10% effect chance provides the user with informational advantage over their opponent, but that advantage inconsistently generates a large enough momentum swing in a battle, especially in higher-rated matchups (except the Evasion items). There are cases where the item + Pokemon creates a problematic situation (Sand Veil Garchomp, Cloyster, Quick Claw Ursaluna), but I think the Pokemon should be assessed before the item.
The point about Pokemon + item creating hard-to-counterplay situations extends to the other side of the line: RNG elements that are hard to play around even when information is (mostly) symmetrical to both players. The first example I think of is ParaSwagger, especially with Prankster users. Before tiering action was taken in Gen V, there was a high likelihood that seeing Thundurus + Sableye on the same team meant that you were gonna run into ParaSwagger shenanigans. Even if you accounted for that in teambuilding to some degree, the opponent mostly had the advantage by manipulating in-game probabilities to generate free turns. The similar argument can be extended to the Pokemon + item mentioned earlier. Gen IV Garchomps under Sand were always Sand Veil, but you knowing that information couldn't help you better stop Garchomp from gaining a massive momentum swing.
With all that said, Hypnosis Darkrai (and by extention IV) straddles that line. The set of lead Hypnosis Darkrai has initial effectiveness because most people did not initially prepare for it (informational assymetry). Now that this issue is now highly publicized across the community, most builders will be keeping Hypnosis Darkrai in mind when building teams. If Darkrai is still generating a net positive for its user even with the set being public knowledge, then the TLs will need to assess Darkrai only. FTR, this is not assessing sleep as a mechanic or a commentary on Sleep Clause because I think that the effectiveness of Sleep as an RNG-generated advantage largely depends on the user (e.g. Spore Amoonguss vs. Hypnosis Darkrai vs. something random like Sing Clefable).
Last edited: