So on the topic of sleep and competitiveness, there's one argument in particular that I take umbrage with, in regards to the accuracy of regular attacks vs. the chance of getting enough sleep turns.
TL;DR, Sleep is statistically not any more un-competitive than low accuracy moves, of course you don't choose how many turns you're asleep, but neither do you choose the accuracy of your opponent's moves, you choose the accuracy of your own moves, and in the case of sleep, you choose the chances of how long your opponent will be asleep.
When moves with imperfect accuracy are brought up as a counterargument to sleep being un-competitive as it only has a 66% chance of winning you the match (in match-ups where only free turn is needed to hit win-con), the common rebuttal is so: "When you build your Pokemon, you choose the accuracy of your moves. You don't get to choose how many turns you're put to sleep for, this takes choice and skill out of the hands of the player, and is thus un-competitive." The problem I have with this, mainly lies in inconsistent logic regarding teambuilding. No, you don't get to choose the length of time you fall asleep, but you also don't get to choose the accuracy of your opponent's moves, only your own, and in this context, sleep's chances of success are very much something that the teambuilder chooses.
I think we can all agree that when you put Focus Blast on Gardevoir, you are accepting that you will have a 30% chance of losing matches where you need a Focus Blast to win, when you put Stone Edge on Durant, you're accepting that you will have a 36% chance of losing matches you need a Stone Edge to win, et cetera, and that neither these moves, nor these Pokemon, are un-competitive.
Yet for example, let's take the match-up of Hustle Durant vs. Mega Charizard Y.
Durant outspeeds Charizard, and easily OHKOs with a single Stone Edge, winning them the match. However, should that Stone Edge miss, all Charizard has to do is hit a single fire attack in order to OHKO Durant. The player using Durant, when building their team, chose Stone Edge as a coverage move for this match-up, and when making the Durant, decided that 64% accuracy to win a majority of the time against Charizard was acceptable. From the perspective of the Charizard player however, they have no choice but to rely on an RNG roll in order to win. In effect, they have been forced to leave this match-up entirely in the hands of whether or not the move lands or misses, which is entirely up to chance. No matter the skill of the Charizard player, all they can do is hope to get lucky and have the move miss. The skill and player choice was in building your Charizard and choosing it against your opponent, in effect, the majority of the time, you lost when you chose Charizard.
Now let's look at Snorlax vs Charizard Y.
Snorlax easily bulks any move that Charizard can throw at it, and yawns the same turn. A protect the following turn ensures that Charizard falls asleep, and Snorlax survives. On the third turn, Snorlax is free to Belly Drum, regardless of how much damage they've taken, thanks to Normalium. No RNG, outside of the errant crit or burn chance, has happened yet. Snorlax, with maximized attack, can now easily OHKO Charizard with one Double-Edge, provided Charizard does not wake up on the first turn it is able, of which it has a 33.34% chance of doing. The player using Snorlax, when building their team, chose Yawn as a status move for these match-ups, and when making the Snorlax, decided that a 66.66% chance of getting one free turn of sleep to win a majority of the time against its opponents was acceptable. From the perspective of the Charizard player however, they have no choice but to rely on an RNG roll in order to win. In effect, they have been forced to leave this match-up entirely in the hands of whether or not they wake up or not, which is entirely up to chance. No matter the skill of the Charizard player, all they can do is hope to get lucky and wake up. The skill and player choice was in building your Charizard and choosing it against your opponent, in effect, the majority of the time, you lost when you chose Charizard.
Do you see what I'm getting at here? In match-ups where a Yawn user only needs one free turn in order to win, there is ZERO qualitative difference between relying on sleep rolls to be in your favor, and relying on a low accuracy move landing. Both are reliant on RNG, and from the perspective of the player facing it, there is little to nothing that can be done once the match has started aside from clicking a button and hoping the opponent gets unlucky. Yes, when choosing a low accuracy move, you are choosing that with the knowledge that it may fail you and miss, and you accept that risk. However if a sleep move achieves a win-con the majority of the time, again when choosing the sleep move, you are choosing it with the knowledge that it may fail you and not give you enough sleep turns, and you accept that risk. Zero qualitative difference, and in the extreme of examples like Stone Edge Durant, Zap Cannon Magnezone, or Gardevoir needing to land two Focus Blasts in a row (a 49% chance), Yawn actually beats it out in terms of reliability, making it quantitatively less haxy.
This of course, all applies to Yawn. Other sleep moves have differing accuracy, and many users rely on more or fewer RNG rolls to hit their wincon, there's a sliding scale of RNG when it comes to sleep users, and it is entirely possible that the higher end, Gengar for example, might have far too many match-ups where more than one turn of sleep is needed, making it rely on 50/50s more often than not, but in order for that to be the case, there'd need to be a clear line set up for how many 50/50s against the VR does it take for a Pokemon to be considered completely RNG and hax reliant, and it would need to be proven that Gengar or any other Sleep user crosses that line.
However, that sliding scale also goes down well below the RNG reliance of 'regular' moves and strategies, with Snorlax, Jumpluff, Vivillon, Smeargle, and Breloom being just as reliable, if not more so, than 'regular' moves in many cases.
TL;DR, Sleep is statistically not any more un-competitive than low accuracy moves, of course you don't choose how many turns you're asleep, but neither do you choose the accuracy of your opponent's moves, you choose the accuracy of your own moves, and in the case of sleep, you choose the chances of how long your opponent will be asleep.