As Bulbapedia's article on Acid Weather is a bit confusing (does Fog do damage or not, after all?), I'm going to use Elevator Music's thread's OP as my reference.
First, we need to determine how the host feature will be implemented. It is vital to the glitch, as only him can activate it. The host only has to faint an opposing pokémon with Pursuit while it's switching out, but it does not have to summon weather by himself. This means that we could make the host always the one without Pursuit on the team (and randomly choose before the beginning of the battle when both have or don't have Pursuit) and say so on the battle log when the battle begins ("PlayerX is the host"), but such implementation would give out information of one's team to the opponent (depending of the scenario):
A - Player1 (Pursuit) vs Player2 (Pursuit)
"Player1 is the host."
1: "Good, I have Pursuit and am the host, this means it was randomized because the other guy has Pursuit."
2: "OK, was it randomized or was he chosen because he doesn't have Pursuit?"
B - Player1 (Non-Pursuit) vs Player2 (Non-Pursuit)
"Player2 is the host."
1: "Great, I have no Pursuit, which means it was randomized as he doesn't have Pursuit either."
2: "OK, I have no Pursuit and was chosen, but I have no idea whether it was randomized or not."
C - Player1 (Non-Pursuit) vs Player2 (Pursuit)
"Player1 is the host."
1: "Cool, I don't have Pursuit, but what about him?"
2: "Cool, I have Pursuit, but what about him?"
As we can see, there would be only one scenario out of three where there would be no "information leaking" when using a, huh, "hierarchical method" (first X, then Y, randomize if equals). This means this method obviously has to be discarded, and that the best way would have to be unrelated to Pursuit, making the host:
a) Whoever challenges the other (in Unrated battles), or whoever presses the "Find Battle" button (on the Ladder) first. The host's name would be the first on the window title ("Player1 vs. Player2 - ShoddyBattle");
OR
b) Randomize the host every time and have it declared on the battle log, after the rules and before the pokémon are sent out.
In both ways, we'd always have scenario C's players' thoughts in play, not mattering if one, two of none of them have Pursuit. This emulates as perfectly as possible what happens on Wi-Fi/wireless play, as both players know who the host is but have no idea which one shouldn't be the host for no glitch.
Second, we have the battle itself. As expected, many are afraid of what will be of the metagame with messed up weather in every battle. I think it is, with all respect, an irrational fear, though. Rain does not activate it, Sun teams are really uncommon and TR/Gravity only last for five turns, are even more uncommon than Sun and are learned by only a handful, so we are left with Sandstorm and Hail. Acid Sand will obviously be the most common weather due to Tyranitar while Acid Hail will be the most dangerous one (not every team is "ready" to face a Hail team already, Cleffa line, Froslass and Mamoswine will be immune to it and Leftover-less Walrein will recover freaking 25% HP every turn it uses a Substitute).
However, both will be double-edged swords. First, if the Acid player is not the host, then the other one will have full control of Pursuit and Acid triggering. And even when the Acid player is chosen as the host and has a strong enough Pursuiter to faint something, the other side also has a brain. If they see Hail is on the field and that you have a Weavile in, why the hell would they switch out? I mean OK, if the pokémon is slower than it and dies also to Night Slash then they might switch out, but then it's a guessing game and Pokémon is filled with them, but what if the pokémon is going to die even to a 40 BP Pursuit? They would stay in and die with a smile on their faces, only to Scizor to come in and fuck up with Weavile and bury the Acid Hail menace for the rest of the battle. And even if the Acid Hail is triggered, it also activates the Sp. Def boost to Rock-types in Sandstorm and (apparently) every weather-related ability. I'm sure it wouldn't be nice to bring Acid Hail in only to realize the opponent had a, say, Kabutops waiting on the wings. Oh hell, it wouldn't.
Then we have Acid Sand. It will be much, much more common as already said, and just as in Acid Hail, Kabutops (and Omastar and Relicanth) will love it. Steel-types will also have less fear of Salamence as Fire Blast will be neutral and the dragon will take even more residual damage every turn and, most importantly, Water-types will have double STAB on Hydro Pump. Now tell me, what kind of Sandstorm team would want to give Starmie (or Kingdra, or even Ludicolo) a way to fuck up everybody on the team for free? Bringing Tyranitar in to clear the weather wouldn't help much unless it comes out revenge killing, as it would eat a nice Hydro Pump while switching in (as Starmie is obviously firing off Hydro Pumps as there's no tomorrow) and then another if it's not a Scarftar. Not that the non-Tyranitar team has nothing to worry about, though, as any non-immune pokémon will eat up 12.5% damage every turn; if there are no Water or Steel types on the team (even though that's highly unlikely...), Acid Sand will be the last thing the player will want to see. And how will he avoid it? Just like in the aforementioned Weavile scenario: avoiding switching out or, at least, outpredicting the opponent in a mind game of "should I stay or should I go".
Basically, it means both sides, in every battle, will have to think. Acid Weather won't be an insta-win/lose to one side, just like Garchomp, Salamence, Triple Dragon+Triple Steel and BP Chains aren't (even though some are allegedly "more broken" than the others). Just like these other features, Acid Weather will become yet another layer of the game, where team building, skill, prediction, "information control" and luck will be essential. "But what if it shits everything up?". Well then, we'll have to analyze it just as we did with Abomasnow and the current Rain issue on UU. "Is it Pursuit? Tyranitar? Abomasnow? The one who switches out in bad timing? Maybe only Acid Hail? Maybe, just maybe, the whole Acid weather?". We cannot just ignore it because it seems "easier". I mean, it exists since September 2008, just a couple months after our Suspect Test started. Wasn't it made to "fairly analyze certain possibly wrongly-tiered pokémon" (OK this wasn't exactly the wording but you get my point)? If we went all the way to test those 600 BST Ubers (even though none of them ended up unbanned), why should we not implement and, if necessary, test Acid Weather? You know, there's even a slight (really slight) chance this glitch will pass on to 5th Gen. Pursuit had a glitch involving Choice items in ADV, and it passed on to early Japanese D/P copies. They also managed to have the Japanese D/P Transform glitch making a comeback in HGSS, for heaven's sake. I wouldn't be surprised if Sky Attack MissingNo. becomes OU in a year from now.
And now, third and lastly, the Castform/Cherrim issue. I suggest both critters should not be banned, but having them switching in during Acid Sun and above be considered as forfeiting. If you send it out during the battle, it's obviously because you want to force a draw as someone would have to turn their DS off had it happened on Wi-Fi/wireless. If the opponent defeats all the other five pokémon and you have to send it out, well, tough luck. You knew from the moment you put Castform/Cherrim in your team that not only you were crazy, but that you'd be playing 5x6 the moment Tyranitar Pursuited something under Sun and above and did not come back to clear its own mess.
tl;dr version: Make the battle host random, inform both sides, Acid Sand/Hail will be the most common ones, Acid Weather will require brains just like any usual aspect of the metagame, we should test it just like Abomasnow/rain in UU if so is needed, sending Castform/Cherrim out during it means you forfeited, and we could avoid it all by merely banning Platinum.
First, we need to determine how the host feature will be implemented. It is vital to the glitch, as only him can activate it. The host only has to faint an opposing pokémon with Pursuit while it's switching out, but it does not have to summon weather by himself. This means that we could make the host always the one without Pursuit on the team (and randomly choose before the beginning of the battle when both have or don't have Pursuit) and say so on the battle log when the battle begins ("PlayerX is the host"), but such implementation would give out information of one's team to the opponent (depending of the scenario):
A - Player1 (Pursuit) vs Player2 (Pursuit)
"Player1 is the host."
1: "Good, I have Pursuit and am the host, this means it was randomized because the other guy has Pursuit."
2: "OK, was it randomized or was he chosen because he doesn't have Pursuit?"
B - Player1 (Non-Pursuit) vs Player2 (Non-Pursuit)
"Player2 is the host."
1: "Great, I have no Pursuit, which means it was randomized as he doesn't have Pursuit either."
2: "OK, I have no Pursuit and was chosen, but I have no idea whether it was randomized or not."
C - Player1 (Non-Pursuit) vs Player2 (Pursuit)
"Player1 is the host."
1: "Cool, I don't have Pursuit, but what about him?"
2: "Cool, I have Pursuit, but what about him?"
As we can see, there would be only one scenario out of three where there would be no "information leaking" when using a, huh, "hierarchical method" (first X, then Y, randomize if equals). This means this method obviously has to be discarded, and that the best way would have to be unrelated to Pursuit, making the host:
a) Whoever challenges the other (in Unrated battles), or whoever presses the "Find Battle" button (on the Ladder) first. The host's name would be the first on the window title ("Player1 vs. Player2 - ShoddyBattle");
OR
b) Randomize the host every time and have it declared on the battle log, after the rules and before the pokémon are sent out.
In both ways, we'd always have scenario C's players' thoughts in play, not mattering if one, two of none of them have Pursuit. This emulates as perfectly as possible what happens on Wi-Fi/wireless play, as both players know who the host is but have no idea which one shouldn't be the host for no glitch.
Second, we have the battle itself. As expected, many are afraid of what will be of the metagame with messed up weather in every battle. I think it is, with all respect, an irrational fear, though. Rain does not activate it, Sun teams are really uncommon and TR/Gravity only last for five turns, are even more uncommon than Sun and are learned by only a handful, so we are left with Sandstorm and Hail. Acid Sand will obviously be the most common weather due to Tyranitar while Acid Hail will be the most dangerous one (not every team is "ready" to face a Hail team already, Cleffa line, Froslass and Mamoswine will be immune to it and Leftover-less Walrein will recover freaking 25% HP every turn it uses a Substitute).
However, both will be double-edged swords. First, if the Acid player is not the host, then the other one will have full control of Pursuit and Acid triggering. And even when the Acid player is chosen as the host and has a strong enough Pursuiter to faint something, the other side also has a brain. If they see Hail is on the field and that you have a Weavile in, why the hell would they switch out? I mean OK, if the pokémon is slower than it and dies also to Night Slash then they might switch out, but then it's a guessing game and Pokémon is filled with them, but what if the pokémon is going to die even to a 40 BP Pursuit? They would stay in and die with a smile on their faces, only to Scizor to come in and fuck up with Weavile and bury the Acid Hail menace for the rest of the battle. And even if the Acid Hail is triggered, it also activates the Sp. Def boost to Rock-types in Sandstorm and (apparently) every weather-related ability. I'm sure it wouldn't be nice to bring Acid Hail in only to realize the opponent had a, say, Kabutops waiting on the wings. Oh hell, it wouldn't.
Then we have Acid Sand. It will be much, much more common as already said, and just as in Acid Hail, Kabutops (and Omastar and Relicanth) will love it. Steel-types will also have less fear of Salamence as Fire Blast will be neutral and the dragon will take even more residual damage every turn and, most importantly, Water-types will have double STAB on Hydro Pump. Now tell me, what kind of Sandstorm team would want to give Starmie (or Kingdra, or even Ludicolo) a way to fuck up everybody on the team for free? Bringing Tyranitar in to clear the weather wouldn't help much unless it comes out revenge killing, as it would eat a nice Hydro Pump while switching in (as Starmie is obviously firing off Hydro Pumps as there's no tomorrow) and then another if it's not a Scarftar. Not that the non-Tyranitar team has nothing to worry about, though, as any non-immune pokémon will eat up 12.5% damage every turn; if there are no Water or Steel types on the team (even though that's highly unlikely...), Acid Sand will be the last thing the player will want to see. And how will he avoid it? Just like in the aforementioned Weavile scenario: avoiding switching out or, at least, outpredicting the opponent in a mind game of "should I stay or should I go".
Basically, it means both sides, in every battle, will have to think. Acid Weather won't be an insta-win/lose to one side, just like Garchomp, Salamence, Triple Dragon+Triple Steel and BP Chains aren't (even though some are allegedly "more broken" than the others). Just like these other features, Acid Weather will become yet another layer of the game, where team building, skill, prediction, "information control" and luck will be essential. "But what if it shits everything up?". Well then, we'll have to analyze it just as we did with Abomasnow and the current Rain issue on UU. "Is it Pursuit? Tyranitar? Abomasnow? The one who switches out in bad timing? Maybe only Acid Hail? Maybe, just maybe, the whole Acid weather?". We cannot just ignore it because it seems "easier". I mean, it exists since September 2008, just a couple months after our Suspect Test started. Wasn't it made to "fairly analyze certain possibly wrongly-tiered pokémon" (OK this wasn't exactly the wording but you get my point)? If we went all the way to test those 600 BST Ubers (even though none of them ended up unbanned), why should we not implement and, if necessary, test Acid Weather? You know, there's even a slight (really slight) chance this glitch will pass on to 5th Gen. Pursuit had a glitch involving Choice items in ADV, and it passed on to early Japanese D/P copies. They also managed to have the Japanese D/P Transform glitch making a comeback in HGSS, for heaven's sake. I wouldn't be surprised if Sky Attack MissingNo. becomes OU in a year from now.
And now, third and lastly, the Castform/Cherrim issue. I suggest both critters should not be banned, but having them switching in during Acid Sun and above be considered as forfeiting. If you send it out during the battle, it's obviously because you want to force a draw as someone would have to turn their DS off had it happened on Wi-Fi/wireless. If the opponent defeats all the other five pokémon and you have to send it out, well, tough luck. You knew from the moment you put Castform/Cherrim in your team that not only you were crazy, but that you'd be playing 5x6 the moment Tyranitar Pursuited something under Sun and above and did not come back to clear its own mess.
tl;dr version: Make the battle host random, inform both sides, Acid Sand/Hail will be the most common ones, Acid Weather will require brains just like any usual aspect of the metagame, we should test it just like Abomasnow/rain in UU if so is needed, sending Castform/Cherrim out during it means you forfeited, and we could avoid it all by merely banning Platinum.