This is a long post, and a lot of what I want to say has already been touched on, but I want to try to outline my responses to others' comments as well as why I feel SR warrants a ban or at least a test. Here are some points that have repeatedly come up:
Usage vs brokenness issue: Usage =/= brokenness. Yes, we all agree with this. The critical point here is that SR is so incredibly common that it is hard to evaluate its impact. As I've said before, SR is literally considered "standard battle conditions". This doesn't mean that it is broken because it is so used. What it means is that we have to consider its ubiquity in thinking about what consequences it has. Infernape has become much less viable this gen in large part because of rain's dominance; similarly, many mons are made much less viable because of SR's existence--and many many more mons, and much more so, because SR is so much more common (and arguably more crippling) than rain. This is not an argument for brokenness, don't attempt to take it a such. I'm just noting that SR's ubiquity means it has a nearly inestimable effect on the meta.
Mons that would become broken because of SR: it may very well be true that dnite and volc would be too hard to check without rocks (personally, I would be close to sure about dnite since multiscale is truly a scary ability; volc, really only an issue for very offensive teams that try to prevent it from setting up in the first place, every other check can usually beat it at 100% or 50%). Hopefully we can all agree that this is no reason not to ban rocks (if it was found broken, I will get to this below). If terrakion and keldeo were found broken and removed from OU (not inconceivable, I hope, although of course unlikely), it is possible that volc, for example, would just become too hard for offensive teams to check and would need to be banned. OK then, but this is no reason not to consider banning terra and keldy. We deal with broken things one at a time. I personally don't really see this or anything as an argument that says that SR benefits the meta, and I'll talk about why I feel it harms it below.
Actual effect of SR on a prepared team: this is semi-debatable, but really against an intelligent opponent with a reliable SR user (which are abundant due to the move's incredible distribution), it is not hard to repeatedly set up SR when the need arises (including, critically, at the very beginning of the match). In contrast, rapid spinning is much much harder--not only are there far less viable users, but there is a method to stop it. Furthermore, the spinners that can get past good blockers (jelli and gar essentially, so tenta in rain and starmie in rain) are exceedingly vulnerable to being trapped, either by pursuit or by goth/duggy. And don't you dare say that letting your spinner get trapped is your fault--if your opp brings their spinblocker in on your RS and attacks, you have two choices: switch or attack. If you switch, you don't get the spin off and come in at lower health next time, while the spinner takes no damage (unless there are also hazard up on their side, in which case you're basically arguing that SR checks SR). If you attack, you either beat them (KO the spinblocker) or are beat (have your spinner KOed). Obviously the second version is unfavorable, but even if the first occurs, the opp can sjut bring in their trapper and KO your spinner. In fact, predicting the spin-blocker switch doesnt help in the slightest, since being able to kill it sooner just shortens this.
Magic Bounce: OK, if GF had given this to something with a decent defensive typing and stats, things might be so different. As it is, espeon and xatu cannot switch in on the majority of offensive SR setters, and easily pursuit-trapped, and are close to a lost spot on the team. If I send out my SR terrakion lead, you have to first guess what set I am and then predict absolutely perfectly vs me to prevent me getting rocks up. Not reliable. Ttar is even worse.
Taunt: poorish distribution, and most of the time a waste of a valuable moveslot. The fact that you by definition have to be faster than your opp's lead for it to work is a downside, and the fact that it basically gets no use outside of the lead slot (and on stuff like jelli, but jelli's never gonna be preventing a first-turn SR). Also the fact that, because taunt is so much rarer than SR, it is much easier for me to guess that you carry taunt than for you to guess that I carry rocks--so when you send out terrakion vs my hippo, for example, you can bet im going straight for the eq (or better example: your hydreigon vs my heatran. I just go straight for HP ice). I really don't see either of these as anything approaching reliability.
OK, so in my mind rocks are quickly up and almost always there to stay. So how much does it really affect the game? I'm gonna make an estimation here, let's say the average game is ~30 turns long. Not a super important number, I don't have an actual statistic and it varies depending on playstyle, but 30 seems reasonable to me. Let's say, of those 30 turns, even half of them are spent switching (and this number is certainly higher in higher-level games as well). Let's assume 6 SR-neutral mons. So 15 switches in a game, 12.5% health lost each time. That's 187.5% health lost in a match. Obviously this number is not perfectly accurate, due to stuff like lefties that wouldn't activate if not for SR, mons dying, and various resistances/immunities, but on the whole these are rather balanced out by the fact that having a single SR-weak mon on your team adds way more damage than having a resistant one. So on average, something approximating ~200% is lost by the single move of SR--and of course, this number can be much higher and will in fact be if you engage in a lot of switching, which is often a hallmark of high-level play (indicative, not necessary--don't quote me lol). Personally, this seems like a ridiculous effect for a move that takes one turn and that about every pokemon in the game somehow knows.
Over-centralizing: I don't know if this is actually a word, but in my mind it means roughly this: when a metagame shifts (as a direct result of whatever is purportedly "over-centralizing") to be much more accommodating to certain mons and playstyles (that do well vs whatever is purportedly over-centralizing), and to be much less accommodating to other mons and playstyles (that fair poorly vs the over-centralizer) to too significant a degree. Obviously this is a subjective definition, and feel free to disagree with me on the wording and such, but I think we can all agree somewhat on the definition? The critical word here is the "too"--what is too much of a shift? However, I think it's fair to say that SR has an absolutely enormous effect on the meta--more than any other move, for sure, and I would argue more than drizzle even--which can be seen at the very least from the usage stats I and others have pointed out.
Banning a move: Here's the way I'd like to think about it. We're not banning the move SR--we're banning the condition of having rocks up. Sure, you can teach your mons the move, but it can't be used in battle (in practice, obviously the move would be banned, but that's just a matter of trustworthiness and implementation). This is the same idea as with sleep clause--we are not banning any moves, because the moves are not broken. Being able to freely spam sleep moves is what is broken. So we ban that. Mostly a technicality, and not super important imo--obviously, always a time for firsts, and we've done complex bans, ability bans, even item bans, so I don't see why a move ban is a stretch.
Brokenness: OK, so this is the big issue that a lot of people keep mentioning. I believe SR has a distinctly negative effect on the meta and that it is too powerful. A couple of big reasons:
Sheer damage and ease of use: As I estimated above, in a standard match, if SR is set up first turn, it will cause roughly 200% damage over the course of the game. Yes, it helps that it is distributed over all 6 mons, but even so. I simply don't see a way in which that could not be a little scary. 1 turn of set-up with no cost, more pokemon that learn it than like any other move except tackle (ok, not literally), very very hard to stop reliably, very hard to remove at all, and even harder to remove reliably.
Over-centralization: I don't see why this shouldn't be grounds for banning something, it's one of the foremost arguments for rain's banning in the debate thread ("if you don't abuse weather, you're out of OU" translates quite easily to "if you're weak to SR and aren't a weather-starter, a former uber, or have access to one of the best stat-boost moves in the game, you're out of OU"). SR limits diversity, it makes certain mons and strategies much less viable (I'm not arguing that hail would be particularly good if not for SR, just that it would certainly be more viable). There's also a funny effect that, of course, the more over-centralized on something a metagame is, the less of an effect it has. So in other words, if your team has all SR-resistant mons, you won't take nearly as much damage by switching around a lot (still close to 100% though on average), but you are proving how over-centralizing SR is.
These are my main two reasons for feeling SR is broken, though there are other, smaller ones. One kinda interesting one is that it discourages switching, perhaps the most intellectually challenging and interesting part of battling. Anyway, I personally do see plenty reason for SR to be at least deserving of a test, and little reason for that not to happen.
Usage vs brokenness issue: Usage =/= brokenness. Yes, we all agree with this. The critical point here is that SR is so incredibly common that it is hard to evaluate its impact. As I've said before, SR is literally considered "standard battle conditions". This doesn't mean that it is broken because it is so used. What it means is that we have to consider its ubiquity in thinking about what consequences it has. Infernape has become much less viable this gen in large part because of rain's dominance; similarly, many mons are made much less viable because of SR's existence--and many many more mons, and much more so, because SR is so much more common (and arguably more crippling) than rain. This is not an argument for brokenness, don't attempt to take it a such. I'm just noting that SR's ubiquity means it has a nearly inestimable effect on the meta.
Mons that would become broken because of SR: it may very well be true that dnite and volc would be too hard to check without rocks (personally, I would be close to sure about dnite since multiscale is truly a scary ability; volc, really only an issue for very offensive teams that try to prevent it from setting up in the first place, every other check can usually beat it at 100% or 50%). Hopefully we can all agree that this is no reason not to ban rocks (if it was found broken, I will get to this below). If terrakion and keldeo were found broken and removed from OU (not inconceivable, I hope, although of course unlikely), it is possible that volc, for example, would just become too hard for offensive teams to check and would need to be banned. OK then, but this is no reason not to consider banning terra and keldy. We deal with broken things one at a time. I personally don't really see this or anything as an argument that says that SR benefits the meta, and I'll talk about why I feel it harms it below.
Actual effect of SR on a prepared team: this is semi-debatable, but really against an intelligent opponent with a reliable SR user (which are abundant due to the move's incredible distribution), it is not hard to repeatedly set up SR when the need arises (including, critically, at the very beginning of the match). In contrast, rapid spinning is much much harder--not only are there far less viable users, but there is a method to stop it. Furthermore, the spinners that can get past good blockers (jelli and gar essentially, so tenta in rain and starmie in rain) are exceedingly vulnerable to being trapped, either by pursuit or by goth/duggy. And don't you dare say that letting your spinner get trapped is your fault--if your opp brings their spinblocker in on your RS and attacks, you have two choices: switch or attack. If you switch, you don't get the spin off and come in at lower health next time, while the spinner takes no damage (unless there are also hazard up on their side, in which case you're basically arguing that SR checks SR). If you attack, you either beat them (KO the spinblocker) or are beat (have your spinner KOed). Obviously the second version is unfavorable, but even if the first occurs, the opp can sjut bring in their trapper and KO your spinner. In fact, predicting the spin-blocker switch doesnt help in the slightest, since being able to kill it sooner just shortens this.
Magic Bounce: OK, if GF had given this to something with a decent defensive typing and stats, things might be so different. As it is, espeon and xatu cannot switch in on the majority of offensive SR setters, and easily pursuit-trapped, and are close to a lost spot on the team. If I send out my SR terrakion lead, you have to first guess what set I am and then predict absolutely perfectly vs me to prevent me getting rocks up. Not reliable. Ttar is even worse.
Taunt: poorish distribution, and most of the time a waste of a valuable moveslot. The fact that you by definition have to be faster than your opp's lead for it to work is a downside, and the fact that it basically gets no use outside of the lead slot (and on stuff like jelli, but jelli's never gonna be preventing a first-turn SR). Also the fact that, because taunt is so much rarer than SR, it is much easier for me to guess that you carry taunt than for you to guess that I carry rocks--so when you send out terrakion vs my hippo, for example, you can bet im going straight for the eq (or better example: your hydreigon vs my heatran. I just go straight for HP ice). I really don't see either of these as anything approaching reliability.
OK, so in my mind rocks are quickly up and almost always there to stay. So how much does it really affect the game? I'm gonna make an estimation here, let's say the average game is ~30 turns long. Not a super important number, I don't have an actual statistic and it varies depending on playstyle, but 30 seems reasonable to me. Let's say, of those 30 turns, even half of them are spent switching (and this number is certainly higher in higher-level games as well). Let's assume 6 SR-neutral mons. So 15 switches in a game, 12.5% health lost each time. That's 187.5% health lost in a match. Obviously this number is not perfectly accurate, due to stuff like lefties that wouldn't activate if not for SR, mons dying, and various resistances/immunities, but on the whole these are rather balanced out by the fact that having a single SR-weak mon on your team adds way more damage than having a resistant one. So on average, something approximating ~200% is lost by the single move of SR--and of course, this number can be much higher and will in fact be if you engage in a lot of switching, which is often a hallmark of high-level play (indicative, not necessary--don't quote me lol). Personally, this seems like a ridiculous effect for a move that takes one turn and that about every pokemon in the game somehow knows.
Over-centralizing: I don't know if this is actually a word, but in my mind it means roughly this: when a metagame shifts (as a direct result of whatever is purportedly "over-centralizing") to be much more accommodating to certain mons and playstyles (that do well vs whatever is purportedly over-centralizing), and to be much less accommodating to other mons and playstyles (that fair poorly vs the over-centralizer) to too significant a degree. Obviously this is a subjective definition, and feel free to disagree with me on the wording and such, but I think we can all agree somewhat on the definition? The critical word here is the "too"--what is too much of a shift? However, I think it's fair to say that SR has an absolutely enormous effect on the meta--more than any other move, for sure, and I would argue more than drizzle even--which can be seen at the very least from the usage stats I and others have pointed out.
Banning a move: Here's the way I'd like to think about it. We're not banning the move SR--we're banning the condition of having rocks up. Sure, you can teach your mons the move, but it can't be used in battle (in practice, obviously the move would be banned, but that's just a matter of trustworthiness and implementation). This is the same idea as with sleep clause--we are not banning any moves, because the moves are not broken. Being able to freely spam sleep moves is what is broken. So we ban that. Mostly a technicality, and not super important imo--obviously, always a time for firsts, and we've done complex bans, ability bans, even item bans, so I don't see why a move ban is a stretch.
Brokenness: OK, so this is the big issue that a lot of people keep mentioning. I believe SR has a distinctly negative effect on the meta and that it is too powerful. A couple of big reasons:
Sheer damage and ease of use: As I estimated above, in a standard match, if SR is set up first turn, it will cause roughly 200% damage over the course of the game. Yes, it helps that it is distributed over all 6 mons, but even so. I simply don't see a way in which that could not be a little scary. 1 turn of set-up with no cost, more pokemon that learn it than like any other move except tackle (ok, not literally), very very hard to stop reliably, very hard to remove at all, and even harder to remove reliably.
Over-centralization: I don't see why this shouldn't be grounds for banning something, it's one of the foremost arguments for rain's banning in the debate thread ("if you don't abuse weather, you're out of OU" translates quite easily to "if you're weak to SR and aren't a weather-starter, a former uber, or have access to one of the best stat-boost moves in the game, you're out of OU"). SR limits diversity, it makes certain mons and strategies much less viable (I'm not arguing that hail would be particularly good if not for SR, just that it would certainly be more viable). There's also a funny effect that, of course, the more over-centralized on something a metagame is, the less of an effect it has. So in other words, if your team has all SR-resistant mons, you won't take nearly as much damage by switching around a lot (still close to 100% though on average), but you are proving how over-centralizing SR is.
These are my main two reasons for feeling SR is broken, though there are other, smaller ones. One kinda interesting one is that it discourages switching, perhaps the most intellectually challenging and interesting part of battling. Anyway, I personally do see plenty reason for SR to be at least deserving of a test, and little reason for that not to happen.