OK, cool. As you might guess if you've seen my posts in the initial thread, I'm definitely on the ban side. Hopefully someone on the pro-SR side will respond to jpw, but for now I just want to answer SDS a bit more formally and add a couple points that haven't really been made yet.
Ease of use: I think LN was making a slightly different point here that was definitely valid, and I want to elaborate on this one a bit. LN's point here was that, compared to other features of OU that define the tier to a similar degree as SR does, SR requires much less changing of one's team. I honestly see (and I think few would deny--let me know if you do) SR as shaping the meta as much as if not more than a major weather. It hugely changes how viable many mons are, how you go about countering various threats (which, as you noted, is showcased by the fact that getting SR up and then RKing dnite is the popular strategy to check something like DDnite since SR naturally breaks multiscale and lowers it to KO range). I'm not saying that any of this stuff is bad. Just that SR has an incredibly huge impact on the meta...and because of its distribution and lack of drawbacks, it requires basically nothing of a team. This is not true for anything with remotely SR's metagame-shaping ability.
The second issue in terms of ease of use I want to touch on is linked to the relative difficulties of RSing and SRing, which I'll get to below.
Permanence of use: The crucial difference I see here between DM, for example, and SR, is that, sure that -40% will be there the whole game, but the opp isn't being attacked every turn. SR is a damaging move that activates repeatedly through the match, and does MORE damage each time. Conversely, sure that damage may remain for the rest of the match, but that's all it's going to do. The "effect" that is permanent is that it will damage, not the damage itself. I mean, are you saying that if DM does less to the opp because they brought a blissey out that it's not having its effect? No--it is the act of dealing damage, not the damage itself, that is the move "activating". And yes, I'd agree that this is true for status effects as well (although the fact that it affects the entire team is huge), and for spikes--but this is not enough to ban something, it's just yet another point against SR in OU. Even if something like freeze was judged to have the same "strength" as SR does on a team (note that you're weighing a debilitation of one mon vs passive damage on an entire team, although I might even still compare freeze's effect on a match to SR's), it still would almost certainly not be broken, because there is nothing that has more than a slight chance to cause freeze. Tl;dr, SR's "permanence of effect" is much more lasting than any other damage-dealing move, and arguably more than most status effects, while it has far more effect on a team than any status effect or similarly permanent move.
Strength of support: I went though a rough calculation in the previous thread, so I'll try to abridge it here. Based on experience, I'd estimate that the mean game length is something approximating 30 turns (I find mine usually range from 20-50ish on average, with some exception obviously and leaning towards the shorter end). Again, I'd estimate something around half of my turns are spent switching. Sometimes a couple mons will duke it out, lowering that average, but then sometimes there'll be double-switching that prompts another (double) switch, and then you're lowering the average, but I've found I switch something approximating every other turn. So 30 turns, 1/2 of them spent switching, is 15 switches per game on average. Let's assume an entire team is neutral to SR. If it gets set up turn 1, and does 12.5% each switch, that team will accrue 187.5% damage over the course of the match. Obviously, this is a very rough estimate. It is lower if a mon has lefties that activate because of SR damage, or heals "past full". Then again, if a team has a single SR-weak mon (which the majority do, I feel confident in guessing), the total damage is raised considerably, particularly given that one SR-weak + one SR-resistant =/= two SR-neutrals. Meanwhile, higher-level games tend to have a markedly higher ratio of switching, often including multiple back-to-back turns of double-switching with both players trying to maneuver into an advantageous position. So all in all I'd estimate an overall effect of something in the vicinity of 200% if set up turn 1. Feel free to disagree with any specifics of this, but I think this serves as an accurate estimate?
Honestly, I get kinda scared looking at this number. One move, taking one turn to set up, that in an average game can deal close to 1/3 of an entire team's health. I can't think of anything with close to this effect (sleep and freeze, for instance, are the only things that stand out in my mind like this. They are separated from SR by the fact that a sleeped/frozen mon is certainly not dead, the conditions are not necessarily permanent, there are viable ways to counter them (a natural cure/magic bounce mon for sleep, for instance, using scald or switching in on a fire attack for freeze, and above all distribution, particularly in the case of freeze. And even still, 100 certainly =/= ~200%). I really see this as where SR is just crazy. Burn and paralysis are strong effects, certainly, but I hope no one will claim that they average an effect of ~200% lost per game. This is the first strike of bannability in my mind.
Discouraging switching: what I'll say is this: free switching is something that should be initially assume in my mind, and should only be lost through being outplayed by the opp's use of such. Here's what i mean. Pursuit is, in some ways, uncompetitive. If the opp is in a checkmate situation, there is literally no play they can make to improve it. But what this also means is that you had to arrange situations so that YOU could be in this situation. For example, consider LOmie vs bandtar. Mie 2HKOs ttar if it predicts the switch-in, so this is not simply a path you can take towards uncompetitive victory. Instead, you have to set up a path where you've established that you always bring your heatran in on their ferrothorn and try to lava plume it. Instead, pivot to ttar as they go to mie to absorb the hit. Obviously, this is a simplified version, but the point is, the only way to accrue that "uncompetitive advantage" is by smart competitive play. Same with spikes--sure, you can lead with your spikes user, but if you send out ferrothorn, they had better not have sent out anything with a fire attack, a powerful fighting attack, or a taunt user, or have a magic bounce user on their team, because in that case ferro is getting up a layer at best. And this doesnt give you nearly the competitive advantage of having SR up--you do 12% max and don't do any damage to anything with flying or levitate (11/top 30 OU mons, fwiw). This is so so much less of a disincentive for switching than SR.
So what makes SR essentially uncompetitive while these others are not? I'll phrase it like this: it does NOT take skill to execute, and once set up, reduces emphasis on skill.
Pursuit/other trapping: very much decreases emphasis on skill (if the opp has their tenta in vs duggy, there's not anything they can do). However, you either have to sac a mon (decreasing the magnitude of the effect considerably) or you have to predict right to get into the advantageous position. So it requires skill to execute, and sure it decreases what the opp can do, but that's fine--their task was to avoid that situation.
Spikes: one layer takes little skill, three takes much more--and their effect on the opp's ability to abuse switching is proportional. Three layers of spikes punishes the opp for switching arguably at the same level as SR (especially since many mons are outright immune to spikes, they can be abused as well--something that is truly impossible with SR, barring the extremely rare magic guard guys), but take three turns to set up, which is huge especially since one turn is almost always guaranteed given the ability to simply lead with a SRer. On the other hand, one layer is similarly easy as SR to set up (although not in truth, due to an issue below), but limits switching much much less.
The other issue here is distribution, which is totally crucial. Part of the reason why it is so much harder to set up three layers of spikes than to set up SR (even if you had to do it 3 times for it to take effect) is because SR users are just so much better than spikes users, due to SR's ridiculous distribution. There are 4 OU mons that can use spikes (skarm, forry, ferro, cloy). First of all, there is a remarkable similarity between them: they are all physical walls with very low SpD and a marked weakness to fire on the whole. While these guys do get ample opportunities to switch in on physical attackers, they are forced out just as often by special attackers. So while they may get a layer up as a physical attacker switches, they're certainly not getting two--and the next time they have an opportunity to come in, they run the risk of being out-predicted, in which case your team has to take a hit from a special attacker or lose momentum. if you can pull it off three times, great--but this is a very difficult task. Compare this with the fact that like half of OU can viably use SR (in fact, 10 out of top 20), and it simply is so much easier to get your SRer in for a turn. Also note that ferrothorn, the one exception to the rule of spikers being frail specially and thus being forced out easily, basically requires rain support for this to be the case since without it any fire user OHKOs it. Notice how even this is somewhat dictating your playstyle (similarly, forry is primarily used on sun, since it needs the cushion on water attacks, fire being a lost cause). The fact that these guys are all defensive is even more restraining to playstyle, since they lose momentum and thus dont fit at all on offensive teams. It also makes them unfit as leads, delaying the time from which spikes can do its work.
Keeping otherwise broken things in check: This is an interesting one since it goes both ways. I think the main thing to keep in mind is just that, even though many threats like volc get harder to contain, they also gain new, much more reliable counters. I don't see this as a huge issue and in fact, my guess is that dnite would be almost certainly broken if SR was banned (unless sand/hail became massively more popular) because multiscale basically always being present is hard to deal with. Volc, I'm not so sure, mostly just because it doesn't become fundamentally different without SR in play--it just has more chances to set up/more health to do so with. Anyway, I don't see this as an important issue, and as you pointed out SDS, it shouldn't affect a decision--and this goes both ways.
Countered by rapid spin: I firmly believe that a well-constructed team (perhaps one, as you phrased it, that is "heavily reliant on SR") that wants to keep SR down on the opp's side will be able to guarantee this 99% of the time vs even a well-constructed team that very much does not want SR down on their side. SR coming sooner than RS by definition is a factor, yes, but in my mind these are the two major factors that make it so so much easier to set up SR than get rid of it: countering tactics and distribution.
Countering SR--basically, you can taunt/magic coat or send out a magic bouncer. That's it. Taunt/magic coat are both very rare, meaning you don't encounter them much. Furthermore, when you DO see them, it's very easy to predict them coming because of their rarity and simply attack the opp. Meanwhile, magic bounce being distributed to two FE mons, both of which are mediocre...meaning that in addition to basically entirely wasting a teamslot to counter SR (already a step towards overcentralization), they cannot even switch into offensive SR setters. terrakion, ttar, mamo, chomp, etc. all decimate both xatu and espy and can set up SR, while even stuff like forry can volt switch out of them. Doesn't help that both are liable to be pursuit-trapped (although BP on espy is nice).
Countering RS--first of all, they have to find a free turn to do so. This is much harder because a SR lead can just rely on their sash, while a spinner by definition won't have one. So option 1 is just kill it or don't let it have a free turn--this is what ideally happens with most offensive teams, and it is very very difficult to get a spin in. The other method is using a ghost-type. Admittedly, there are basically two good options: jellicent and gengar, with sableye sorta counting as well. And the thing is, it doesn't even really matter if the RSer can beat them--although I'll go into this a bit more below--because just sending them in blocks the spin. And saccing a spinblocker to kill the spinner is basically always worth it.
Distribution: This is what really tips the scale in my mind. Basically, there are two viable OU spinners: tenta and mie. Donphan and forry fill specific niches, but basically only function on sun/stall respectively, and neither can do anything to touch the aforementioned spinblockers. And tenta and mie are both highly prone to being trapped: tenta by duggy and mie by ttar/weavile/scizor/pursuiter. Because of this, a core of something like thunder gengar + scarftar is close to 100% guaranteed to beat all spinners: tenta only 2HKOs gar in rain, in which case it can abuse thunder to close to OHKO tenta (252 SpA Life Orb Gengar Thunder vs. 252 HP / 0 SpD Tentacruel: 294-346 (80.76 - 95.05%) -- guaranteed 2HKO). If it's not in rain, it can just sball twice while tenta barely recovers health. Then jsut send in ttar to finish up. Mie is even simpler: sac gar, bring ttar in. Pursuit does 60% min to offensive mie (the only kind that has a chance to beat jelli), so you win guaranteed. It does get a spin off if they predict right (you can always crunch, of course, especially given that they took 22.5% switching in and killing gar at the very minimum), but just bring your SRer back in and set up rocks, saccing it as you do so. This is possible because the distribution of SR is simply too wide, allowing top OU offensive threats to set it up, while RS's narrow distribution means only very few mons are able to effectively use it, and these are so specific that they are easily dealt with.
Stealth rock's effect on typing: Yes, it happens to be true that many of the types that SR hits hardest are already mediocre. But that's no reason to think that SR has not played a huge role in making many mons of those types non-viable in OU. Additionally, this seems really important given the goal of tiering, which is to maintain a balanced meta that is as diverse as possible, including in regard to the viability of different types. Sure, ice has lots of disadvantages compared to something like fighting, but it has a niche in helping to beat dragons--this is what gives the meta flavor: a complex, intricate rock-paper-scissors. But when you switch the rules and say that now rock beats paper as well, or even that paper only has a 50% chance to beat rock, you're essentially invalidating it as an option (and by doing so, scissors as well, which loses its niche...). Obviously a super-simplified parallel, but you get my point. Adding SR does nothing but exacerbate the fragility of what are already some of the weakest and least viable types, which is contrary to our goal regarding OU.
Stealth rock hurting playstyles with a lot of switching: Honestly, I see this as a subset of the "switching as the competitive element of pokemon" argument. In many ways, stall is the most intelligent playstyle, and more crucially involves more intelligence in playing correctly. As I already explained, SR really discourages this, and thus discourages stall as a result. In and of itself, not a bad thing--I just think this is demonstrative of the non-competitive nature of SR and is a big reason I feel it should be banned/suspected.
@Epsilon: Just want to point out that the number of mons lost because of a SR ban is far outnumbered by the number gained--I'm thinking sandslash, donphan maybe, things like sableye and hitmontop would be less used as well. Starmie would def stay OU imo, it's been OU for 5 gens including 3 in which SR did not exist, and it's still an amazing rain attacker. A SR ban would not hurt stall because stall is the kind of playstyle that can make use of spikes, so it wouldn't alter their strategy that much while they lose a ton of residual damage on their members. As I mentioned above, this would not only help stall, it would help make stall's success rate much more connected to the skill level of the player. Finally, we can't just wait until next gen because: BW is gonna be relevant way after XY comes out; GF has never done anything to equalize the meta, we have no reason to think they'll start now; SR is not gonna go through a mechanics change; and finally, that's not how tiering works. We have months before XY is released and if we believe that SR is broken, we should test and potentially ban it.