Announcement Slowbro and the OU/UU cutoff.

[Hipmonlee]

The RBY OU council have been discussing where to draw the line between OU and UU.

Previously, this was working on the basis of anything ranked B or above would be OU and C or below would be UU. Unfortunately, the VR system does not automatically assign names to ranks. Up until now this had fallen to whoever was organising the VRs, but, thanks to fairly consistent ratings this has been a fairly straightforward call to make each year.

This year, Slowbro has thrown a spanner in the works.

I think we can pretty much all agree that we have a rank which is very clearly B, and a rank that is very clearly C, and one Pokemon hanging out in between them.

So it has fallen to the council to determine what should be done, and we have not come to an agreement on the best way forward, and so we are seeking input from the community.

We have discussed the following options:

​

1. The Council assigns ranks.
This system is easy to organise and is flexible enough to cope with any possible arrangement of ranks that a VR might show up. It also covers scenarios where a pokemon is hanging out between A and B ranks for instance. However it seems somewhat at odds with the spirit of community VRs for the council to just make a ruling.

2. We define a number (n) of Pokemon that should be the minimum size of OU, and the rank that includes the nth Pokemon in the VR is OU and below that is UU.
This system is the easiest to implement. It mirrors the usage cutoff used in modern tiers. However, the number is ultimately arbitrary, and as with any arbitrary system it is possible that this will lead to scenarios in future that we might consider a bit weird.

3.The people chosen to submit VRs vote on where the cutoff should be.
This system is also very flexible, and is the most democratic, but is by far the most work. Quite frankly it sounds like a huge pain in the ass.

4. Some kind of statistical method where VR submitters provide their own cutoffs, and an aggregated cutoff is calculated.
Unfortunately the council does not have the capability to design such a system. If someone has an idea for such a method, please suggest it, and we will take it into consideration. Obviously it is too late to apply this method this year, so in this case the Council will make a decision this year, but for future years we will follow this method.

Another idea we considered was that whatever cutoff will result in the fewest Pokemon changing tier, but you could imagine a scenario where for multiple years in a row, the previous year’s top C rank and two lowest B ranks end up forming a rank together, and OU ends up adding a new Pokemon to B rank every year. It just could result in weird scenarios like this.

If anyone has any other ideas please post them in this thread.

It should also be noted that the Council is unanimous in agreeing that Slowbro should remain OU. So, for instance, when defining the number of OU pokemon, we suggest that the number should be 13. However, in the spirit of community VRs, we will consider all input in good faith, and act accordingly.

 

[The Faz]

The root of the problem here has nothing to do with where we draw the line between OU and UU, the problem here is how the system broke down when a tier between B/C appeared.

Simply calling that zone "B/C Edge", make a poll between the council if B/C Edge always falls into UU or stays in OU and there.
This way if another mon ends in a similar limbo, like a if a mon ends in a tier between A- and B+ for example we'd have the terminology for it.

The big problem here was Slowbro falling into the unnamed no-precedent tier; just give it a name and a precedent and that's it. No council deciding on a case-by-casis, no extra work needed to change the method, no extra stuff on the submissions, nothing.

The current methodology for the tiers don't name them, leading to it sometimes creating weird tiers between + and - tiers or even lower. This solves the issue, simply say they're "(higher tier)/(lower tier) edge tier" and always round it into the higher/lower tier, whatever tier the council decided to be the precedent.

 

[Hipmonlee]

2 is abhorrent because it doesnt care about meta changes or actual pokemon strength. if we defined OU as top 15 when vic and lapras were OU then today lapras and articuno would be OU, and imo clearly both are quite below OU level.

If we felt that 14 was the appropriate number of pokemon to be OU, then of course Lapras would be OU. Having decided Lapras is OU, the VRs tell us it would be unreasonable to differentiate Lapras from Victreebel.

The thing is, the idea of being "OU level" is mostly defined by where we have placed the cutoff. ADV, for instance, lists (in my opinion) much weaker mons than RBY as OU. Articuno, Lapras and Moltres all have seen top level tournament success recently, its not outrageous that they would be called OU. It's just a matter of how we want to arrange our tiers.

The root of the problem here has nothing to do with where we draw the line between OU and UU, the problem here is how the system broke down when a tier between B/C appeared.

Simply calling that zone "B/C Edge", make a poll between the council if B/C Edge always falls into UU or stays in OU and there.
This way if another mon ends in a similar limbo, like a if a mon ends in a tier between A- and B+ for example we'd have the terminology for it.

The big problem here was Slowbro falling into the unnamed no-precedent tier; just give it a name and a precedent and that's it. No council deciding on a case-by-casis, no extra work needed to change the method, no extra stuff on the submissions, nothing.

The current methodology for the tiers don't name them, leading to it sometimes creating weird tiers between + and - tiers or even lower. This solves the issue, simply say they're "(higher tier)/(lower tier) edge tier" and always round it into the higher/lower tier, whatever tier the council decided to be the precedent.

The issue with this is that for a precedent to be of value the VRs need to consistently produce a clear B and C rank, which there are absolutely no guarantees of ever happening. Next year we could see everything from Gengar down to Victreebel all fall into one super rank. Or we could see them stratified into individual rankings. Either way, any precedent we set here would be pointless.

 

[The Faz]

The issue with this is that for a precedent to be of value the VRs need to consistently produce a clear B and C rank, which there are absolutely no guarantees of ever happening. Next year we could see everything from Gengar down to Victreebel all fall into one super rank. Or we could see them stratified into individual rankings. Either way, any precedent we set here would be pointless.

From what we gather from how the tiering system works wouldn't this just work perfectly?

From what we remember from discussions in the discord, a situation like this would just endup with "all the mons from Gengar all the way to Victreebel are on the same tier", no? "All these mons are the B- rank and are roughly the same strength". If any mon was consistently ranked higher than another mon it'd appear on the graph as a separate tier from the other mon.

The problem here is your second example, if every mon ended being squarely on different tiers. Tauros S+, Chansey S, Snorlax S-, Starmie A+, etc as an example. There wouldn't be nearly enough tiers in the slightest, something my system would fix; assuming the council votes Edge mons round down for the sake of example:

If Eggxecutor fell into the pit between S- and A+, it'd be S/A Edge but legally an A-tier mon.

If Lapras fell between B- and C+, it'd be a B/C Edge and thus C (and thus UU)

If Alakazam fell between S/A Edge and A+ it'd be S/AEdge / A Edge, thus A+. (This is a very extreme example tho, and one not likely to happen. This would only happen if everyone decided Lapras is objectively better than all C+ mons but objectively worse than Slowbro.)

It's not a big change, being just a little addendum that covers when mons fall between two already designated tiers like this

 

[Sabelette]

I don't have a clear solution but more of a general thought - what is the point of the cutoff here? Is it to mimic usage-based tiering as best as possible without actually having usage-based tiering? Is it to assign all pokemon to the highest tier that they do at least decently well in? Is it to assign all pokemon to the lowest tier they don't break in half? Is it to make all the tiers feel "fun" or "healthy?" Based on the discord discussions around this it feels more like people are trying to post-hoc justify Slowbro being an OU mon and the cutoff being stuck at exactly wherever Slowbro it is by whatever argument is convenient - it'd break UU, it does fine in OU, it gets used in tours often enough, the "Slowbro Standard," etc - rather than having a standard reasoning for why cutoffs should be where they are, which would be able to apply to other tiers and other mons. That might be fine now but I'm sure it'll bite UU/NU/PU in the ass later when these drops or lack thereof have cascading effects on those tiers and their own drops, and the reality is that the letters assigned to tiers are arbitrary as is saying "B- is the cutoff" because that means different things to different people. We could just as easily have had S not exist, or +/- tiers not exist, or said C is the cutoff, or whatever else; B- being the standard feels to me more like tradition and it feels as though people have very different ideas of what B- or C tier means, but have been able to avoid issues with it in RBY OU just because of how ancient and relatively stable the tier is.

At some point I suggested, for example, setting the cutoff at B instead of B-, or after 10 mons instead of 13, which drops Gengar and Jolteon (I know it sounds wild but I had a certain logic to it), and got a lot of discussion (and pushback) on it, but none of the arguments against it addressed why smaller tiers or more stringent cutoffs would be worse, instead coming back to those mons being too strong for current UU or to them being B tier. Looking the opposite way, there's really no reason that we can't justify extending OU further to basically any mon getting tour wins, which adds the birds and Lapras to OU on the basis that they can do something good in the tier. Basically anything can be justified, so the question isn't about which point we draw the arbitrary line at, but rather what is this arbitrary line meant to do? Is it meant to make both tiers as healthy as possible, is it meant to completely ignore the lower tier and entirely focus on whether the mon fits the higher tier, or what? What happens with the myriad mons that fit multiple tiers well (or equally badly) such as Electrode, or mons whose best tier is OU or Ubers, where they're still niche, but better than they are in NU or PU?

There's no right way to choose a rationale for tiering cutoffs - ultimately, it's subjective and we can't just mathematically deduce a perfect tiering system that objectively makes Pokemon goodTM, funTM, or balancedTM. Even if it's subjective, though, people need to be on the same page about what rationale is being used, or this is gonna repeat every single time there's a mon on the borderline and that mon will end up just getting banlisted and treated as though it never dropped. We don't need it to be objective - literally every Smogon policy is based on subjective ideas of how to make Pokemon competitive, and the huge diversity of pet mods and other metagames shows there's a lot of ways to do so. There will definitely be disagreement about what the rationale of a tier cutoff is, but deciding on a rationale will help avoid repeating this with every future VR, especially given that UU and NU are likely going to see a lot more volatility than OU.

Anyway, hope that thought-dump ends up being helpful to someone, but it's probably word salad. Gl with the tiering stuff, I do not envy y'all right now trying to figure this out

 

[Plague von Karma]

I think we can pretty much all agree that we have a rank which is very clearly B, and a rank that is very clearly C, and one Pokemon hanging out in between them.

Yeah, it's called B-.

I strongly believe that this point is being made way more complicated than it needs to be.

Yes, Slowbro is far away from either regular rank, but if we look at the historical context of OU, I think the method is actually working the way it's intended, just being misunderstood. Even on the most ancient, crusty Viability Rankings where we were playing the game wrong, Slowbro has consistently been at or near to the bottom of OU.

Here's where the problem comes in.

This rationale of Slowbro being at the bottom transitioned into how the method spat out its result. I believe that because many of these cut-offs use Slowbro as the metric, this is why it got ranked so much lower than everything else and created its own tier. We are aware of this ideology, it has been a tradition for well over a decade. Ergo, this data is being misinterpreted due to not factoring in the community context.

The weird rank you're seeing, factoring in "The Slowbro Standard", is B-. The results gave B- to you. Take it.

The B- rank system works, it is working as intended. This feels like a miscommunication between the robot and the worker.

------------------------------------------------------------------------------------

So let's get into the non-solutions.

1. The Council assigns ranks.
This system is easy to organise and is flexible enough to cope with any possible arrangement of ranks that a VR might show up. It also covers scenarios where a pokemon is hanging out between A and B ranks for instance. However it seems somewhat at odds with the spirit of community VRs for the council to just make a ruling.

Terrible. Utterly terrible. Antithetical to everything this community is based around. The community has historically had extreme distrust in the RBY OU Council due to repeated organisational screw-ups and I doubt any decision made would have integrity behind it. While yes, it has changed and improved in many areas - the delegation of lower-tier councils is amazing - I still don't see this working out with the community at all. There's a timeline where people respect the intelligence and integrity of the RBY OU Council, but as-is, I don't think this would work.

This isn't a call-out, I just don't think it works right now. You're under no obligation to be liked, you're here to make the right decisions, it's just...hard.

2. We define a number (n) of Pokemon that should be the minimum size of OU, and the rank that includes the nth Pokemon in the VR is OU and below that is UU.
This system is the easiest to implement. It mirrors the usage cutoff used in modern tiers. However, the number is ultimately arbitrary, and as with any arbitrary system it is possible that this will lead to scenarios in future that we might consider a bit weird.

Given Lapras and Golem - which were OU for over 15 years, have dropped - I don't think this works either. I echo melanie's opinion that this would inadvertently trap some Pokemon in OU when they should not be. You could say you pick an arbitrary number based on projected amounts, but in practice, changing the line every year from the council's own opinions would just feel like shady picking and choosing. Moltres has been on the rise with the best player of this tier - arguably of all time - saying it should be OU, and if this trend continues, that could happen. What then? What if Lapras suddenly becomes a great Pokemon again? What if a mechanical discovery suddenly drops Chansey stonks by a trillion? What if the number was just wrong?

------------------------------------------------------------------------------------

So now for the bits I actually think work.

3.The people chosen to submit VRs vote on where the cutoff should be.
This system is also very flexible, and is the most democratic, but is by far the most work. Quite frankly it sounds like a huge pain in the ass.

4. Some kind of statistical method where VR submitters provide their own cutoffs, and an aggregated cutoff is calculated.
Unfortunately the council does not have the capability to design such a system. If someone has an idea for such a method, please suggest it, and we will take it into consideration. Obviously it is too late to apply this method this year, so in this case the Council will make a decision this year, but for future years we will follow this method.

I like both, but I think we actually want a mix of these two.

I think the consistency of B- that's been used for the past couple of years for lower tiers is superior. It allows for the community to have a focused idea of how a VR should look. I don't think we should move on this: the problem is everything else. The consistency adds confidence to the rankings as has been seen, and having it change every year just makes people confused. B one year, C the next, what the hell, Jenkins, what is this?! We should have a clear system in place, and the B- Rank is a part of that. However, I do think that moving the definition of B- - which would effectively change it every year - would be better optically. You'll understand by the end of this, I promise.

You see, as I mentioned before, Slowbro ended up dropping because it is the OU cut-off for many players. Of course it's going to get ranked unusually low but always barely be away from C Rank. So let's take this into account.

We should more strongly take into account the Pokemon the users are designating as their OU cut-off. ​

We do not need to leave out this crucial information. Right now, the script lacks this context, and this is why Slowbro fell, as it literally did not know how we approach this Pokemon. The ranks are working, so we should move the Pokemon. If there are clusters that end up being too far to contain in our 3-slot ranking system, then we should use the S1 to B5 system, expanding the numbers as we go. So we look at everyone's cut-off, take the Pokemon directly above it, and shove it in some stupid algorithm.

So really, what I'm proposing is we synchronise the B- Rank with an effectively voted Pokemon to be used as the cut-off. The Rank would move where the cut-off Pokemon needs to be, effectively forcing it into the correct position. Voting on the worst OU-tiered Pokemon should always spit out Slowbro, and for lower-tiers, this could actually end up making for an interesting study. If the time comes for Slowbro to drop, then it would work, as people would shift to the next Pokemon up. The script needs to find the meaning to the B- Rank, so dictating what the worst is should do that, no?

Now there is obviously a danger of the voted Pokemon being unclear or something, unintentionally dropping something. Perhaps unless there is no supermajority, this system is disabled for that slate or something.

To summarise:

• The Pokemon at the bottom of everyone's OU cut-off should be more strongly taken into account when determining the final product. This would ensure the cut-off functions correctly.
• B- Rank, or a numerical equivalent like GSC's rankings and so on, should be defined by that Pokemon, to make the rankings consistent to someone who does not care to discuss these things ad infinitum.
• If there is no clear Pokemon to use - let's say we want a 60% supermajority like with a suspect test - the function is disabled.

This allows the script to consistently nail the cut-off, as it now has some form of rankings provided.

"PvK, isn't this the same purpose as the B- Rank?"

​

Sort of, but also not quite, but I'm having trouble articulating this in a way that makes sense. It's complicated. I guess what I'm saying is to redefine the B- Rank in a more precise way. I believe that the script's contextless, rankless system doesn't jive with a metagame this optimised, as we have indisputable opinions on certain subjects, such as our case with Slowbro. Ergo, we should provide a safety net that allows a cut-off to keep that Pokemon there, and then weigh the Pokemon above it more carefully, even if it is sometimes dipping into a lower tier. In this case, it isn't Slowbro being the problem - we know this because of player opinion - it's the other Pokemon being considered OU. Does that make sense?

This would improve vapicuno's method, but it also ends up being a bit weird. The problem vapicuno had is that everyone's standards are different, and thus, the system could break in some way. So, maybe it doesn't make sense to use it for lower tiers. It's why a rankless system is used in the first place; for example, some people see S as "A but a bit better", while I would personally see it as "mandatory on every team".

Regardless, this solution would answer the community's desires and solve the problem, in my eyes.

Also, for what it's worth, there have historically been blind voting slates like this one. Maybe this could help?

 

[vapicuno]

Purpose needs to be figured out first. Tiering methods, which can be statistical, can then be devised to suit the intended purpose. For example, deciding cutoffs in order to

1. Determine which mons you need to prepare for in an OU tournament - look at usage stats, find some high percentage cutoff, find the subtier that said mon is in, draw a cutoff below that subtier.
2. Determine which mons can be viably used in OU - ditto but lower percentage, or council based tiering
3. Tie it to some intuitive understanding of what is OU - select one or several reference OU mons and draw the line below the subtier they are in. The selection of such reference mons might be able to be inferred statistically from an OU usage table.
4. Tie it to some intuitive understanding of what is UU - select one or several reference UU mons and draw the line above the subtier they are in. The selection of such reference mons might be able to be inferred statistically from an OU usage table likewise.
5. Mimic usage-based tiering as per new gens - same as (1) with different cutoff.

If someone can provide me a list of usage stats for relevant tournaments, I can try to do some exploring.

 

[vapicuno]

Sorry for double posting, just felt like I had quite a bit more to share.

I am providing these graphs for information. It's probably best that I not make any suggestions until the community decides on the intended purpose of tiering. Based on the criterion above, I will elucidate what different tiering methods might involve or result in if we were to use usage stats

The distribution of usage stats tends to follow an exponential (trendline). Based on the criterion stated in my previous post, we could be looking out for

1. Cutoffs that have been used by Smogon in other tiers
2. Deviations (kinks) from the trendline, which may imply a cutoff

The following graph is an average of 4 years of SPL data. Errorbars are the peak-to-peak variation from 2019-2022. Note

1. The cutoff of 3.406% corresponds to the Pokemon appearing in 20 games with 50% probability.
2. The cutoff of 2.28% is for 30 games with 50% probability.
3. I was informed that these are the cutoffs for upgrading / downgrading a Pokemon in Smogon's usage-based tiers.
4. Notice the plateauing at Jolteon and dropoff thereafter to Slowbro. And the dropoff from Articuno to Persian. This could potentially be another cutoff depending on the aims of tiering.
5. From the peak-to-peak variation, the community could agree on several benchmark Pokemon that will always remain above some cutoff to remain in OU, or below some cutoff to remain in UU, and this can be used in tandem with the VR subtiers to delineate the boundaries.

These are all some ideas for tiering. Again, method chosen depends on intent, which needs to be figured out first imo.

This chart can be plotted for the various years as well (see spoilers below). I will continue using the trendline from the aggregate data across the years. Observations

• 2022: Significant dropoffs from Jolteon > Slowbro and Slowbro > Lapras
• 2021: Slowbro > Lapras
• 2020: Gengar > Lapras
• 2019: Gengar > Slowbro

2022

2021

2020

2019

 

[Amaranth]

I think there's a relatively common feeling amongst high level players that Slowbro is better and more impactful to OU than its usage reflects.

Determine which mons you need to prepare for in an OU tournament - look at usage stats, find some high percentage cutoff, find the subtier that said mon is in, draw a cutoff below that subtier.
Determine which mons can be viably used in OU - ditto but lower percentage, or council based tiering

It'd always be highly advisable to run either a fast Thunderbolt or multiple Explosion users to avoid getting ran over by Slowbro completely, even if it doesn't show up too often, because it is a mon that can easily 6-0 if you underprepare for it to a larger degree than anything else in the tier.
And Slowbro is always going to be viably used in OU, but it doesn't get brought too often in SPL because it puts a lot on match up / not getting crit and people don't like to bank their high stakes tour games on such things.

It could be argued that this is fair and that in all usage-based tiers this effect exists, but it's not quite the same pulling ladder stats vs tournament stats because the high stakes do end up warping team choices imo.
Basically I don't believe that SPL usage stats are a fair representation of how viable a pokemon is and how much you need to respect it in the builder.

 

[Hipmonlee]

Speaking for myself in this post, not council.

Thinking about this, in terms of the philosophical goal of the cutoff, we actually have a great de facto explanation already.

B+ Rank: Reserved for Pokemon who are very good in the RBY metagame, yet are somewhat inconsistent. These Pokemon may have flaws, but they have positive aspects which makes them stand out, making them valuable members of a team.

[...]

C+ Rank: Reserved for Pokemon that have solid niches in the RBY metagame, but nonetheless are generally inconsistent. These Pokemon have definite flaws, but may have positive aspects which can make them worth using.

I think this is as good a conceptual basis for when a Pokemon is OU and when a Pokemon is UU as we are ever going to find. And it quite nicely preserves the status quo, so we dont end up accidentally completely retiering RBY lower tiers.

So we would need to find a way to determine whether the VR suggests that Slowbro is very good, but somewhat inconsistent, or if it has a niche but is generally inconsistent.

 

[Ika Ika Musume]

Slowbro feels a lot more impactful than Lapras and garners significantly more respect from the OU community in both conversation and in the builder. Lapras has essentially zero respect in the builder. If people are going to mention that they are bringing a fast Electric-type for a Water, it is going to be for Starmie or for Slowbro. I think tiering based on feel to an extent is also valid.

People bring Lapras because it is safer and more consistent especially if you have better gameplay than your opponent. It has minorly negative matchup spreads against a lot of 6ths but very little that is unplayable. Lapras does not however and will never make the butthole clench like slowbro when it's starting to set up its amnesias and you are there praying for crits in this cruel cruel world or you will lose the game. Slowbros matchup spread is a lot more dynamic and Slowbro due to the way it is played can sometimes have low impact which is why it is generally avoided at tours like SPL. It is however clearly better than Lapras and shouldn't be grouped with it.

 

[vapicuno]

I think there's a relatively common feeling amongst high level players that Slowbro is better and more impactful to OU than its usage reflects.

Lapras has essentially zero respect in the builder. If people are going to mention that they are bringing a fast Electric-type for a Water, it is going to be for Starmie or for Slowbro.

Okay, I will reinterpret, correct me if I'm wrong - usage is agnostic to distinctiveness / uniqueness. Even if Lapras and Slowbro are used as much as each other, Lapras is more of the same Starmie, whereas Slowbro is unique kind of threat itself. This quality is neither captured in usage nor captured in the VR-based rankings.

So I stand corrected about my assumptions earlier about preparedness vs viability as higher vs lower percentages, because I had not considered uniqueness.

1. Determine which mons you need to prepare for in an OU tournament - look at usage stats, find some high percentage cutoff, find the subtier that said mon is in, draw a cutoff below that subtier.
2. Determine which mons can be viably used in OU - ditto but lower percentage, or council based tiering

As for this,

I think this is as good a conceptual basis for when a Pokemon is OU and when a Pokemon is UU as we are ever going to find. And it quite nicely preserves the status quo, so we dont end up accidentally completely retiering RBY lower tiers.

1. This means you are in favor of drawing cutoffs from an OU-centric point of view.
2. VR rankings are presumably made by players based on subjectively weighted considerations of niches, consistency, and other factors, so extracting this information based on VRs alone is impossible, and based on other data like tour stats, I can't think of a solution right now.
3. You can consider using Slowbro as (one of) the "reference OU benchmark(s)", but note the implications: Slowbro (as it appears to me) earns its eternal place in OU from the dimension of uniqueness, which can make it a bad yardstick for the mons nearby, since Lapras is basically a boring fella that does some RBY things, comparing apples to oranges.
4. If the council is unanimous that Slowbro is OU, then the question is really not about whether Slowbro is OU, but rather whether it makes sense to compare other mons around B-/C+ to Slowbro, and if not, to what?

 

[Hipmonlee]

1. This means you are in favor of drawing cutoffs from an OU-centric point of view.
2. VR rankings are presumably made by players based on subjectively weighted considerations of niches, consistency, and other factors, so extracting this information based on VRs alone is impossible, and based on other data like tour stats, I can't think of a solution right now.
3. You can consider using Slowbro as (one of) the "reference OU benchmark(s)", but note the implications: Slowbro (as it appears to me) earns its eternal place in OU from the dimension of uniqueness, which can make it a bad yardstick for the mons nearby, since Lapras is basically a boring fella that does some RBY things, comparing apples to oranges.
4. If the council is unanimous that Slowbro is OU, then the question is really not about whether Slowbro is OU, but rather whether it makes sense to compare other mons around B-/C+ to Slowbro, and if not, to what?

1. Yeah, pretty much. This is the philosophically pure way to do it.

3. I think the dimension of uniqueness is a tricky one, because Victreebel is also a very important mon to at least know about, because there isnt much else like it. There's a lot of factors making Bro where it is: other stuff has more noticeably moved up around it, I personally think it is just generally underated, and Bro is high risk, high reward, and a lot of players dont like that. Its not just uniqueness that puts it in OU in my view.

2. What I like about the VRs is that it is somewhat agnostic to all of that. It doesnt matter what factors you use to rank mons, you just rank them, and we get an aggregated result. We trust the wisdom of the crowds to mash all of that together.

So, in that spirit, after a convo with pvk and based on her system, here is what I propose:

So, accepting we have the philosophy more or less square, that things with a solid niche but that are not very good outside of that niche are considered UU. What we need is a system that gives us a practical community consensus on "exactly how niche is too niche?"

So, our VR submitters each draw their personal line on their own VR. This serves as essentially a pokemon by pokemon vote on whether each mon should be OU. Everything above the line is voted OU and everything below the line is voted UU.

Then, for the overall VR we have a few options:

1a. Everything above the highest ranked Pokemon that is consensus UU is OU.
1b. Everything above the rank of the highest ranked Pokemon that is consensus UU is OU.
2a. The lowest ranked Pokemon that is consensus OU and everything above it is OU.
2b. The rank of the lowest ranked Pokemon that is consensus OU and everything above it is OU.
3a. We aggregate the votes for every Pokemon in each rank and every Pokemon in a rank above the highest rank that is consensus UU is OU.
3b. We aggregate the votes for every Pokemon in each rank and every Pokemon in the lowest rank that is consensus OU and above is OU.

Personally, I really hate the idea of splitting ranks.

 

[Hipmonlee]

Ok, having put a little thought into it, I have decided that 2b is the best solution (again, my opinion, not council's).

Firstly, it doesnt split ranks. If we cant say concretely that one Pokemon is more viable than another, then those Pokemon should be in the same tier.

Secondly, it ensures that every Pokemon that is considered OU by the community is OU. This also does mean Pokemon that the community considers UU will be OU, but I think it is correct to err on the side of making things OU.

The tiers serve two purposes:
1. They designate what pokemon cannot be used in UU. But, as we have discussed here, cutoffs are pretty much arbitrary. UU will adapt to whatever cutoff it is presented with, and, the consensus seems to be that that is the correct way to handle it.

2. They inform new users about what Pokemon they can expect to see in OU battles. And I think it should be reasonably uncontroversial that it is better to warn people about things they dont encounter, than it is to not warn people about things they do encounter.

Putting all that together, there is only one option that makes sense.

 

[Useless Uses]

So in my opinion, the only good answers to this problem are 2a and 2b. I'm not really sure which one would be healthier for the tiering system. I'd go with 2a, so anything widely considered better than Slowbro is OU material. (Excluding Slowbro of course)

Also depending on how Lapras shapes UU, it could stay as UUBL.

 

[Hipmonlee]

Yes, I am out of town though, I will post later on in the week. Short version is bro is OU for now, next time you do VRs indicate your cutoff.

 

[phoopes]

This may be out of scope of this thread (maybe should lock it at this point) but will the same be planned for lower tiers? Aka will the UU/NU line be similarly decided?

I think we can probably do case-by-case basis because this is a fairly unique situation. For the UU VR which is upcoming in a few months though I don't think it would be a terrible idea as to where to indicate cutoff though just in case something like this happens again where something is very clearly below UU but very clearly above NU (like this could very well happen with, say, Charizard, as I've seen some players rank it bottom of UU while nearly everyone ranks it in the top 3 in NU). I guess I could talk to vapicuno/Shellnuts/whoever is compiling the rankings at the time to see what they think/how feasible it is.

Good question!

 

[juoean]

i dont really understand what ‘this is a fairly unique situation’ means.
the vapicuno method is a way of ordering mons by rank (“ordered set” whatever) along with a set of conditions limiting where rank and subrank divisions might or might not be drawn consistent with the voting.
it doesnt in itself establish where the tier cutoff should be, thats why this thread was created.

i dont play ou but in this case it seems that there was only one mon in question, for various reasons. in uu which as i understand is going to be retiered within the next month or two, there is not remotely consensus about anything once you get past #11 or so, including about the tier cutoff. there are some players who think that the tier should be in the #11-13 mons range, some who see #26-28 or so mons as uu, and probably others anywhere in between; and those of us who are v unsure what we think about that question.

personally i def think having an at least mostly clear process for the tier cutoff before the uu vr submission process begins, rather than figuring it out afterward, is preferable to say the least.

~ ~ ~
this last part is rambly sorry trying my best w writing

i do want to note a couple things about any form of “lowest pokemon that is consensus in the tier” standard. these are observations / neutral-descriptive thoughts unless otherwise indicated (eg ‘i think’ ‘should’). first, that is going to be significantly influenced by whether voters choose to draw the tier cutoff lines higher or lower. if consensus means unanimous, then one voter drawing the cutoff much higher than others would drastically impact the size of the tier. if it doesnt mean unanimous, then the threshhold should be set before the vr process begins. second, if using the variation where the line is determined based on the tier that contains the lowest consensus [uu] mon, which i do think makes more sense than splitting up a tier, lets say for example that dodrio were to be the lowest consensus-uu mon (which i think is plausible, dodrio is mostly agreed as #11 and cutting off the tier at that place is not unreasonable, plus mons below that vary a lottt in where ppl rank them), dodrio is a mon that could easily end up either in the ‘A’ ranks with the mons typically ranked above it (eg gyara vap), or where it is in the current vr at the top of ‘B’ ranks. whether dodrio ends up in ‘A’ or ‘B’ ranks is in part independent of whether it is the lowest consensus-uu mon (by whatever metric of consensus). bc we each submit our own tier cutoffs individually.
if dodrio were to be the lowest consensus-uu mon, and by policy whatever rank contains dodrio were the lowest rank of mons that is tiered as uu, then obviously whether dodrio were ranked with mons above it or below it would majorly change where the tier cutoff is.
(or if dodrio is in its own tier from the vr process like what happened with slowbro.)

to be clear maybe this is good, if dodrio is the lowest consensus vr but is tiered top of “B” maybe that indicates that the uu tier should in fact be a bit larger (it was closer to the stuff below it than the stuff above, but maybe the reason it ended up as the lowest-consensus uu was because the mons below it are controversial and so a few ppl ranked each one of them quite a bit lower even if no one actually had #11 as their lowest uu-mon), while if its tiered bottom of “A” that maybe indicates that a smaller uu tier is in fact accurate to how ppl see the metagame. im j trying to use some hypothetical possibilities so we can think about what might happen and evaluate if this ‘matches’ how we envision tier cutoffs to work; or if it doesnt then maybe it suggests some thing(s) we might want to incorporate in the tier cutoff aspect of the vr process. idk