Rebooting Smogon RBY

[Mirabel_]

While Showdown's RBY OU has been steadily growing in popularity for a good 12 months, there was an explosive influx of new users climbing the ladder from February onward, which is now gradually subsiding (almost 16000 matches played in March alone). I attribute that to the RBY re-release.

We now have a recent usage sample size to draw from that most Old Gens can only dream of one day obtaining. Armed with this new information, I took to evaluating the eligibility of every mon with a presence in RBY OU at the 1760 rating-weight threshold.

I did this because...
1) The usage-based tiers of Smogon RBY were drawn when the game was being simulated incorrectly over a long period of time, and include mons that occasionally or perpetually are out of usage. RBY UU and especially BL are for this reason very illegitimate, and that partially stems from what might be an incorrectly drawn OU.
2) We may never get an opportunity like this again. RBY may very well fade into obscurity after this spike.
3) I wanted to see if we could argue for a ban by ubiquity (96.6%) on the big 4 or something. (edit: I'm not saying I want this or think it'd be smart)

So here are the mons that you have had a decent chance of seeing in any given day of battling, by Smogon's official modern definition of OU. (~3.41%)

16 Pokemon that stayed OU for May 2015 to April 2016:
Tauros, Chansey, Exeggutor, Snorlax, Alakazam, Starmie, Jynx, Lapras, Zapdos, Golem, Gengar, Rhydon, Slowbro, Dragonite, Cloyster, Jolteon

So those 16 mons are not up for evaluation; they're to be permanently OU-designated, and as it stands are all in the previous RBY OU as well. What about those stragglers, though?

7 OU visitors since May 2015:
-Victreebel (8/12 months)
-Persian (7/12 months)
-Moltres (3/12 months)
-Articuno (1/12 months)
-Machamp (1/12 months)
-Blastoise (1/12 months)
-Ninetales (1/12 months)

Obviously we can't just tally the months up to split them into OU and UU, as May to December 2015 combined is worth less than March 2016. So let's just scale the %s up by each month's approximate weight. (March and April were by far the heaviest months).

-Victreebel 3.631%
-Persian 3.326%
-Articuno 1.535%
-Machamp 0.847%
-Moltres 0.611%
-Ninetales 0.356%
-Blastoise 0.149%

Now if we were retiering, which I think we should really consider doing, this is how it would look. (note, BL is effectively part of UU)

-Victreebel remains OU, bringing the total count to 17 overused mons.
-Persian and Articuno drop to UU.
-Blastoise, Moltres, Machamp, and Ninetales all remain UU.


Conclusions

Yes, we should re-tier RBY with usage stats on an accurate simulator (it'd be the first time Smogon touched RBY, technically), but - I think we need another month, maybe three. Wait for the huge wave to recede a little further before we change anything. Persian dropped out of OU because it's had two (very heavy) bad months lately, and Victreebel climbed higher up due to April being a particularly good month for it. Victreebel tends to hover above and Persian right below the threshold, and we can decide later what it means for it.

Please use teams independently of this project in the meantime. Not that it'll make a large difference if you lose rating over it as bel and cat do a lot better when they're a surprise anyway.

On a side note, no mon has ever even touched that ubiquity threshold I set. Tauros is still OU by me unless you guys think it's unhealthy gameplay or something.

 

[The Immortal]

I think the idea of tier shifts based on long term usage stats for past gens is potentially a good one. Obligatory Antar tag.

 

[HereToHelpRBY]

Should definitely use weighted usage stats too if possible. Pretty clear, if machamp and stuff are at points getting above the cutoff, that there should be some weighting. Edit: you're already doing that

Also I know Jellicent wanted to start moving on this issue.

Banning any of the big 4 is ridiculous. Leave it untouched.

-

Edit2: Also whilst there's a large playerbase on PS, how much of it will take part on here? And will they be invested? Some top players are already involved in tiering underway elsewhere.

 

[Mirabel_]

Also whilst there's a large playerbase on PS, how much of it will take part on here? And will they be invested? Some top players are already involved in tiering underway elsewhere.

As far as I know this form of tiering doesn't require top players or structured tournaments, just tallies on what everyone's seeing in high-rated blind matches on a crowded ladder over a long period of time.

 

[HereToHelpRBY]

As far as I know this form of tiering doesn't require top players or structured tournaments, just tallies on what everyone's seeing in high-rated blind matches on a crowded ladder over a long period of time.

that comment is more about exploring lower metas than the initial tiering.

Disputably, Articuno and Moltres may or may not make the cutoff and more importantly may or may not be broken in BL or UU (whatever it's being called now) but you need good players interested to make those decisions and structure the process and stuff.

 

[HereToHelpRBY]

Also would there be a ladder? And would the lower tiers be active enough that usage would work with them?

 

[Mirabel_]

Also would there be a ladder? And would the lower tiers be active enough that usage would work with them?

If you ask me, once we determine OU we can talk about getting an accompanying RBY UU ladder and figuring out who's voting on BL. We don't even know where OU's activity will be at that point since there's not much of a precedent for our situation.

 

[Ortheore]

Something something check out pokemon perfect. This is genuinely relevant, since we're currently in the process of tiering rby, so there's lots of cool insights to be had

While Showdown's RBY OU has been steadily growing in popularity for a good 12 months, there was an explosive influx of new users climbing the ladder from February onward, which is now gradually subsiding (almost 16000 matches played in March alone). I attribute that to the RBY re-release.

We now have a recent usage sample size to draw from that most Old Gens can only dream of one day obtaining. Armed with this new information, I took to evaluating the eligibility of every mon with a presence in RBY OU at the 1760 rating-weight threshold.

I did this because...
1) The usage-based tiers of Smogon RBY were drawn when the game was being simulated incorrectly over a long period of time, and include mons that occasionally or perpetually are out of usage. RBY UU and especially BL are for this reason very illegitimate, and that partially stems from what might be an incorrectly drawn OU.

I was under the distinct impression that RBY has never ever seen usage based tiering. Unless I'm woefully mistaken, smogon's rby tiers are based off rby2k10's which definitely aren't usage based. Same goes for pokemon perfect. Personally I'm not really a fan of usage based tiering in general, but I might talk about that below

2) We may never get an opportunity like this again. RBY may very well fade into obscurity after this spike.
3) I wanted to see if we could argue for a ban by ubiquity (96.6%) on the big 4 or something. (edit: I'm not saying I want this or think it'd be smart)

As much as you could make a theoretical argument here, I think I speak for the vast majority of people in saying that we really do not want this. On top of that, I don't think ubiquity is a good basis for banning. Maybe it would be in newer gens where the playerbase has different expectations in terms of diversity, but for old gens I think most people have the perspective that some ubiquity isn't inherently bad provided there's other room for diversity, the meta's still healthy and enjoyable, and so on (also helps that while the big 4 are justifiably very common, the only one I personally approximately mandatory is Tauros). Furthermore, I have doubts about whether it's possible to avoid having these ubiquitous pokemon in the meta- every level of rby tiering done on pokemon perfect thus far has seen at least one pokemon that's pretty much ubiquitous without being broken (not experienced enough to comment on rby2k10's lists, except I will note that water types in UU are ridiculous). Further bans I think would simply mean something else fills the void. I guess it all boils down to the fact that there just aren't as many options in rby. As a bit of theorymon, Zam sticks out as being more than capable of filling the void left by a potential big 4 ban as the new ubiquitous mon.

So here are the mons that you have had a decent chance of seeing in any given day of battling, by Smogon's official modern definition of OU. (~3.41%)

16 Pokemon that stayed OU for May 2015 to April 2016:
Tauros, Chansey, Exeggutor, Snorlax, Alakazam, Starmie, Jynx, Lapras, Zapdos, Golem, Gengar, Rhydon, Slowbro, Dragonite, Cloyster, Jolteon

So those 16 mons are not up for evaluation; they're to be permanently OU-designated, and as it stands are all in the previous RBY OU as well. What about those stragglers, though?

7 OU visitors since May 2015:
-Victreebel (8/12 months)
-Persian (7/12 months)
-Moltres (3/12 months)
-Articuno (1/12 months)
-Machamp (1/12 months)
-Blastoise (1/12 months)
-Ninetales (1/12 months)

Obviously we can't just tally the months up to split them into OU and UU, as May to December 2015 combined is worth less than March 2016. So let's just scale the %s up by each month's approximate weight. (March and April were by far the heaviest months).

-Victreebel 3.631%
-Persian 3.326%
-Articuno 1.535%
-Machamp 0.847%
-Moltres 0.611%
-Ninetales 0.356%
-Blastoise 0.149%

Now if we were retiering, which I think we should really consider doing, this is how it would look. (note, BL is effectively part of UU)

-Victreebel remains OU, bringing the total count to 17 overused mons.
-Persian and Articuno drop to UU.
-Blastoise, Moltres, Machamp, and Ninetales all remain UU.


Conclusions

Yes, we should re-tier RBY with usage stats on an accurate simulator (it'd be the first time Smogon touched RBY, technically), but - I think we need another month, maybe three. Wait for the huge wave to recede a little further before we change anything. Persian dropped out of OU because it's had two (very heavy) bad months lately, and Victreebel climbed higher up due to April being a particularly good month for it. Victreebel tends to hover above and Persian right below the threshold, and we can decide later what it means for it.

Please use teams independently of this project in the meantime. Not that it'll make a large difference if you lose rating over it as bel and cat do a lot better when they're a surprise anyway.

On a side note, no mon has ever even touched that ubiquity threshold I set. Tauros is still OU by me unless you guys think it's unhealthy gameplay or something.

The first thing that sticks out to me is that Toise, Tales and Champ are in this conversation at all. We all know that these pokemon are pure shit, and that they're even in this discussion speaks volumes of the quality of the stats being used. I am aware that unfortunately there's little to be done about it, as if you restrict the data to people that have a higher rating and therefore presumably know what they're doing, you limit your sample size and run into other issues in terms of accuracy, so some compromise has to be made. I guess this is just one of my complaints about usage-based tiering- it's influenced by people that don't know what they're doing, something that imo is especially relevant for RBY because there a lot of bad teams out there (and it's really obvious too).

Secondly, you're describing a long-term tier list that's heavily influenced by short-term trends. Unless you propose we maintain and active and ongoing tiering process similar to current gen metas, this will mean that we have a tier list that is merely a snapshot in time rather than something that genuinely reflects the capabilities of various pokemon. If you wanna talk about legitimacy, there are serious questions to be asked here. There's a very notable distinction to be had between short-term trends and genuine long term shifts in our understanding of the meta

Lastly, I find it pretty amusing that something as terrible as Rhydon gets straight into OU

 

[Mirabel_]

Something something check out pokemon perfect. This is genuinely relevant, since we're currently in the process of tiering rby, so there's lots of cool insights to be had
I was under the distinct impression that RBY has never ever seen usage based tiering. Unless I'm woefully mistaken, smogon's rby tiers are based off rby2k10's which definitely aren't usage based. Same goes for pokemon perfect. Personally I'm not really a fan of usage based tiering in general, but I might talk about that belowAs much as you could make a theoretical argument here, I think I speak for the vast majority of people in saying that we really do not want this. On top of that, I don't think ubiquity is a good basis for banning. Maybe it would be in newer gens where the playerbase has different expectations in terms of diversity, but for old gens I think most people have the perspective that some ubiquity isn't inherently bad provided there's other room for diversity, the meta's still healthy and enjoyable, and so on (also helps that while the big 4 are justifiably very common, the only one I personally approximately mandatory is Tauros). Furthermore, I have doubts about whether it's possible to avoid having these ubiquitous pokemon in the meta- every level of rby tiering done on pokemon perfect thus far has seen at least one pokemon that's pretty much ubiquitous without being broken (not experienced enough to comment on rby2k10's lists, except I will note that water types in UU are ridiculous). Further bans I think would simply mean something else fills the void. I guess it all boils down to the fact that there just aren't as many options in rby. As a bit of theorymon, Zam sticks out as being more than capable of filling the void left by a potential big 4 ban as the new ubiquitous mon.

This really has very little to do with Pokemon Perfect as it does not affect their tiering, will require very little attention from the same player base as it develops throughout, and is based on a different philosophy, data set, AND competition format (ladder vs tournament). Although, if it is me who is, as you say, woefully mistaken and I am pushing the first usage based tiering in RBY history in order to create a less embarassing Smogon UU when we could just implement PP 2U as our own and call it a day (at least it'd be an improvement, and Smogon would be breaking fewer of its old gens design paradigms) then more power to those opposed. Despite that potential misconception of mine I still believe there is merit to usage-based tiering when limited to situations like our current one, and its drawbacks can be balanced out.

I also think the deliberate viability-based metas like PP's are limited in what information they can automatically receive from the showdown server, and unfortunately they would run into all the same bellcurve issues as I tried to mitigate here and you mention, unless their data was from as small a sample size as monthly deep-bracket tournament results.

I also believe ubiquity is not inherently bad unless the mon reduces the depth of the game at every layer. But the thought of it was part of what motivated me to gauge 12 months of usage and draw conclusions compatible with what i thought to be smogon's tiering philosophy so I thought it'd be honest to throw it in the intro.

And yeah I was priorly aware a Tauros ban would just power the special tanks up.

Can you clarify what precisely the vast majority of the community does not want? I assume you are referring to a switch to recent-usage-based tiers from smogon's current bizarre tiers.

The first thing that sticks out to me is that Toise, Tales and Champ are in this conversation at all. We all know that these pokemon are pure shit, and that they're even in this discussion speaks volumes of the quality of the stats being used. I am aware that unfortunately there's little to be done about it, as if you restrict the data to people that have a higher rating and therefore presumably know what they're doing, you limit your sample size and run into other issues in terms of accuracy, so some compromise has to be made. I guess this is just one of my complaints about usage-based tiering- it's influenced by people that don't know what they're doing, something that imo is especially relevant for RBY because there a lot of bad teams out there (and it's really obvious too).

As everyone knows and the stats pages verify, these weirdmons get ironed out as the sample size increases. I included them because I only considered mons that had a chance of breaking OU even as a flash in an empty pan. At a glance I immediately knew the birds and the weirdmons below them weren't going to be remotely significant after the heavier months were factored in because those recent months had the smallest and most predictable bunch of OU mons, but I just followed through with my intention to include any mon that passed the first test: having ever crossed the OU line.

Secondly, you're describing a long-term tier list that's heavily influenced by short-term trends. Unless you propose we maintain and active and ongoing tiering process similar to current gen metas, this will mean that we have a tier list that is merely a snapshot in time rather than something that genuinely reflects the capabilities of various pokemon. If you wanna talk about legitimacy, there are serious questions to be asked here. There's a very notable distinction to be had between short-term trends and genuine long term shifts in our understanding of the meta

I don't think we'll be switching mons in or out of OU a lot. Brace for theorymon, but actually I expect RBY OU to die back down to the point of obscurity, until our results over any 6 months are so scattered and dwarfed that we have 30-40 OU mons at 1760 weighting, Tauros is at 55% usage, and people are once again asking particular known RBY players of vastly different rating to jump into the ladder for a game, further skewing the rosters and illegitimizing further usage-based tiering action. I don't know when or if this will be the state of things, but I want to harness this current wave to replace what Smogon has now - an overly presumptive conversion of a tiering apparently inherited from a wrapless community using whatever wrong sim - with something that the current Smogon and RoA would accept as a modern approach to competitively housing every mon in what we all hope is the real RBY this time around. If we as a community want to outsource that task to Pokemon Perfect, who has done a fantastic job so far and whose solution is ready to go right now, then sure I'm on board. Or we could just do what Smogon and Showdown are best equipped to do and known for, keep it up just until RBY recedes again, and call it our own. I have no personal stake in either position but for what my opinion is worth I'd like to keep both approaches alive (as I value them both) and therefore seperate.

Lastly, I find it pretty amusing that something as terrible as Rhydon gets straight into OU

Yeah I think that weirds everyone out a bit. That niche, people aren't persuaded to let go of it for ladder teams.

 

[Isa]

very minor detail but ortheore you got it backwards. rby2k10 tiers came after smogon's tiers, then when PO added lower tiers (with rby2k10 tiers as basis) hipmonlee removed the BL tier from smogon. it seems now that it's back online on the dex though...?

ps rhydon deserves ou lol

 

[Ortheore]

This really has very little to do with Pokemon Perfect as it does not affect their tiering, will require very little attention from the same player base as it develops throughout, and is based on a different philosophy, data set, AND competition format (ladder vs tournament). Although, if it is me who is, as you say, woefully mistaken and I am pushing the first usage based tiering in RBY history in order to create a less embarassing Smogon UU when we could just implement PP 2U as our own and call it a day (at least it'd be an improvement, and Smogon would be breaking fewer of its old gens design paradigms) then more power to those opposed. Despite that potential misconception of mine I still believe there is merit to usage-based tiering when limited to situations like our current one, and its drawbacks can be balanced out.

I also think the deliberate viability-based metas like PP's are limited in what information they can automatically receive from the showdown server, and unfortunately they would run into all the same bellcurve issues as I tried to mitigate here and you mention, unless their data was from as small a sample size as monthly deep-bracket tournament results.

Yeah I think you hit the nail on the head in terms of differences in philosophy, perhaps I was being a bit off-base in drawing a strong parallel between the two systems. I also think I might've been trying to assert my own perspective as being more valid than others, which is a bit of a silly thing to do here, especially that any given approach has strengths and drawbacks, which means that it isn't a matter of being right and wrong, but instead what traits do you value more strongly, it's just I think I sometimes let that get away from me because it seems I'm biased as fuck on this kinda thing.

One area where the PP comparison is good is looking at how metas tend to form in rby tho, which is completely different from this whole question of philosophies but is still of interest if any tiering project were to be undertaken. RBY2K10 would also be of interest, but idk enough about it.

Regarding your stuff on pp's approach and how it relates to collecting data, I think we agree that they're of limited compatibility, especially because of the tour-ladder distinction. I just don't think tournaments can generate enough data to directly use in decision making- off the top of my head RBY in SPL generates a good amount of data by the standard of tournaments, but it's still nowhere near enough to decide on the truly borderline things that don't see huge amounts of usage. That's to say nothing of the fact that collecting data from tours is a lot more problematic.

Also ty Isa for correcting me

I also believe ubiquity is not inherently bad unless the mon reduces the depth of the game at every layer. But the thought of it was part of what motivated me to gauge 12 months of usage and draw conclusions compatible with what i thought to be smogon's tiering philosophy so I thought it'd be honest to throw it in the intro.

And yeah I was priorly aware a Tauros ban would just power the special tanks up.

Can you clarify what precisely the vast majority of the community does not want? I assume you are referring to a switch to recent-usage-based tiers from smogon's current bizarre tiers.

Regarding the majority of the community, I was referring specifically to changes to ou. Beyond that I definitely can't say.

As for stuff like smogon's tiering philosophy again I'm not sure how that applies to old gens if only because my understanding is that gen 4 brought a radical shift in the ideology. Idk how well it applies to old gens. I do agree that if the current mentality were applied to RBY and questions of ubiquity there would probably be bans coming into play, it's just I don't think that applying those current gen mindsets to RBY is ultimately a great way of going about it due to simply having fewer options available.

As everyone knows and the stats pages verify, these weirdmons get ironed out as the sample size increases. I included them because I only considered mons that had a chance of breaking OU even as a flash in an empty pan. At a glance I immediately knew the birds and the weirdmons below them weren't going to be remotely significant after the heavier months were factored in because those recent months had the smallest and most predictable bunch of OU mons, but I just followed through with my intention to include any mon that passed the first test: having ever crossed the OU line.

I think this is the point where my bias caught up with me and I started making shitty points, because those shitmons are a bit of a red herring (tho why is Moltres lumped in with them? It's not that bad). Probably a more sensible line of critique is whether truly borderline pokemon such as Victreebel are having their positions swung one way or another by a disproportionate amount of dodgy data points. Then again, you say the stats are weighted, which probably accounts for a lot of that?

I don't think we'll be switching mons in or out of OU a lot. Brace for theorymon, but actually I expect RBY OU to die back down to the point of obscurity, until our results over any 6 months are so scattered and dwarfed that we have 30-40 OU mons at 1760 weighting, Tauros is at 55% usage, and people are once again asking particular known RBY players of vastly different rating to jump into the ladder for a game, further skewing the rosters and illegitimizing further usage-based tiering action. I don't know when or if this will be the state of things, but I want to harness this current wave to replace what Smogon has now - an overly presumptive conversion of a tiering apparently inherited from a wrapless community using whatever wrong sim - with something that the current Smogon and RoA would accept as a modern approach to competitively housing every mon in what we all hope is the real RBY this time around. If we as a community want to outsource that task to Pokemon Perfect, who has done a fantastic job so far and whose solution is ready to go right now, then sure I'm on board. Or we could just do what Smogon and Showdown are best equipped to do and known for, keep it up just until RBY recedes again, and call it our own. I have no personal stake in either position but for what my opinion is worth I'd like to keep both approaches alive (as I value them both) and therefore seperate.

I'm honestly unsure what my opinion on this kinda stuff is, especially given that my perspective is of someone fairly heavily invested in PP, which is obviously gonna differ a little from plenty of other ppl who might read this. Again, I might be making a mountain out of a mole hill here, since we all know that if there are any metagame shifts that occur, it'll probably be slowly enough that the tier list can last a good few years before being out of date or anything like that. Not sure how usage will translate to lower tiers tho, especially given that it'll take a good bit of time to progress down through the lower tiers which may cause problems if rby's surge in ladder popularity is too fleeting

 

[Lutra]

I appreciate trying to use these usage statistics to be more objective, but I think the danger is cherry-picking a process that meets in-line with already existing expectations. I'd say the problem (for all gens), that as good as the usage stats are now compared to the past, is that we still don't have in-depth enough usage statistics to outweigh the relevant observations players make in battles.

Ultimately as things stand now, I feel that if we gave enough time to a RBY council, say 2 months at every big tiering stage (every Used tier), and 1 month at minor stages (banning of 1 category of moves, 1 Pokémon etc), we could make a good tier list. I'm thinking like a ladder and old gen room tours held at each tiering stage to gain the opportunity to vote borderline Pokémon in and out the tier and ban stuff - along with the elected (democratically?) members of the council. I'd start the tiering from Anything Goes and make it completely transitive (lower tiers inheriting all bans from upper tiers).

In the future, with enough advancement in usage stats and AI, then I think you could do an automated process, but I just think with what we have at the moment, it's no good for a limited playerbase playing with a limited Poképool.

 

[Shax]

We need a GSC and RBY Ubers tier on the ladder. It promotes a healthy meta game in the long run. Or we should consider a RBY OU metagame with Fissure,and,Sheer Cold, Horn Drill.

 

[HereToHelpRBY]

Just for clarity, Pokemon Perfect's RBY tiers [In Smogon Terms]:
Uber - Mew, Mewtwo
OU - Alakazam, Chansey, Cloyster, Dragonite, Exeggutor, Gengar, Golem, Jolteon, Jynx, Lapras, Slowbro, Snorlax, Starmie, Tauros, Zapdos
BL - Articuno, Moltres
UU - Dodrio, Gyarados, Haunter, Hypno, Kadabra, Kangaskhan, Persian, Poliwrath, Raichu, Raticate, Rhydon, Tentacruel, Victreebel
RU - Aerodactyl, Charizard, Clefable, Dewgong, Dugtrio, Electabuzz, Exeggcute, Fearow, Golduck, Kingler, Mr. Mime, Pinsir, Poliwhirl, Sandslash, Vaporeon, Venusaur, Wigglytuff
NU - Abra, Arcanine, Blastoise, Nidoking, Nidoqueen, Omastar, Porygon, Rapidash, Scyther, Slowpoke, Staryu, Tangela, Venomoth
PU - Everything Else

Basically how it works/worked:
In OU a council was made [Disaster Area / Golden Gyarados / Lutra / Marcoasd / Ortheore] to vote on what's OU or not. Rhydon got 1/5 votes, Jolteon got 3/5 votes, everything else got 5/5 votes.
In UU a similar council system was used to ban Articuno and Moltres. The system of voting got adapted at this point, and Viability Rankings were used. Everything in B rank and above became UU, everything in D Rank and below became RU, and everything in C rank was voted on. Only Raticate was voted UU.
In RU and beyond, the Viability Rankings system was used. The top 5 from a tournament were used as an effective council should there be a tie in the votes for any particular Pokémon.

-

Obviously it's up to Smogon's staff to decide what's more appropriate - to use usage, to use Pokemon Perfect's tiers to some extent, or to form their own tiers in a non-usage based fashion so that Isa gets to keep Rhydon in OU. If you don't use PP's tiers, how will whoever manages the process drum up enough support to make either selected process effective?

Also Shax, firstly that's a second conversation and secondly I don't think most players want OHKO moves permitted or even tested; I'd make another thread to discuss that to avoid derailing this one.

 

[pharaohcalvin]

Just for clarity, Pokemon Perfect's RBY tiers [In Smogon Terms]:
Uber - Mew, Mewtwo
OU - Alakazam, Chansey, Cloyster, Dragonite, Exeggutor, Gengar, Golem, Jolteon, Jynx, Lapras, Slowbro, Snorlax, Starmie, Tauros, Zapdos
BL - Articuno, Moltres
UU - Dodrio, Gyarados, Haunter, Hypno, Kadabra, Kangaskhan, Persian, Poliwrath, Raichu, Raticate, Rhydon, Tentacruel, Victreebel
RU - Aerodactyl, Charizard, Clefable, Dewgong, Dugtrio, Electabuzz, Exeggcute, Fearow, Golduck, Kingler, Mr. Mime, Pinsir, Poliwhirl, Sandslash, Vaporeon, Venusaur, Wigglytuff
NU - Abra, Arcanine, Blatsoise, Nidoking, Nidoqueen, Omastar, Porygon, Rapidash, Scyther, Slowpoke, Staryu, Tangela, Venomoth
PU - Everything Else

Basically how it works/worked:
In OU a council was made [Disaster Area / Golden Gyarados / Lutra / Marcoasd / Ortheore] to vote on what's OU or not. Rhydon got 1/5 votes, Jolteon got 3/5 votes, everything else got 5/5 votes.
In UU a similar council system was used to ban Articuno and Moltres. The system of voting got adapted at this point, and Viability Rankings were used. Everything in B rank and above became UU, everything in D Rank and below became RU, and everything in C rank was voted on. Only Raticate was voted UU.
In RU and beyond, the Viability Rankings system was used. The top 5 from a tournament were used as an effective council should there be a tie in the votes for any particular Pokémon.

-

Obviously it's up to Smogon's staff to decide what's more appropriate - to use usage, to use Pokemon Perfect's tiers to some extent, or to form their own tiers in a non-usage based fashion so that Isa gets to keep Rhydon in OU. If you don't use PP's tiers, how will whoever manages the process drum up enough support to make either selected process effective?

Also Shax, firstly that's a second conversation and secondly I don't think most players want OHKO moves permitted or even tested; I'd make another thread to discuss that to avoid derailing this one.

I do not see why we would need any tier below UU for RGBY. The tiers are already super small.

 

[Lutra]

The idea of tiering is to make every Pokémon viable in a tier, so the tiers being super small necessitates more. Newer generations don't necessarily have more tiers - e.g. Generation A has 10 tiers of 60 each, with 600 total Pokémon, while Generation B has 8 tiers of 100 each, with 800 total Pokémon. The total Pokémon in Generation B is larger, but there's less tiers.

 

[Peasounay]

Also it can't really be bad to have a lot of playable tiers. Yes obviously they're small but they're still all different, balanced, and fun. They were tested a lot. Saying there's no need to go down below UU is a bit of a limited vision (imo, no offense) and is purely psychological, it doesn't hurt/harm anything.

 

[teal6]

PP should probably be taken into account in some way if making any changes. The main reason for RBY Ladder increases is the RBY Ladder Tour, I've almost never laddered RBY in my life (just thought it was dead) but played about 50 games myself during the first cycle. Some others did many more.

 

[HereToHelpRBY]

I do not see why we would need any tier below UU for RGBY. The tiers are already super small.

Actually an interesting point to make here is that each tier is of a relatively similar size. OU has 15 Pokémon; UU and NU have the least each with 13 Pokémon whilst RU has the most with 17. So it seems you can carry on for at least a couple more tiers before anything weird happens. And initial playtesting of PU suggests it'll continue to fit this trend.

 

[Antar]

Happy to assist in re-tiering based on usage. If we have a large enough training set (>6 months), I'd recommend performing a general linear regression to get the coefficients (going maybe 6 months out instead of three)?

For those of you who don't speak Data Science, what I mean is that 20-3-1 was chosen as the best predictor of usage in the next month. I'd suggest doing some math stuff to get the ratios usong, say, six months of stats.

 

[Mr.E]

RE: all the shit

We didn't start basing tiers off usage until what, late Gen 4 when Smogon was fairly well entranched as the competitive mecca of mons? Usage stats were not even a thing until late Gen 3 on the NetBattle forums. Until that point, tiers were based on the usual competitive conventions of community agreeing upon what is good and what is not. Obviously, this roughly emulates what gets used more often but it isn't an exact science. Just in case anyone cares.

I, for one, have no interest in redoing the tiers based off some flash-in-the-pan usage stats to "modernize" them. They're interesting nonetheless. Ubiquity is certainly not a thing to worry about tier-wise considering there's just so few fully evolved mons in RBY at all compared to newer gens.

Rhydon will always be better than shitty Golem, get at me scrubs.

 

[HereToHelpRBY]

The RBY tiers on here as they stand were made under a different ruleset [wrap illegal] + before a major mechanical shift [and a bunch of relatively minor ones too like counter] and on a separate site so there's the motivation to redo them in some form [or failing that, at least take from a site that made them with the same ruleset and after the mechanical shifts and with a bunch of RBYers who are respected on here too]... whether you / others have the impetus to do that in GSC / ADV is another question but I know Jellicent would like to do that (also we were talking on IRC about it the other day about what's the challenges in doing that on here).

 

[Peasounay]

tiers were based on the usual competitive conventions of community agreeing upon what is good and what is not.

Rhydon will always be better than shitty Golem, get at me scrubs.

That's one of the limits of just basing tieiring on opinions, not everyone has the same.

Ubiquity is certainly not a thing to worry about tier-wise considering there's just so few fully evolved mons in RBY at all compared to newer gens.

What's the issue with playing with NFEs ? I'm sure you won't be surprised that Kadabra and Haunter are for example very strong in UU...

 

[Mr.E]

Of course tiering is subjective. That's why there's a certain community consensus amongst high-level players in any given competitive community that governs them. Usage stats should roughly follow power level but it's not an absolute thing, e.g. almost everybody uses Golem over Rhydon these days (but I will never relent) but traditionally they are both OU because they're all but interchangable; they serve the exact same role in a team and have only very minor differences between them. Or something like Snorlax is very generally strong and can fit on most teams, but it isn't necessarily "better" than something with a niche role that is incredible at doing more specific things but less well-rounded (like, say, Golem/Rhydon).

As if usage stats aren't poisoned by lesser players, weighted or not, as much as the opinions of people differ with regard to traditional tiering (viability), but honestly it doesn't matter anyway because tiering by either method accomplishes different things. Tiering by usage is great because it gives us more freedom to use the "OU but not really OU" stuff we rarely see in normal (OU) competition. That certainly makes lower tiers a hoot! But it's not necessarily the best from a pure competitive standpoint (use anything within the means of the game as given, except agreed-upon totally broken shit, e.g. Mewtwo, Sleep Clause). It doesn't inform us that Persian is a lot better than its usage indicates, because it's just a budget Tauros but "budget best-pokemon-in-OU" can still be pretty good and simply nobody makes room on a team for both these days, or that Rhydon is better than Golem even though all the SPL scrubs who let Tiba win 800 games this past season pretend it doesn't exist. (maybe they'd have taken more games off him if they embraced Rhydon's superiority tbh)

When the hell did we make Articuno OU anyway?

 

[HereToHelpRBY]

As if usage stats aren't poisoned by lesser players, weighted or not, as much as the opinions of people differ with regard to traditional tiering (viability), but honestly it doesn't matter anyway because tiering by either method accomplishes different things. Tiering by usage is great because it gives us more freedom to use the "OU but not really OU" stuff we rarely see in normal (OU) competition.

Really the difference between usage or not is usually one or two Pokémon...

It doesn't inform us that Persian is a lot better than its usage indicates, because it's just a budget Tauros but "budget best-pokemon-in-OU" can still be pretty good and simply nobody makes room on a team for both these days

It's not that good though; even though the meta has favoured it somewhat recently [more use of reflect, faster rather than bulkier take on the meta than was common previously, etc] it's still pretty crap in OU and difficult to use. And even then in BL Kangaskhan is generally the better pick. Here is the RBY 2U viability rankings on Pokemon Perfect for example that reflects that notion; furthermore in this meta Rhydon is allowed which is only an advantage for persian and is worse for dodrio and arguably worse for kangaskhan. That being said most variations of BL they allow Articuno and/or Moltres; I'd note that the former is banned on rby2k10 where these tiers originated and that moltres probably isn't broken without Fire Spin [i.e. using rby2k10 rules], and also that Kangaskhan's access to Rock Slide may well be yet another advantage it has over Persian if those are permitted.

or that Rhydon is better than Golem even though all the SPL scrubs who let Tiba win 800 games this past season pretend it doesn't exist. (maybe they'd have taken more games off him if they embraced Rhydon's superiority tbh)

You overrate SPL though, quite a few good players couldn't play or didn't play every week [marco tourbanned till midseason, ggfan banned ofc, disaster area banned, alexander on a team with Lutra (Alexander's record on Pokemon Perfect is insane, he won 4 out of the last 5 tournaments and reached finals of the other, and those are all that he has played) so only one was playing RBY each week, and a couple of other players in PP's top 10 didn't get picked up either.] and some relatively mediocre players like [EDITED OUT]

even if u disagree with my assessment (although I did use a list that accounts for over 20 tournaments from the past few years and I'm not sure what else I'm expected to use) you can't disagree that it's subjective as to whether the best 10 players available are the ones who are playing, and you had some picks who have much less experience. And whilst you can argue players care more about the tour and so the games are better quality 1 there's a strong RBY community outside of Smogon so the passion is bound to be less for the RBY players [actually I attribute to the reduced passion for Smogon because of other communities why so much ghosting is by Italians - the italian community is a pretty active one as far as I gather and has a long history] 2 how much is there to gain from prep for your opponent in RBY? There's much less teambuilding burden and the most build-wise you should really be doing in this is smaller tweaks most of the time.

This is no disrepsect to Tiba or any players in particular (though I did mention names :x) just that the whole tournament set-up isn't really good enough to judge the utility of one niche Pokémon fairly. There's a small sample size on top of everything else I said, i.e. that the players are of varying quality and tournament passion. Add in that Tiba might have greater experience / talent with Rhydon than with other niche Pokémon and as such it might appear stronger than it is even with everything else notwithstanding.