Metagame SV Monotype Metagame Discussion [Indigo Disk]

[Eviolite Bisharp]

Daw whats this discussion, rock is the worst type ever thats a monotype law in every gen The end

On another matter I wanted to propose something, this gonna sound like a sort of a joke but hear me out ok? I propose to unban annihilape and ban rage fist instead.

Props to unban Annihilape/ ban Rage Fist:

• Annihilape is balanced without the move
• Monotype win a valid and capable mon to use in the tier
• 2 mid tier types enjoy the unban making them more viable in the meta

But now the obvious con: Primeape, a Rage Fist user as well, isnt broken with the move.

That could be the end of the discussion right there, except there is a precedent: when the move Last Respects was banned in the past, instead of banning the users, the move was banned even if there was a mon, Basculin, that wasnt broken with the move (if you tell me Last Respects Basculin is broken in monotype i'm gonna slap you).
So at the time was decided to ban the move and keep the mons, and it proved to be a good decision, no one complains and it let us use Basculegion in the tier. I think the same can be done with Annihilape. What do you think?

well, yes, but the reason why Last Respects was banned was precisely the reason why Rage Fist isnt banned: both Houndstone and Basculegion were broken with that absurd move, and when Smogon bans a move is because its clear that the move itself it's broken and not the move plus the inherent traits of the pokemon banned that gets it, and it's the same reason why Shed Tail is banned instead of Cyclizar, Orthworm and Sceptile.

If Eviolite Primeape was as same as broken and banworthy as Annihilape with Rage Fist then the movement would be banned, but that's obviously not the case

Also i won't accept any bad comments against the GOATst typing

 

[Giyu]

Daw whats this discussion, rock is the worst type ever thats a monotype law in every gen The end

On another matter I wanted to propose something, this gonna sound like a sort of a joke but hear me out ok? I propose to unban annihilape and ban rage fist instead.

Props to unban Annihilape/ ban Rage Fist:

• Annihilape is balanced without the move
• Monotype win a valid and capable mon to use in the tier
• 2 mid tier types enjoy the unban making them more viable in the meta

But now the obvious con: Primeape, a Rage Fist user as well, isnt broken with the move.

That could be the end of the discussion right there, except there is a precedent: when the move Last Respects was banned in the past, instead of banning the users, the move was banned even if there was a mon, Basculin, that wasnt broken with the move (if you tell me Last Respects Basculin is broken in monotype i'm gonna slap you).
So at the time was decided to ban the move and keep the mons, and it proved to be a good decision, no one complains and it let us use Basculegion in the tier. I think the same can be done with Annihilape. What do you think?

While Basculin wasn't considered unhealthy itself despite learning Last Respects, you still have more than one Pokemon (Basculegion and Houndstone) that were both clearly uncompetitive due to having access to the move. As it stands now, there are only two users of Rage Fist with only one showing signs of being unhealthy (Annihilape) with Primeape being pretty bad. I suppose the precedent here is at least two Pokemon that are deemed unhealthy due to a common trait such as an unbalanced move, can be made legal with the said common trait being banned instead.

Also you don't really have imperical evidence that Annihilape would be balanced without Rage Fist but that is up for discussion.

Finally, unbanning a Pokemon to make a certain type better is against tiering policy. Baxcalibur, Iron Bundle and Chien-Pao would make Ice really good again but they are inherently unhealthy for the metagame.

 

[Jahkem]

Props to unban Annihilape/ ban Rage Fist:

• Annihilape is balanced without the move
• Monotype win a valid and capable mon to use in the tier
• 2 mid tier types enjoy the unban making them more viable in the meta

But now the obvious con: Primeape, a Rage Fist user as well, isnt broken with the move.

The only viable user of Rage Fist is Annihilape. How Primeape does with Rage Fist doesn’t matter since it’s unviable to begin with. If more viable mons are given Rage Fist, then the move will have grounds for a ban.

That could be the end of the discussion right there, except there is a precedent: when the move Last Respects was banned in the past, instead of banning the users, the move was banned even if there was a mon, Basculin, that wasnt broken with the move (if you tell me Last Respects Basculin is broken in monotype i'm gonna slap you).
So at the time was decided to ban the move and keep the mons, and it proved to be a good decision, no one complains and it let us use Basculegion in the tier. I think the same can be done with Annihilape. What do you think?

Houndstone was banned at the beginning of SV because of Last Respects. Even though it was confirmed Basculegion got Last Respects before Pokemon Home was released, the council tiers the format in its current state, not for what is coming in the future.

 

[pikachu<3]

The only viable user of Rage Fist is Annihilape. How Primeape does with Rage Fist doesn’t matter since it’s unviable to begin with. If more viable mons are given Rage Fist, then the move will have grounds for a ban.


Houndstone was banned at the beginning of SV because of Last Respects. Even though it was confirmed Basculegion got Last Respects before Pokemon Home was released, the council tiers the format in its current state, not for what is coming in the future.

In fairness to Day Healer, the council has chosen to ban something over one pokemon in the past. For example, Excadrill is the only broken Sand Rush abuser with smooth rock, but the council chose to ban Smooth Rock over banning Excadrill. I'm not going to state whether banning Rage Fist or Annihilape is the better option, just that the precedent of banning something over only one broken user is already there.

 

[Jahkem]

In fairness to Day Healer, the council has chosen to ban something over one pokemon in the past. For example, Excadrill is the only broken Sand Rush abuser with smooth rock, but the council chose to ban Smooth Rock over banning Excadrill. I'm not going to state whether banning Rage Fist or Annihilape is the better option, just that the precedent of banning something over only one broken user is already there.

That’s a good pull. Smooth Rock ban worked out because Excadrill with 5 turns of Sand is still fine. If it was any stronger, it’d be banned and Smooth Rock would be allowed. The council bans whatever is the strongest element. In this case, Annihilape being the only viable mon with Rage Fist, both are seen as a singular target and that results in the ban. Compared to Sand where it involves Hippowdon, Sand Stream, Sand Rush, Excadrill, and Smooth Rock. You look at those 5 elements and pick Smooth Rock to be the strongest to be removed.

 

[biggruse04]

so smooth rock and damp rock are banned but heat rock, Icey rock, terrain extender and light clay are not banned even though icey rock has the same as excadrill in alolan sandslash i can make same claim with heat rock and it's mons that use the sun and the same with terrain with psychic terrain being the main terrain, screens as well 8 turn long field effects are not healthy for the meta

 

[cpt.kraken]

so smooth rock and damp rock are banned but heat rock, Icey rock, terrain extender and light clay are not banned even though icey rock has the same as excadrill in alolan sandslash i can make same claim with heat rock and it's mons that use the sun and the same with terrain with psychic terrain being the main terrain, screens as well 8 turn long field effects are not healthy for the meta

Alolan sandslash is slow even max speed jolly being outspeeded by most common scarf users or doesnt hit hard enough. Issue would be more around Cetitan but ice as a whole isn't perceived good enough because steel and fire hard stop it. Most top tier-types can just play around. Also, it’s far worst than Excadrill under sand or Barra/Kingdra/BascuF under rain that can abuse it way more.
We had a heat rock suspect test and it has been voted to remain. Only Scovillain has access to Chlorophyl and it's not crazy.

Terrains as a whole are pretty limited to psychic, normal, grass and elec types. Normal doesn't need it and psychic, grass and elec types arent considered good enough even with terrain.

We could talk about light clay but maybe abuse a bit more about knock off. I dont think it's a real issue in the meta rn. Like you can take advantage of Kelfki main screens setters cause very passive and cant run twave + foul play + spikes. Ninetales-a can be a good setter and threatens water a lot but is then very weak to hazards and Avalugg is pretty easy to play around. Grimmsnarl for dark is fun but not as reliable than a regular dark bulky offense core. Sableye is just a free setup fodder for dark mons. Generally those mons are too passive and have to come on field several times, taking hazards, hits, giving a free setup to oppo

 

[Jahkem]

so smooth rock and damp rock are banned but heat rock, Icey rock, terrain extender and light clay are not banned even though icey rock has the same as excadrill in alolan sandslash i can make same claim with heat rock and it's mons that use the sun and the same with terrain with psychic terrain being the main terrain, screens as well 8 turn long field effects are not healthy for the meta

That's because the main abuser of Sand is Excadrill and for Rain it's Barraskewda and Kingdra. Excadrill doesn't always have to run Choice Band with 8 turns of Sand, it can run SD sets and be even more threatening that it already is. There's only a select few Choice Scarf users that can outspeed Jolly Excadrill, those being Flutter Mane and Dragapult which they don't OHKO Excadrill and get OHKO'd in return. 8 turns of Rain is an even bigger problem not just because of Barraskewda and Kingdra, but also with the bonus of having their STAB boosted by Rain. The whole type becomes an issue when they get Choice Specs on their STAB for 8 turns. Things like Urshifu, Walking Wake, Manaphy, Primarina, etc. become more dangerous than they already are. The arguments for why these particular weather rocks are removed are valid enough to keep them out of the metagame.

Looking at the other remaining items you mentioned starting with Icy Rock: The main Snow abuser is Cetitan. Belly Drum + Ice/Ground coverage allows it to do a great amount of work. The other viable Slush Rush user being Alolan Sandslash is usually a utility mon for hazard removal, item removal in Knock Off, and the occasional threat of Iron Head on a Fairy. It's not as threatening, but it has it's uses. However, even with 8 turns of Snow with a Defense boost, it doesn't push Ice to be a viably consistent type to use. Even when Baxcalibur was allowed, 8 turns of Snow wasn't the main option support Baxcalibur, it was Light Clay. Which I'll segway into that.

Screens as a whole isn't nearly as strong in Monotype compared to the usage based tiers below OU. Only one viable type consistently runs Light Clay Screens and that's Fairy which is currently on fraud watch. Light Clay helps Fairy have a reliable defensive backbone for 8 turns but that's about it. Having to reapply to backbone multiple times in a game can put you in some tough spots on a type that wants to end games within 25 turns. Dark could runs Light Clay Screens as well but it'd be subpar to the current offensive iterations that are ran.

And finally with Terrain Extender, the main reason for it being banned in the first place isn't around currently. Tapu Koko + Alolan Raichu pushes Electric to an unhealthy level in just about any metagame for our tier. The closest indication being NatDex Mono having it banned solely because of the combination of Tapu Koko + Alolan Raichu. The only abuser of Terrain Extender in the current meta is Indeedee-M and Psychic is one of the most subpar types this generation.

Not every argument for one thing is linear to everything.

 

[TheKart017]

Hi, I have been silent over the past few months because initially I coded a matchup chart to be integrated into the usage stats of pokemon showdown, that way we can see the actual win rates of matchups and actually create a type chart that is based on data instead of vibes. This was back in september 2025. Since I have not had any message from marty because she is very busy, I had the idea of taking sample teams and performing a simulation using bots. However, because of school and the fact that merging poke-env(simulation API) and poke-engine(AI bot API) is very annoying, I'm still doing that. If you are still interested in the results(because it can cover old gen MU charts) please like this post so I know who is actually willing to see the results. I want to thank the works of people like ethereal sword who have tried to give us better insight into the meta through data during the long wait. Today, I was contacted by marty who was finally able to test the code and give some preliminary results. The code has already been integrated into usage stats and the charts will be in monotype/matchupcharts/ and will include 8 charts, 4 for gen 9, and 4 for nat dex, capped by glicko. The format will be a bit different as it will have the raw count of games won as well. She was also kind enough to run the premilinary results for the first 13 days of march, so I will base the below analysis of gen 9 on said data. I will use 1630 glicko as that is a good benchmark for intemediate to good players. If you hate numbers go to the end of this post.

NOTE: this chart only considers 13 days worth of data, so it may not necessarily represent the meta in general. Also this is ladder, so please take that in mind.

MU Chart

The way to read this is that the percent is the win rate of the left side vs the top. Based on this chart, we can see the worst MUs in the game, where the matchup is so skewed that the type wins more than 80% of the time:
dragon vs electric: 90.026
This is very evident as without tapu koko electric is kinda screwed vs dragon

fighting vs rock: 87.663
You have no way to switch into close combat spam well, i don't think more needs to be said

ice vs flying: 87.484
ice vs ground: 87.241
ice is just good against types that don't switch into ice moves well. Although ground has sand, snow can compete and can even got buffed with the defense bonus

rock vs fire: 84.338
fire already sucks against scarf kleavor stone axe so you can imagine what a whole rock team does.

psychic vs electric: 84.321
You have no psychic switch in as zone can't save you and morpeko sucks. Also, competing terrain

ground vs rock: 82.644
No way to switch into moves like earthquake and earth power without losing a lot.

electric vs bug: 82.469
You have no good electric switchin so thunderclap and as well tbolt spam is wrecking your entire team.

ground vs poison: 80.692
if gliscor appears, you're screwed, if not, gweezing is not saving you and most ladder players prob are not running amoonguss

steel vs ice: 80.246
if scizor appears, gg, if not, you have to deal with heatran + dengo

Using this chart, we can also derive 3 scores: the average win rate, usage-weighted win scores, and winrate-weighted win scores. All three are mildly valid to create a tier list and I will show a tier list for all 3. First, based on average win rates

We observe a lot of interesting things. First, steel is best if we consider every type equally, which makes sense. Similar story water and dragon. On the otherside, we see types like electric and grass which most consider to be terrible types. There are 2 surprising facts though. First Ice, Rock, and Fighting are significantly higher than usual. Ice probably that high because they are really good at decimating certain types like flying and lower types like grass which may artificially increase its standings. It's a similar story with rock destroying types like bug and fire as well as rock and somehow bug for fighting. Second surprise is just how low flying is. My best interpretation is that flying is very difficult to pilot so people on ladder usually don't know how to use it.

As you can see, the chart is heavily skewed by the fact that some types have overwhelming win against others. As a result I prefer to use the weighted score option. The weighted score of type_1= Σ [usage(type_2) × winRate(type_1 vs type_2) ]. Since the weight is dependent on how much the opposing types is used, which is usually an ok indicator for how good a type is, winning a lot against weaker types is counted less, with that we get the below results.

Here we see the weakness of using usage as an indicator for viability as for example, steel is ranked very low because a lot of people really don't like steel and will often overprepare for it or use type like fire, ground, and fighting to snipe it. On the contrast, we see that ghost and bug are doing much better than normal while flying and ground are doing worse than normal. This is the simple result of using usage because ladder is filled with types like dark and ghost that sth like lokix just shreds through and the abundance of water on ladder really doesn't help types like ground, steel, and fire.

This bring us to lastly, the winrate-weighted win scores which I believe to be the best metric for measuring how good a type is. The formula is Σ [average_winrate(type_2) × winRate(type_1 vs type_2)]. This uses overall winrate to measure viability which I believe to be the best metric and the result shows.

Personally I feel that this is a much better analysis of the overall meta in general. For example, one of things I have always believed is that normal is an amazing type. Khahara has always stressed as much and I believe this tier list shows. Additionally, I have always been wary of calling fighting a snipe type as it does very well against great types like steel dark and normal. The huge surprise is the placement of rock and ice. I think the main thing is that they have consistently above average win rates against most types and even really good ones like against flying, but I think the problem is that a lot of players are a bit too hooked on the fact that against the types they lose against, they lose HARD. Lastly, I know all the flying players are going to scream at me for this, but I'm very much aware, most good flying players don't ladder with flying much because they much prefer to use it for tournaments. This leaves the remaining flying players to perform much worse on ladder. Flying is a very difficult type to use so consider this as proof that you flying players are good at this game. Lastly, I do want to note a lot of people have been under-rating water. I've always thought of it as one of the best types potentially top 3 and these tier lists prove my point.

TLDR: Lastly, I know a lot of hate numbers so here is the chart but non-math people friendly. -3 is borderline unwinnable, 3 is you should not lose.

 

[Wincon]

Hi, I have been silent over the past few months because initially I coded a matchup chart to be integrated into the usage stats of pokemon showdown, that way we can see the actual win rates of matchups and actually create a type chart that is based on data instead of vibes. This was back in september 2025. Since I have not had any message from marty because she is very busy, I had the idea of taking sample teams and performing a simulation using bots. However, because of school and the fact that merging poke-env(simulation API) and poke-engine(AI bot API) is very annoying, I'm still doing that. If you are still interested in the results(because it can cover old gen MU charts) please like this post so I know who is actually willing to see the results. I want to thank the works of people like ethereal sword who have tried to give us better insight into the meta through data during the long wait. Today, I was contacted by marty who was finally able to test the code and give some preliminary results. The code has already been integrated into usage stats and the charts will be in monotype/matchupcharts/ and will include 8 charts, 4 for gen 9, and 4 for nat dex, capped by glicko. The format will be a bit different as it will have the raw count of games won as well. She was also kind enough to run the premilinary results for the first 13 days of march, so I will base the below analysis of gen 9 on said data. I will use 1630 glicko as that is a good benchmark for intemediate to good players. If you hate numbers go to the end of this post.

NOTE: this chart only considers 13 days worth of data, so it may not necessarily represent the meta in general. Also this is ladder, so please take that in mind.

MU Chart
View attachment 816308
The way to read this is that the percent is the win rate of the left side vs the top. Based on this chart, we can see the worst MUs in the game, where the matchup is so skewed that the type wins more than 80% of the time:
dragon vs electric: 90.026
This is very evident as without tapu koko electric is kinda screwed vs dragon

fighting vs rock: 87.663
You have no way to switch into close combat spam well, i don't think more needs to be said

ice vs flying: 87.484
ice vs ground: 87.241
ice is just good against types that don't switch into ice moves well. Although ground has sand, snow can compete and can even got buffed with the defense bonus

rock vs fire: 84.338
fire already sucks against scarf kleavor stone axe so you can imagine what a whole rock team does.

psychic vs electric: 84.321
You have no psychic switch in as zone can't save you and morpeko sucks. Also, competing terrain

ground vs rock: 82.644
No way to switch into moves like earthquake and earth power without losing a lot.

electric vs bug: 82.469
You have no good electric switchin so thunderclap and as well tbolt spam is wrecking your entire team.

ground vs poison: 80.692
if gliscor appears, you're screwed, if not, gweezing is not saving you and most ladder players prob are not running amoonguss

steel vs ice: 80.246
if scizor appears, gg, if not, you have to deal with heatran + dengo

Using this chart, we can also derive 3 scores: the average win rate, usage-weighted win scores, and winrate-weighted win scores. All three are mildly valid to create a tier list and I will show a tier list for all 3. First, based on average win rates
View attachment 816309View attachment 816331
We observe a lot of interesting things. First, steel is best if we consider every type equally, which makes sense. Similar story water and dragon. On the otherside, we see types like electric and grass which most consider to be terrible types. There are 2 surprising facts though. First Ice, Rock, and Fighting are significantly higher than usual. Ice probably that high because they are really good at decimating certain types like flying and lower types like grass which may artificially increase its standings. It's a similar story with rock destroying types like bug and fire as well as rock and somehow bug for fighting. Second surprise is just how low flying is. My best interpretation is that flying is very difficult to pilot so people on ladder usually don't know how to use it.

As you can see, the chart is heavily skewed by the fact that some types have overwhelming win against others. As a result I prefer to use the weighted score option. The weighted score of type_1= Σ [usage(type_2) × winRate(type_1 vs type_2) ]. Since the weight is dependent on how much the opposing types is used, which is usually an ok indicator for how good a type is, winning a lot against weaker types is counted less, with that we get the below results.
View attachment 816328View attachment 816330

Here we see the weakness of using usage as an indicator for viability as for example, steel is ranked very low because a lot of people really don't like steel and will often overprepare for it or use type like fire, ground, and fighting to snipe it. On the contrast, we see that ghost and bug are doing much better than normal while flying and ground are doing worse than normal. This is the simple result of using usage because ladder is filled with types like dark and ghost that sth like lokix just shreds through and the abundance of water on ladder really doesn't help types like ground, steel, and fire.

This bring us to lastly, the winrate-weighted win scores which I believe to be the best metric for measuring how good a type is. The formula is Σ [average_winrate(type_2) × winRate(type_1 vs type_2)]. This uses overall winrate to measure viability which I believe to be the best metric and the result shows.
View attachment 816338View attachment 816340
Personally I feel that this is a much better analysis of the overall meta in general. For example, one of things I have always believed is that normal is an amazing type. Khahara has always stressed as much and I believe this tier list shows. Additionally, I have always been wary of calling fighting a snipe type as it does very well against great types like steel dark and normal. The huge surprise is the placement of rock and ice. I think the main thing is that they have consistently above average win rates against most types and even really good ones like against flying, but I think the problem is that a lot of players are a bit too hooked on the fact that against the types they lose against, they lose HARD. Lastly, I know all the flying players are going to scream at me for this, but I'm very much aware, most good flying players don't ladder with flying much because they much prefer to use it for tournaments. This leaves the remaining flying players to perform much worse on ladder. Flying is a very difficult type to use so consider this as proof that you flying players are good at this game. Lastly, I do want to note a lot of people have been under-rating water. I've always thought of it as one of the best types potentially top 3 and these tier lists prove my point.

TLDR: Lastly, I know a lot of hate numbers so here is the chart but non-math people friendly. -3 is borderline unwinnable, 3 is you should not lose.
View attachment 816345

I will not stand for all of this electric slander.

 

[Eviolite Bisharp]

Hi, I have been silent over the past few months because initially I coded a matchup chart to be integrated into the usage stats of pokemon showdown, that way we can see the actual win rates of matchups and actually create a type chart that is based on data instead of vibes. This was back in september 2025. Since I have not had any message from marty because she is very busy, I had the idea of taking sample teams and performing a simulation using bots. However, because of school and the fact that merging poke-env(simulation API) and poke-engine(AI bot API) is very annoying, I'm still doing that. If you are still interested in the results(because it can cover old gen MU charts) please like this post so I know who is actually willing to see the results. I want to thank the works of people like ethereal sword who have tried to give us better insight into the meta through data during the long wait. Today, I was contacted by marty who was finally able to test the code and give some preliminary results. The code has already been integrated into usage stats and the charts will be in monotype/matchupcharts/ and will include 8 charts, 4 for gen 9, and 4 for nat dex, capped by glicko. The format will be a bit different as it will have the raw count of games won as well. She was also kind enough to run the premilinary results for the first 13 days of march, so I will base the below analysis of gen 9 on said data. I will use 1630 glicko as that is a good benchmark for intemediate to good players. If you hate numbers go to the end of this post.

NOTE: this chart only considers 13 days worth of data, so it may not necessarily represent the meta in general. Also this is ladder, so please take that in mind.

MU Chart
View attachment 816308
The way to read this is that the percent is the win rate of the left side vs the top. Based on this chart, we can see the worst MUs in the game, where the matchup is so skewed that the type wins more than 80% of the time:
dragon vs electric: 90.026
This is very evident as without tapu koko electric is kinda screwed vs dragon

fighting vs rock: 87.663
You have no way to switch into close combat spam well, i don't think more needs to be said

ice vs flying: 87.484
ice vs ground: 87.241
ice is just good against types that don't switch into ice moves well. Although ground has sand, snow can compete and can even got buffed with the defense bonus

rock vs fire: 84.338
fire already sucks against scarf kleavor stone axe so you can imagine what a whole rock team does.

psychic vs electric: 84.321
You have no psychic switch in as zone can't save you and morpeko sucks. Also, competing terrain

ground vs rock: 82.644
No way to switch into moves like earthquake and earth power without losing a lot.

electric vs bug: 82.469
You have no good electric switchin so thunderclap and as well tbolt spam is wrecking your entire team.

ground vs poison: 80.692
if gliscor appears, you're screwed, if not, gweezing is not saving you and most ladder players prob are not running amoonguss

steel vs ice: 80.246
if scizor appears, gg, if not, you have to deal with heatran + dengo

Using this chart, we can also derive 3 scores: the average win rate, usage-weighted win scores, and winrate-weighted win scores. All three are mildly valid to create a tier list and I will show a tier list for all 3. First, based on average win rates
View attachment 816309View attachment 816331
We observe a lot of interesting things. First, steel is best if we consider every type equally, which makes sense. Similar story water and dragon. On the otherside, we see types like electric and grass which most consider to be terrible types. There are 2 surprising facts though. First Ice, Rock, and Fighting are significantly higher than usual. Ice probably that high because they are really good at decimating certain types like flying and lower types like grass which may artificially increase its standings. It's a similar story with rock destroying types like bug and fire as well as rock and somehow bug for fighting. Second surprise is just how low flying is. My best interpretation is that flying is very difficult to pilot so people on ladder usually don't know how to use it.

As you can see, the chart is heavily skewed by the fact that some types have overwhelming win against others. As a result I prefer to use the weighted score option. The weighted score of type_1= Σ [usage(type_2) × winRate(type_1 vs type_2) ]. Since the weight is dependent on how much the opposing types is used, which is usually an ok indicator for how good a type is, winning a lot against weaker types is counted less, with that we get the below results.
View attachment 816328View attachment 816330

Here we see the weakness of using usage as an indicator for viability as for example, steel is ranked very low because a lot of people really don't like steel and will often overprepare for it or use type like fire, ground, and fighting to snipe it. On the contrast, we see that ghost and bug are doing much better than normal while flying and ground are doing worse than normal. This is the simple result of using usage because ladder is filled with types like dark and ghost that sth like lokix just shreds through and the abundance of water on ladder really doesn't help types like ground, steel, and fire.

This bring us to lastly, the winrate-weighted win scores which I believe to be the best metric for measuring how good a type is. The formula is Σ [average_winrate(type_2) × winRate(type_1 vs type_2)]. This uses overall winrate to measure viability which I believe to be the best metric and the result shows.
View attachment 816338View attachment 816340
Personally I feel that this is a much better analysis of the overall meta in general. For example, one of things I have always believed is that normal is an amazing type. Khahara has always stressed as much and I believe this tier list shows. Additionally, I have always been wary of calling fighting a snipe type as it does very well against great types like steel dark and normal. The huge surprise is the placement of rock and ice. I think the main thing is that they have consistently above average win rates against most types and even really good ones like against flying, but I think the problem is that a lot of players are a bit too hooked on the fact that against the types they lose against, they lose HARD. Lastly, I know all the flying players are going to scream at me for this, but I'm very much aware, most good flying players don't ladder with flying much because they much prefer to use it for tournaments. This leaves the remaining flying players to perform much worse on ladder. Flying is a very difficult type to use so consider this as proof that you flying players are good at this game. Lastly, I do want to note a lot of people have been under-rating water. I've always thought of it as one of the best types potentially top 3 and these tier lists prove my point.

TLDR: Lastly, I know a lot of hate numbers so here is the chart but non-math people friendly. -3 is borderline unwinnable, 3 is you should not lose.
View attachment 816345

Wow, rock is much higher than i expected lol

 

[TheKart017]

I will not stand for all of this electric slander.

you can blame most electric players being terrible

 

[Sugarbear]

The problem with Electric is that it wants to be an all out hyper offensive terrain abusing type, but it doesn’t have the tools. The loss of Tapu Koko was devastating (Pincurchin is lol), as it lost a Pokemon that both set terrain and was a fast cleaner in its own right. No more Zeraora was another huge blow. Oh and Rising Voltage’s distribution being devastated (only Raging Bolt). It doesn’t have the bulk and resistances to pull off balance and bulky offensive. Oh, and as hyper offensive types go, Fire, Water, Ice, Psychic and Fairy all do that better.

It’s still better than Grass though.

 

[TheKart017]

The problem with Electric is that it wants to be an all out hyper offensive terrain abusing type, but it doesn’t have the tools. The loss of Tapu Koko was devastating (Pincurchin is lol), as it lost a Pokemon that both set terrain and was a fast cleaner in its own right. No more Zeraora was another huge blow. Oh and Rising Voltage’s distribution being devastated (only Raging Bolt). It doesn’t have the bulk and resistances to pull off balance and bulky offensive. Oh, and as hyper offensive types go, Fire, Water, Ice, Psychic and Fairy all do that better.

It’s still better than Grass though.

the chart disagrees with you. I also disagree with you. For example, you saw the data, by every metric possible, grass is better than electric and it makes sense. For example, ogerpon-cornerstone is not very fun for types like flying and fire to deal with as they have very little switch to ivy cudgel. The same can be said of ogerpon hearthflame against types like steel. In addition, meow is a fantastic option for speed control and rilaboom can help a lot when you don't have the speed advantage. Lastly, grass has a terrain while elec does not. Rillaboom is an actual pokemon while pincurchin very much so is not. Now it's not better by much as grass as a type is just worse than electric but elec just doesn't have the tools to function.

 

[Giyu]

the chart disagrees with you. I also disagree with you. For example, you saw the data, by every metric possible, grass is better than electric and it makes sense. For example, ogerpon-cornerstone is not very fun for types like flying and fire to deal with as they have very little switch to ivy cudgel. The same can be said of ogerpon hearthflame against types like steel. In addition, meow is a fantastic option for speed control and rilaboom can help a lot when you don't have the speed advantage. Lastly, grass has a terrain while elec does not. Rillaboom is an actual pokemon while pincurchin very much so is not. Now it's not better by much as grass as a type is just worse than electric but elec just doesn't have the tools to function.

Using raw ladder statistics like this makes for a pretty shallow argument. As you mentioned this is 13 days of data, so the sample size is next to nothing. Player skill distribution, type popularity, and even certain team compositions are all factors here, you can't just base type viability off dry numbers. Electric is objectively better than Grass, it has significantly better tools to deal with meta staples, substantially more winnable matchups, and overall a more balanced type. Even with Ogerpon-C/W, the Fire matchup is next to unwinnable if your opponent is semi-competent. Grass even lost to Water recently, in a competitive setting mind you, not ladder. Grass has quite literally nothing neutral to Ice except Abomasnow, nothing neutral to Fire other than Ogerpon formes, and then you have Heatproof Sinistcha... So yeah Grass has absolutely no tools defensively compared to Electric.

 

[TheKart017]

Using raw ladder statistics like this makes for a pretty shallow argument. As you mentioned this is 13 days of data, so the sample size is next to nothing. Player skill distribution, type popularity, and even certain team compositions are all factors here, you can't just base type viability off dry numbers. Electric is objectively better than Grass, it has significantly better tools to deal with meta staples, substantially more winnable matchups, and overall a more balanced type. Even with Ogerpon-C/W, the Fire matchup is next to unwinnable if your opponent is semi-competent. Grass even lost to Water recently, in a competitive setting mind you, not ladder. Grass has quite literally nothing neutral to Ice except Abomasnow, nothing neutral to Fire other than Ogerpon formes, and then you have Heatproof Sinistcha... So yeah Grass has absolutely no tools defensively compared to Electric.

You know, my rust is starting to show. I can see your point. To be honest, at this point, I kinda wanna spam flying on ladder just so I can 1. shake off this rust, 2. raise flying's standings by the end of march.

 

[pikachu<3]

the chart disagrees with you. I also disagree with you. For example, you saw the data, by every metric possible, grass is better than electric and it makes sense. For example, ogerpon-cornerstone is not very fun for types like flying and fire to deal with as they have very little switch to ivy cudgel. The same can be said of ogerpon hearthflame against types like steel. In addition, meow is a fantastic option for speed control and rilaboom can help a lot when you don't have the speed advantage. Lastly, grass has a terrain while elec does not. Rillaboom is an actual pokemon while pincurchin very much so is not. Now it's not better by much as grass as a type is just worse than electric but elec just doesn't have the tools to function.

I'd personally advise people not to look into what the chart says too closely, especially since the matchups on the chart don't always follow the eye test. The chart's a nice experiment, imo, to see what raw ladder data tells us, but I wouldn't rely on the chart as validation for *anything* at the moment until the data passes the eye test at least.

To understand what I'm talking about when I say "the eye test", just looking at the ice matchups for example:

Ice vs Fire: The chart says this is -1 in favor of Fire. Anyone who has played this matchup as Ice will tell you this is actually unwinnable unless you get absurdly lucky (assuming a good opponent). You're consistently getting outsped (Fire's mons are faster than yours) and getting OHKOed by super effective STAB; you can't set up Veil easily due to being contested on weather; your Avalugg is useless because Ceruledge exists, and to add insult to injury, even when you do get Veil or Rocks up it doesn't mean shit because Cinderace has a balanced move called Court Change. If you do run Cloyster, that doesn't OHKO heatran at +2 and gets OHKOed in return by Magma Storm. Oh, if that isn't enough, Will-O-Wisp Volcarona 6-0s from preview. If this isn't considered a -3, no matchup should be -3 lmao.

Ice vs Fighting: The chart says this is an even MU; this is not even by any stretch of the imagination. A lot of the same issues exist here as with fire; you're consistently getting outsped and OHKOed by super effective STAB. Veil's nice here but Urshifu ignores it and just clicks Surging; while this isn't unplayable for Ice it's definitely fighting favored.

Ice vs Bug: Scizor, Volcarona, and Kleavor exist. This is one of Ice's worst mus and is basically an autoloss at preview. How is this not a -3?

Ice vs Water: This is somehow 0 on the chart, but among high level play this is a super good MU for ice due to Freeze-dry being broken. Empoleon builds smash but that mon isn't common on water. How is this considered even?

Ice vs Rock: The chart says this is -1 which suggests only unfavored; but this is an atrocious matchup for Ice for various reasons; I would go as far as to say it's borderline unwinnable (barring stone miss) against someone who knows the matchup. Veil can't be setup safely due to TTar's sand stream, and rock has too many mons that are just too fast and too strong (Iron Boulder, Ogerpon-C), and Ice's main nuke, Specs Kyurem, has problems getting anything done here due to the Special Defense boost of Sand. Avalugg's good but has problems clearing hazards consistently due to Glimmora coming in on it (who, incidentally, punishes Body Press with Toxic Debris); ideally Avalugg switches into the Rock setter and spins but that's unfornately TTar who also has Knock Off, and a Knock'd Avalugg is really bad because Glimmora can force it to take status from Toxic Spikes.

Ice vs Fairy: The chart claims this is even but it's atrocious for Ice; Avalugg is useless here due to Flutter existing, Iron Valiant has no switchins and outspeeds the entire Ice team, and Hatterene straight up 6-0s behind screens.

Ice vs Ghost: The chart claims this is even again, but this matchup is awful for ice; Avalugg's useless and Gholdengo + Ceruledge + Flutter is curtains. To add insult to injury, Specs Pult doesn't care about Aurora Veil and 2HKOs everything with shadow ball.

Ice vs Normal: The chart claims this is even, but, again, it's awful for Ice; Avalugg's useless due to Zoro existing and Normal's content to play the hazard game with SR (which Ice can't prevent from coming up; blissey too tanky). Setup sweepers, which is Ice's perferred win condition, are countered by the omnipresent Ditto.

Overall these are just the more blatant things I've noticed on the Ice MU chart; there are other mus that are wrong that I didn't point out (for example, among other things, Ice vs Flying shouldn't be +3), I'm sure the other types also have similar issues on their matchup chart.

 

[ImprobableBlob]

The main issue is of course sample size, especially when it comes to the lower usage types like ice and electric. As an example, if we assume 10000 games played, ice vs electric should have been played about 3 times, whereas dragon vs dark should get 96. Any of the types that get under 3% usage are going to be a bit wonky, but the higher usage types should be giving usable data.

 

[aaaaahowdogamework]

I will not stand for all of this electric slander.

well then get on ladder and boost those numbers

 

[aaaaahowdogamework]

I think the main issue causing the chart to differ from people's experience of what is and isn't a good matchup isn't sample size, but selection effects. For two people to be paired with each other they have to be similar elo, and a 1500 elo fire player is probably a lot less skilled than a 1500 elo ice player, as it's harder to get to that elo with ice. This is making the matchups for bad types look a lot less bad than they actually are, because with skilled pilots a lot of bad matchups can be overcome, and because fundamentally with the way the chart is constructed average winrate can't stray too far from 50% (or where it does it's from people like me and cee spamming the type at high ladder). I think this data is useful for e.g. comparing water's ghost matchup to water's dark matchup, but not for e.g. comparing water's ghost matchup to grass's dark matchup.

I wish there was a good way to control for a player's skill (i.e. what their elo/glicko would be if they weren't playing that type) but I don't think there is one, given how many people mostly play one type anyway.

 

[TheKart017]

I'd personally advise people not to look into what the chart says too closely, especially since the matchups on the chart don't always follow the eye test. The chart's a nice experiment, imo, to see what raw ladder data tells us, but I wouldn't rely on the chart as validation for *anything* at the moment until the data passes the eye test at least.

To understand what I'm talking about when I say "the eye test", just looking at the ice matchups for example:

Ice vs Fire: The chart says this is -1 in favor of Fire. Anyone who has played this matchup as Ice will tell you this is actually unwinnable unless you get absurdly lucky (assuming a good opponent). You're consistently getting outsped (Fire's mons are faster than yours) and getting OHKOed by super effective STAB; you can't set up Veil easily due to being contested on weather; your Avalugg is useless because Ceruledge exists, and to add insult to injury, even when you do get Veil or Rocks up it doesn't mean shit because Cinderace has a balanced move called Court Change. If you do run Cloyster, that doesn't OHKO heatran at +2 and gets OHKOed in return by Magma Storm. Oh, if that isn't enough, Will-O-Wisp Volcarona 6-0s from preview. If this isn't considered a -3, no matchup should be -3 lmao.

Ice vs Fighting: The chart says this is an even MU; this is not even by any stretch of the imagination. A lot of the same issues exist here as with fire; you're consistently getting outsped and OHKOed by super effective STAB. Veil's nice here but Urshifu ignores it and just clicks Surging; while this isn't unplayable for Ice it's definitely fighting favored.

Ice vs Bug: Scizor, Volcarona, and Kleavor exist. This is one of Ice's worst mus and is basically an autoloss at preview. How is this not a -3?

Ice vs Water: This is somehow 0 on the chart, but among high level play this is a super good MU for ice due to Freeze-dry being broken. Empoleon builds smash but that mon isn't common on water. How is this considered even?

Ice vs Rock: The chart says this is -1 which suggests only unfavored; but this is an atrocious matchup for Ice for various reasons; I would go as far as to say it's borderline unwinnable (barring stone miss) against someone who knows the matchup. Veil can't be setup safely due to TTar's sand stream, and rock has too many mons that are just too fast and too strong (Iron Boulder, Ogerpon-C), and Ice's main nuke, Specs Kyurem, has problems getting anything done here due to the Special Defense boost of Sand. Avalugg's good but has problems clearing hazards consistently due to Glimmora coming in on it (who, incidentally, punishes Body Press with Toxic Debris); ideally Avalugg switches into the Rock setter and spins but that's unfornately TTar who also has Knock Off, and a Knock'd Avalugg is really bad because Glimmora can force it to take status from Toxic Spikes.

Ice vs Fairy: The chart claims this is even but it's atrocious for Ice; Avalugg is useless here due to Flutter existing, Iron Valiant has no switchins and outspeeds the entire Ice team, and Hatterene straight up 6-0s behind screens.

Ice vs Ghost: The chart claims this is even again, but this matchup is awful for ice; Avalugg's useless and Gholdengo + Ceruledge + Flutter is curtains. To add insult to injury, Specs Pult doesn't care about Aurora Veil and 2HKOs everything with shadow ball.

Ice vs Normal: The chart claims this is even, but, again, it's awful for Ice; Avalugg's useless due to Zoro existing and Normal's content to play the hazard game with SR (which Ice can't prevent from coming up; blissey too tanky). Setup sweepers, which is Ice's perferred win condition, are countered by the omnipresent Ditto.

Overall these are just the more blatant things I've noticed on the Ice MU chart; there are other mus that are wrong that I didn't point out (for example, among other things, Ice vs Flying shouldn't be +3), I'm sure the other types also have similar issues on their matchup chart.

I completely agree with you on this one. I was talking to a few people on the discord and this idea is very prevalent. The main cause could be what aaaa was talking about. Most of the players using types like ice and rock are probably top tier players that are way more skilled that usual, allowing them to make a type feel stronger than it actually is. One thing ethereal sword was talking about is maybe compounding the results of multiple months. I think that's a good idea, however I think the better option would be to compound the result from multiple sources, that way it sort of lessens the effects. For example someone could go to the resource section of past tournaments and compile all the data. The issue is that replays that are too old usual don't work and writing code to process these links is not the easiest thing to do. This is why I also want to try the simulation method as AI usually have the exact same skill level meaning the issue of skill gap is compensated for. This is also a sort of update on the simulation code. I currently have the conversion from poke-env to poke-engine working, however, there are a lot of bugs that I need to iron out as currently, there seems to be a lot of syncing issues. I will try to work on this as much as I can but I do have school stuff that I need to prioritize first.

 

[Ethereal Sword]

Firstly, I want to thank TheKart017 for their incredible work in obtaining these stats. This opens a whole new dimension for usage stats analysis and will for sure be useful as a high-level overview of the Monotype meta.

Continuing the discussion on Discord from a few days ago, I wanted to briefly illustrate just how bad the sample size issues can be when looking at only a single month (or less) of data. Ordinarily the preferred stats to use are 1630 Glicko stats since this best captures mid-ladder usage. 1760 stats are too heavily influenced by individuals, while 1500 stats compromise on quality. Normally this is fine; for instance, last month's Monotype ladder saw 115k games played (so 230k teams) with around 17.5% played above the 1630 threshold. This gives us 40351 teams in February, which is usually more than enough. However, when looking at individual matchups, the number of games gets divided among the 324 matchups, leading to severe sample size issues, especially among rare matchups.
(Disclaimer: some of the steps that follow are not fully rigorous. Also I'm not a statistician so there could be things I missed.)

The win rates in a particular matchup can be treated as a sample proportion from which a confidence interval can be constructed. I will be using this website for that: https://www.statskingdom.com/proportion-confidence-interval-calculator.html. We'll also need type usage stats which can be found here: https://www.smogon.com/stats/2026-02/metagame/gen9monotype-1630.txt.

The rarest types used at the 1630 level are Ice and Electric, with 1.85780% and 1.53617% usage respectively. That means, assuming independence, the Electric vs Ice matchup occurs with probability 0.00028538966. As previously mentioned, 40351 teams were considered for the 1630 stats, meaning the Electric vs Ice matchup occurred only 11 times. For illustrative purposes, let's say that Electric's win rate against Ice is 22.480% as given by the data above (I'm aware that the type usage stats and the matchup stats are not taken from the same games, but we will ignore that). Now this seems pretty bad for Electric, but constructing the Wilson score interval gives a 95% confidence interval of [0.072, 0.52]. In other words, we can't rule out the possibility that Electric might actually be favored in this matchup!
(In reality, as already discussed by Giyu and aaaaahowdogamework, it is not safe to assume independence due to repeated games between players, metagame shifts within a month, team-specific tech, and skill differences. So these numbers should not be taken literally.)

Admittedly, this was the "worst case scenario" in that I deliberately chose the two types with the lowest usage. Repeating the same steps with the two most commonly-used types, Dragon (10.14434%) vs Dark (9.45323%), gives us 387 games played, and using Dragon's observed win rate of 39.942% gives a 95% confidence interval of [0.35, 0.45], which can be interpreted in a more useful manner.

The reason this happens is because the margin of error scales as 1/sqrt(n). So aggregating two months of data (assuming all else remains the same) will cut the margin of error by around 30%, and aggregating three months of data will cut the margin of error by around 43%. (This does not mean that the data should only be reported once every three months, but rather that a running average of multiple months may be a better way to interpret the stats.) This would also require the actual number of games played in each matchup to be reported as well.