ジムリーダーの城　対戦考察Ｗｉｋｉ

I am from Japan and am an expert at GSC NC2000 (Nintendo Cup 2000, Poke Cup).
And I have a habitation of trying to achieve stable and good results with defining the work process (This is true not only in competitive GSC, but in my job).

At Japanese NC2000, I usually got good results by thoroughly team management, but the methods were not very sophisticated, so I got bad results in several torunaments.
For example, I built teams that did not have enough countermeasures against the top tactics of the metagame, like L55 Whirlwind Zapdos with Spikes. I also overestimated passive teams against players whose tendencies were not obvious.

In order to reduce these mistakes, I applied the quality control methods of Japanese companies and refined team management methods.
I applied it for short term testing purposes on NC2000, and then I applied that to tournaments for GSCOU from 2024 (Winter Seasonal #6).

The methods were divided into two (at this time).
One was the method to describe quality information for each team, and the other was the method to manage each team integratively.

In this document, I specifically describe the management methods I applied.

I have gotten used to GSCOU metagame and tournament play.
And, as I adapted to the intense competition of GSCOU tournaments, I established the new method that I had not originally planned.

About the two methods originally defined, the meaning of the former has changed, and the latter has replaced the other method due to almost lost its significance.

The process was rebuilt on 03/23/2025.

I rebuilt the processes that were added during the actual tournament on 03/23/2025.
The main purpose was to delete the bloated "Supplement" section.

In the end, the process was defined in four major sections.

• Team description method (Building Managed Team)
  • Define methods for building and maintaining high quality teams
• Quickly building method (Building Temporary Team)
  • Define methods for building the optimal team for the current specific game
• Amoebic team management method
  • Define methods for appropriately using Managed Teams and Temporary Teams
• Team selection method
  • Define methods for scouting your opponent and selecting appropriate teams

One md file is created for each team to manage quality information.
It is preferable to edit md files with an editor "VS Code", but you can also edit them with any text editor.

Teams managed using this method are referred to as "Managed Team" in this document.

Team information is divided into the following sections.

• Team Paste
  • Write a team paste that can be used in Teambuilder of Pokemon Showdown!.
• Team Revision History
  • List the dates and contents of team revisions.
• Note On Especial Sets
  • For the moveset included in the team, write down the reason for its adoption and how to use it if necessary.
• Weakness Of Team
  • Enumerate team weaknesses. Details will be described later.
• Strength Of Team
  • In the same way as Weakness, Strength is also listed.
• Mutate Options
  • If it is planning to make small changes to the team, write that information. If the change causes a fluctuation in weakness or strength, also write it.
• Quality Control
  • Include information that objectively ensures quality, such as review records by others and game results.

Details regarding Weakness and Strength management are particularly important, so I will provide details.
The image below is an image of actual management.
This is an excerpt of the certain GSCOU team currently in operation.

When enumerating team weaknesses, I divide them into the following five aspects to prevent omissions.
(No such aspect exists for Strength)

The following aspects have used in NC2000 to find the weakness of a team.
（キラパの作り方 (This is written in Japanese)）

• To Be Stalled
  • Weakness caused by team stalled.
• To Be Excluded And Sweeped
  • Weakness caused by certain mons being easily removed by things like Boom, and being exploited.
• To Be Baton Passed
  • Weakness caused by lack of means to stop Baton Pass.
• To Be Switched In And Sweeped
  • Weakness caused by allowing the sweeper to switch in, which can cause the team to collapse.
• To Be Exploited The Selection Limit Of The Three (only NC2000)
  • Weakness caused by selection of 6 -> 3, which aparts the necessary cooperation between certain mons.

In my case, I use the below method on paper for completely enumeration.
(Additions, corrections, and deletions will be made from the actual play, but this should be done before the actual play)

• Intuitive enumeration
  • The five aspects above support intuitive enumeration.
• Automatic checks using tools
  • If you are familiar with Japanese language, you can use the following tools created by CH.
  • This is made for NC2000, but it can also be applied to GSCOU, and I actually use it.
    • 対策くん (This is written in Japanese)
• Reference from analysed similar teams
  • If you have analyzed similar teams in the past, you can save time by referring to that.
• Comprehensive check using VR ranking
  • In the end, I comprehensively check the VR ranking. However, this check does not guarantee completeness, and whether or not you can analyze completely depends on your skill (especially about combinations of multiple pokemons).
  • VR rankings exist for both NC2000 and GSCOU, and the following ones which have been analyzed by multiple skilled players are particularly reliable.
    • Nintendo Cup 2000 Viability Rankings
    • GSC OU Viability Rankings mk. 4

However, there is a problem in that it is difficult to establish clear standards regarding the granularity of enumeration.
For example, if there is a weakness with Clefable of Fire Blast+Belly Drum, there is a question of whether Snorlax of Fire Blast+Belly Drum should be listed separately.
Regarding this, I believe that the essence should be grasped, and in this example the essencial weakness is with the Normal sweeper of Fire Blast + Belly Drum, so I will list only one representative example and not list others.
However, each player may have a different sense of this standard, and even the same player may not always be able to make decisions based on a certain standard.
Problems caused by the lack of clear standards regarding granularity may have a negative impact on scoring and management using it, which will be described later.

Each weakness and strength is scored.
It will be given a provisional score of 1-3 in terms of frequency and 1-3 in terms of severity, and the final score will be the product of these two.

For details on the format of the team description method, please refer to the md file below.
https://image01.seesaawiki.jp/p/d/pbs-thread/JdJhh...

This is an actual team's md file that went out of operation after being in operation for over a year.
This md file contains the word "Temporary Team", which will be mentioned in detail later in this document.
Also, some replays were lost due to an incident in December 2023 （https://www.smogon.com/forums/threads/re-replays.3...）.
https://image02.seesaawiki.jp/p/d/pbs-thread/oNEot...

This is the RMT that I wrote for other players based on the contents managed in the md file.
This RMT is packed with the essence of management content.
https://www.smogon.com/forums/threads/venuquag-off...

Regarding GSCOU, in order to prevent misplaying, I define the playing for the first few turns depending on the opponent's lead.
This is a anonymized version of the Lead Plan that I actually used.
(This was written in the detailed way to explain to teammates in a team tournament)


There is a possibility that the playing will change depending on the opponent's tendency, but the default playing when the tendency is unknown is subject to be managed.

If the opponent's tendency is clear and you can judge that you can take a big risk at the time of team select, instead of teams managed according to the method previously described, you can build and use a disposable Temporary Team.
However, to control risk, the following rules must be observed.

• Be sure to test play the built Temporary Team against the team that is expected to be used.
  • A team built ignoring normal methods may not work as intended on paper.

• Don't give up on your efforts to ensure consistency.
  • Temporary teams also require a minimum level of consistency, as there is a possibility that an opponent may suddenly change his tendency. Consistency can be easily ensured by building a Temporary Team based on the structure of a team that guarantees consistency, such as a managed team or sample team.

• Be sure you have a plan which uses Managed Team.
  • If unexpected changes in tendency are observed, you should immediately switch to Managed Team after this game to reduce risk.

This is a new team integrated management method that replaces the "Integrated management of teams method" and is based on the combination of Managed Team and Temporary Team.
This method consists of "Team Classification" which supports your team building and "Managed Team Portfolio" which supports the lower limit of your performance.

Played against humans on equal conditions, this game requires you to take risks in situations with the highest expected value in order to win.
The way you build a team differs greatly depending on whether you take risks during the game or before the game begins. For both approaches, the methods for building teams are more divided.
The following classification is for GSCOU. This will be especially useful as a guide when you build a new Temporary Team.
GSC OU Team Classification

This is a portfolio of Managed Teams.
This portfolio must have the following characteristics:

• All teams are Managed Teams to guarantee team's consistency
• Each team has a different architecture, and weakness is distributed
• Each team has outstanding versatility worthy of the name "Standard" or some important strength, to pose some kind of threat to an opponent

If you can create a team portfolio like this, you can simply select a team from this portfolio and pose some kind of threat to your opponent with keeping the low possibility of fiasco of your team building/playing and low exploitability.
Even a random selection can be effective, and of course it's even more effective if you can see the vague tendency of your opponent and select based on them.
This is effective to ensure your minimum performance in situations where it is not appropriate to use Temporary Team, such as cases where the opponent's tendency is unclear, cases where sufficient preparation time cannot be keeped, cases where an unknown tendency is observed in Game 1 or Game 2 of BO3, etc.

This does not sense that the more teams which make up this portfolio the better.
From my experience with NC2000 and GSCOU, 8 to 10 is probably the appropriate number.
If there are more than that, there is a risk that low quality teams which are unsuitable for Managed Team will be mixed in, or increasing exploitability due to increase teams with similar architecture.

To form a suitable portfolio, the numerical management described in the "Integrated management of teams method" section can be helpful.
(This is only useful as a guide and is not required)

Below is my Managed Team Portfolio as of 03/01/2025.
(Since this document is not intended to explain individual teams and requires minimum countermeasures against scouting, the description is kept to a minimum)
Now that I'm no longer in a position to vigilar to scouting, I'm posting this with a specific team paste.

• Lax - Zap - Cloy - Golem - Gar core modern standard offense team （https://pokepast.es/23813b5953e16eed）
• TarNidoChamp's standard offense team which is close to the sample team （https://pokepast.es/981a71e6faff6e5a）
• Borat standard offense team which is close to the sample team （https://pokepast.es/b01ac5194207a411）
• Double Elec offense team as a modern standard offense with P2 as a niche element （https://pokepast.es/12f5dd5f83c4762e）
• Niche offense team consisting of Heracross to punish Nidoking, Tyranitar, Machamp and Marowak （https://pokepast.es/609d18f5265ecaa2）
• Niche offense team consisting of Rhydon to punish Curse RT Lax etc and Venusaur to punish Vaporeon etc （https://pokepast.es/6012fd1933b24a1d）
• Passive offense team which is tuned to be especially strong against Stall teams and Double Elec teams （https://pokepast.es/96a005c216184861）
• Passive offense team which is tuned to be especially strong against Boom offense teams and Marowak offense teams （https://pokepast.es/708a6a64faba19da）
• Stall team which is highly responsive to modern Spikes War （https://pokepast.es/bd72e9421794d1f4）

I have been writing about the process of building and managing a team.
And if you are a player with good experience, you will also realize that it presupposes scouting.
I also define a process for team selection, which includes scouting.

I define the team selection method in the following five steps.
Some steps will be skipped depending on the time given before the game starts.

The remaining time may be suddenly lost or increased due to your personal circumstances, your opponent's circumstances, or the circumstances of a tournament.
Therefore, if you repeat these steps several times and gradually improve your level of preparation, you will be able to respond flexibly with your best perfection on your current circumstance.
For example, the progress is as follows (if you are planning a game in 5 days and can spend about 2 hours a day preparing).

• On the first day, watch some replays of recent tournaments (Steps 1-2) and select teams (Step 3).
• Over the next two days, watch and analyze a sufficient amount of replays (Steps 1-2), and if you have any concerns, build a new team or select teams again (Step 3), or analyze your opponent's play for optimizing your play (Step 4).
• On the last two days, do test play the uncertain cases (Step 5). Depending on the results, perform additional work on the previous steps.

The research targets are mainly replays, but if there are articles like RMT, it will also be included.
The research targets are follow things.

• Public listed replay (as a guide, from the past six months to one year)
• Posts on Smogon Forums (This can be obtained from the Smogon Forums profile page)

If necessary, you can also research the following.

• Public listed replay (over one year ago)
• Public unlisted replay (If there are tournament participation records which can be researched by Posts, it may be possible to obtain them by scanning the replay URL number)

Also, while gathering the above information, you will naturally find out whether the opponent is an experienced smogon player.
(This will be explained later in Step 2)

In addition to the above, if you have any personal information (such as play style or whether you know each other), that would also be helpful.

If it's not a smogon tournament, it may be difficult to even know the Pokemon Showdown account.
In that case, for example, the tournament is on the specific Discord server, you may be able to access the information on that server and use it as an alternative.

And If it's a team tournament, you should get the information with the help of your team.

You find an exploitable tendency from the opponent's information.
In particular, it's important to investigate what cores, leads, species, and movesets teams prefer.

Also, important information is whether the opponent will prepare a specialized team for you.
This can often be inferred from posts such as tournament participation reports or whether or not you know each other.
In the case of judging from replays, if there is a sudden change in tendency or adopting weird core, there is a higher possibility that a specialized team will be prepared. Hiring niche mons can also be a sign of a specialized team.
If the opponent is an experienced smogon player (this can be inferred from the number of replays of all gens, number of reactions on Smogon Forums, number of batches on Smogon Forums, presence of custom avatars, etc), the possibility of being prepared a specialized team against you is increase.

If you're planning to surprise the opponent, you'll need to watch their replays and check their playing habits.
In particular, the first few turns of play for each lead combination are highly reproducible and serve as a key confirmation point.
If your opponent is the type of player who brings in a specialized team, you may be able to surprise it by bringing in a team which you don't normally use (even if it's a standard team).

Please note that shallow career players are more likely to sudden change in tendency. Past information may no longer be useful.

You build and select a team which will exploit the opponent's tendency confirmed in Step 2.
It is desirable to build a temporary team which can aim for a higher win rate, but if the opponent's tendency is unclear or there is no time, select an appropriate team from Managed Team.

If the opponent is likely to prepare a specialized team, you should think about protecting you from it.
Be aware of what kind of tendency other players think you have, imagine what a specialized team is for you, and refrain from using teams which are weak against it.
Even if the opponent is a newcomer to this format, if that player is an experienced smogon player, it's dangerous to ignore the countermeasures of specialized teams (and specialized play mentioned in Step 4).

If you are participating in a team tournament, it is a good idea to ask for help in coming up with ideas in this step.
Teammates may suggest better tools than you think, and it has the effect of mixing the tendencies of multiple players, which naturally protects you from specialized teams.

Against shallow career players, rather than following past tendency, it is better to exploit the lack of knowledge, for example, by using the following strategies.

• Curse EQ Rest Lax
  • It is not only stable, but also has the power to end the game when the opponent makes a mistake.
• Spikes War Oriented Stall Team (like Umbreon Stall)
  • Stall teams can passively respond with observing the opponent's play, so they tend to show differences in ability itself. Furthermore, this type of Spikes War requires knowledge specific to GSCOU.
• P2 Offense
  • Some niches, such as P2, have few fatal weaknesses, but the opponent may not know the correct way to play against them.

Regarding the team selection randomization function introduced in the "Calculation of selection probability for each team" of "Integrated management of teams method" section, I don't feel that it has any effect on alleviation of risk at all.
This is probably because the number of games in a tournament is far from being a sufficient number from a view of statistic, so the outcomes are discrete.
Furthermore, simple numerical management such as the one introduced here may not be able to cope with actual activities around specialization against the opponent.

You can specialize not only teams, but also play.

Checking playing habits was mentioned in Step 2, but this is important even if you don't plan on surprising them.
In addition to playing the first few turns, the timing of using moves such as Sleep Moves, Thief, and Boom, which can only be used a limited number in the game, should be checked.
Once you check this, you can plan your advantageous plays accordingly.

However, if this habit is unclear, do not force making the play plan.
This is a characteristic which sometimes found in high level players.
That player will probably try to plan specialized plays just like you and will play just like usual you.
(I don't know if that's actually the case, but it's better to think so)

You test the team selected in Step 3.

It's recommended to perform this test play by loading the opponent's assumed team according to the opponent's tendency confirmed in Step 1-2.
Also, if possible, it is desirable to play according to the opponent's play habit confirmed in Steps 2 and 4.

When you think of test play, you may imagine having someone help you, but it's probably more comfortable to do this type of test play alone.
This test play requires an understanding of the opponent (you should know more about your opponent than any of your teammates), and will likely require multiple redos.

However, if you want to test a completely unexpected situation, you may be better to ask someone to help you.