Copyediting Battle Tree Mechanics and Guide [GP 0/2]

Part 1: Introduction, Mechanic and AI


Greetings and welcome on this article focused on the latest iteration of Pokemon post-game Battle Facilities: the BATTLE TREE!

After beating the Elite 4, Poni Plains finally become accessible, and after beating Dexio and getting the ability to use Mega Evolution, the player is finally asked to confront Red or Blue (it's possible to choose either with no difference outside of the Pokemon used) in order to gain access to the facility. The Battle Tree reintroduces most mechanics we are familiar with from the ORAS Battle Maison, however reducing the amount of modes available to the player due to Triples and Rotation battles not being present at all in Sun/Moon. In similar fashion to other "first set of titles" of previous gens, there's no Tutors available for BPs (they are however available in UltraSun and UltraMoon), but several evolutionary items, held combat items, and Mega Stones are available for buying. The character which enables IV checking is also present at the facility, and will enable IV checking from the PC after the player has hatched 20 Eggs and talks with him.

Victories in any mode will award the player with Batte Points (or BPs) that can be used either at the nearby vendors or the Royale Dome to buy items. In Normal battles, the player will get 1 BP per victory upwards to 10, then 2 BP per victory in battles up to the 20, and a final bonus of 20 for beating the Legend. In Super battles, the player will receive 2 BP for battles 1-10, 3 for battles 11-20, 4 for battles 21-30, 5 for battles 31-40, 6 for battles 41-50, a bonus of 50 BPs for beating the Battle Legend at 50, and 7 BP per victory afterward.

The Battle Tree allows the player to challenge the AI in Singles, Doubles, and Multi Battles, the latter also allowing to play with any AI Trainer beaten in the other 2 modes if you choose to invest 10 BP in "recruiting" them. In similar fashion to Maison, the player is asked to complete the normal mode before being allowed into the Super mode of a category, once more having to win 20 battles in a row and then face a easier version of the "Battle Legend", in this case Red for Singles, Blue for Doubles, and both at same time for Multis at the 20th battle. Red will also gift the Kanto starters' Mega Stones once beaten for the first time in Singles. The difficulty, however, is considerably higher compared to previous battle facilities, as enemy Trainers will have access to Z-Crystals and eventually Mega Stones and legendary Pokemon, with the only actual roster advantage for the player being access to the Tapus and Ultra Beasts that the AI does not have in any set.

[note for HTML editing: the following section would greatly benefit for readability if put on some tables rather than a list]
Every time a streak ends (to loss or retreat), the following day you will be rewarded with the following prizes, depending on how many battles in a row you went through before the ending.
5 Wins: Moomoo Milk
10 Wins: PP Up
20 Wins: Rare Candy
30 Wins: Bottle Cap
40 Wins: PP Max
50 Wins: Ability Capsule (Gold Bottle Cap in USUM)
100 Wins: Lansat Berry
200 Wins: Starf Berry


There's also 3 vendor NPCs on the left, which can trade your Battle Points for evolutionary items, important battle items, or Mega Stones.
Specifically, you can get the following evolutionary items from the first vendor:
King's Rock 32BP
Deep Sea Tooth 32BP
Deep Sea Scale 32BP
Dragon Scale 32BP
Up-Grade 32BP
Dubious Disc 32BP
Protector 32BP
Electirizer 32BP
Magmarizer 32BP
Reaper Cloth 32BP
Whipped Dream 32BP
Sachet 32BP

The second vendor will offer the following equippable items:
Toxic Orb 16BP
Flame Orb 16BP
Iron Ball 16BP
Ring Target 16BP
White Herb 32BP
Mental Herb 32BP
Power Herb 32BP
Focus Sash 32BP
Air Balloon 32BP
Red Card 32BP
Eject Button 32BP
Weakness Policy 32BP
Choice Band 48BP
Choice Specs 48BP
Choice Scarf 48BP
Life Orb 48BP
Rocky Helmet 48BP
Assault Vest 48BP
Safety Goggles 48BP
Terrain Extender 48BP
Protective Pads 48BP

And the third vendor will offer all the Mega Stones for the Pokemon present in the base Alolan Pokédex for 64 BPs, or all the Mega Stones except Diancite if playing on UltraSun or UltraMoon.
Specifically, Gengarite, Scizorite, Pinsirite, Aerodactylite, Lucarionite, Kangaskhanite, Gyaradosite, Absolite, Alakazite, Garchompite, Sablenite, Metagrossite, Sharpedonite, Slowbronite, Glalitite, and Salamencite.



BATTLE TREE MECHANICS

In this section, we'll cover on the rules used by the facility. Those rules affect both the player and the AI.
First, a quick rundown of what is allowed to be used on Battle Tree.

1) Pokemon allowed
The Battle Tree allows the use of the entire Pokedex, excluding Mythical Pokemon, Mewtwo, Lugia, Ho-Oh, Kyogre, Groudon, Rayquaza, Dialga, Palkia, Giratina, Reshiram, Zekrom, Kyurem, Xerneas, Yveltal, Zygarde, Necrozma, and the Cosmog evolutive line. In fact, several post-40 victories AI trainers will feature several legendaries if not mono-legendary team compositions, while also featuring Mega Evolutions occasionally, so the player has to keep in mind that using suboptimal Pokemon will inevitably result in a pure power disadvantage once the first easy AI sets are left behind. As usual, the Species Clause is active, so your team must feature up to 6 Pokemon with different Pokedex number.

2) Items allowed
Every item available is allowed on the Tree, with the usual Item Clause (a team cannot have item duplicates). It's worth mentioning that while the AI Trainers are coded to respect the Item Clause, there's still the possibility of an AI trainer generating with several Mega Stones or Z-Crystals, resulting effectively in Pokemon generated with items they will not be able to use. Consumable items are not consumed on use, nor is using Fling, Thief, or similar mechanics permament, since the Facility replicates PvP mechanics. Note that the Item Clause applies to the entire team box submitted, even Pokemon that are in the team but not selected to be used, so you are not able to register a box with two Pokemon holding the same item even if you only intend to use one at time.

3) Moves/Abilities allowed
Differently from what one might be used to by playing on Showdown, there's no actual ruleset in regard to how moves are used. You can put as many Pokemon to sleep as you wish, evasion and OHKO moves are allowed, and Smeargle can essentially be run with any move existent in the game. The same goes for abilities; all abilities are allowed, including the usually competitively banned Moody, Snow/Sand Veil, and even unreleased ones. The AI can in fact generate with unreleased abilities (currently, the only ones being the Hidden Abilities of Oranguru, Passimian, and Heatran), and there's a few AI sets that do run items not available in the current games (such as Custap and Enigma Berries).

4) Pokemon levels
Also a returning feature from the latest facilities, all the Pokemon in the player's team will be temporarily downleveled to 50, even if normally they could not exist below that level as they evolve later. However, they are not leveled up to 50, so any Pokemon below 50 will remain at that level.

The way the player selects Pokemon in Battle Tree is identical to Battle Spot and VGC: you are asked to enter with a selected team that fits the Clauses (either one of the saved teams, a QR team, or your party), and select 3 for Singles, 4 for Doubles, 2 for Multis. For Multi Battles, the Clauses are applied per Trainer, which means the 2 Trainers (valid for both AI and players) can have duplicated Pokemon, duplicated items, and as well use one Mega Evolution and Z-Move each if they so please. Once your Party is selected, you will continue to battle with it in same order until you either decide to take a break or lose. Taking a break has no effect on the ongoing streak, so you are allowed to freely swap things around in between battles if you decide so, and even change your partner for Multi Battles. You are also allowed to Scout the Trainer you just beat to use in Multi Battle with AI for 10 BP, and it will use the first 2 Pokemon it had in that specific battle (so the lead + the 2nd Pokemon if it was Singles, or the two leads if it was Doubles).

Every AI Trainer has a fixed set of Pokemon to access, with the IVs depending on the number of current battles. The Pokemon that the Trainer is generated with are completely random and vary each time you battle them, so for Trainers with particularly big rosters, every battle will likely be different from the others. Most Trainers also feature several sets of the same Pokemon, which means it's uncertain which set the Pokemon is running until it either reveals a move or the item equipped. The abilities the AI Pokemon are generated with is also random, with 33% chance for each ability including the Hidden Ability. A Pokemon that only has a normal ability and a Hidden one will have 66% chance to generate with the normal ability and 33% to have its Hidden Ability. However, if you Scout a Trainer, his Pokemon will always have the abilities they had at the moment of the recruitment when used in a later Multi battle.



AI TRAINERS GENERATION AND RANDOMNESS

Before going into the details of how the AI chooses what to do, a little emphasis must be given to the biggest cause of frustration and "cheat impression" from Battle Tree: randomness. As it can be understood from the previous paragraph, a lot about what you will be facing will be randomly generated. Pokemon will have random sets, random abilities, and are also not 100% predictable. When facing a Pokemon, you should also be wary of what movesets and stat spread each specific Trainer can carry, as well as the possible items it can have and what abilities the Pokemon has access to if they are not revealed istantly by a screen message (like Intimidate or Pressure).

1) AI Pokemon IV, EV, Natures, and Abilities
In contrast with the random nature of the rest of the Pokemon generation, all sets have already fixed EVs and Nature. The IVs are also fixed, but they are based on the Trainer you are facing rather than tied to the specific set. It is important to note that this includes Pokemon designed around the usage of Trick Room (usually very slow with a nature that reduces Speed), which results in those Pokemon having Speed IVs higher than 0, so if you are running correct spreads your Pokemon of same speed tier will be always underspeeding them outside of the rare Iron Ball sets. The IVs for the AI will progressively increase as you progress through the facility. While it is not perfectly omogeneous, the enemy IVs will start at 19 in Super Battles, then will eventually hit 23, 27 and finally 31 IV once you reach the Trainers that appear only past battle 40. As mentioned in the previous chapter, the Ability every Pokemon gets is random between the ones available, with 66% chance to have a normal ability and 33% to have its Hidden Ability. This also occasionally leads to weird bad sets where, for example, a Guts or Poison Heal Pokemon holding a Toxic Orb is generated without the needed ability, ending up only harming itself. To balance out situations like those, several sets that are based on a specific ability of the Pokemon also have a "Backup plan": for example, the sets that rely on a combination of Sturdy and Metal Burst or Endeavour will also often feature a Focus Sash as equipped item, and sets reliant on weather will often also have access to the related weather setting move.

2) Team generation
Every time a new Trainer is generated, the game picks one of the Trainers available for the number of battle you are at, then picks the required amount of Pokemon from its team one by one making sure that the Clauses are respected. Since those are all the conditions for creating the enemy team, the AI can generate with teams that do not synergize, or even Pokemon that harm each other in some cases, as well as have perfectly synergic compositions if the dice rolls are favorable. Some Trainers have dedicated teams who are guaranteed to synergize, often Weather or Trick Room based, while others have a very big roster of different Pokemon to pick from, resulting in very unpredictable compositions to face.

3) Apparently unpredictable AI behaviours
There are instances which will often create confusion or frustration for the player, with apparently inconsistent behavior by the AI. As your streak continues, you will inevitably witness odd actions such as consecutive uses of Protect, weird Pokemon swaps, using resisted or nullified moves while clearly superior alternatives exist, and so on. Typically, the AI prioritizes moves based on power and effect (accuracy is not a factor considered), but the move selection is based on a potentially weighted roll, so expecially in situations where there's no guaranteed OHKO, there's a significant chance for the AI to not select the highest damage attack. It's these instances you'll notice the AI selecting suboptimal moves which merely inflict any kind of damage or status, unlike the choices a person might make. In Doubles, they may not select the most obvious target for their attacks either. The AI may also switch one of its active Pokemon with one from its backline in order to resist the last attack that connected, to absorb it with an immunity, or to circumvent move-locking into a ineffective move. If a swap into a resistance is performed, the incoming Pokemon will also have a move that can hit your active Pokemon for supereffective damage.


AI BEHAVIOUR: ATTACK PRIORITY, RECOGNITION AND COMMON CHOICES

Even though the AI is generally never fully predictable, there are a few events that appear to have higher priority, which can allow you to have a general idea of what will happen. It's still important that even if such events have a higher chance to happen, , the AI can still do something else, so do not act blindly and always ponder your actions before clicking on the DS screen, expecially late in the streak where an error can easily turn in a loss and streak drop. Note that several of those behaviours are limited to Doubles/Multis since they are triggered by interaction of multiple Pokemon.

Part 1: AI move selection priority and common behaviours

- Setup moves
Most AI sets that are based on boosting moves tend to attempt to set up until a certain breakpoint before committing to attack. For moves that buff several stats like Quiver Dance and Dragon Dance, the breakpoint looks like either outspeeding your Pokemon or being able to OHKO them. Several Stockpile+Rest/Roost users will usually attempt to reach 2 stacks and then heal up if needed, occasionally going to 3 if not damaged. Other boosting moves aren't as consistent and cannot be predicted reliably.

- Weather and Terrains
If an AI set is running a Weather or Terrain setting move and the specific Weather or Terrain is not up, the AI generally priorizes setting it up if there's no option to OHKO an active enemy right away. This also occasionally leads to funny interactions of AI Pokemon in Doubles taking turns setting different Weathers, often Sandstorm and Rain since a lot of Trick Room specialists run both slow Rain setters and some Sand setters.
It's worth mentioning that most Drizzle, Drought, Sand Stream and Snow Warning Pokemon also have the dedicated Weather setting move in their movesets to artificially balance out the chance of not having the necessary Ability, with the notable exception of Aurora Veil users who do not have direct access to Hail themselves aside from Alolan Ninetales' Hidden Ability.

- Speed control
Moves that affect Speed such as Tailwind and Icy Wind are very high in the priority list for the AI. Tailwind nearly always is set anytime is not up and Trick Room is not active, and Icy Wind is generally spammed by the Pokemon who have it, expecially in Doubles where the AI priorizes spread moves. Trick Room is also usually priorized if your active Pokemon are faster than both the AI's, and occasionally the AI might try to revert Trick room if it is active and your active Pokemon are slower than their. Note that in case the AI threatens a KO that turn, a Trick Room setter can still choose to go for the attack instead. It is important to note that if two Trick Room setters are active at same time for the AI, they will never try to Trick Room at same time, one will always use another move instead.

- Phazing
Moves that will force a swap (Roar, Whirlwind, Dragon Tail) have high priority for the AI if your Pokemon have boosts active. While they aren't particularly dangerous on their own outside of denying your boosts, it is important to keep track of phazing spam when the AI happens to have had the chance to setup the occasional entry hazard or you are relying on setup moves. Take advantage of the negative priority of those moves to use Taunt or take out the Pokemon before it can become a issue.

- Paralysis and Sleep
Status inducing moves are very high on the priority list for the AI, expecially Paralysis since it doubles as status and speed control. If a set runs those, you can be expecting it will be attempting to status you first unless your active Pokemon is immune to it. Most sets featuring Hypnosis also have Zoom Lens or Wide Lens as equipped item to patch up the low accuracy. Will-o-Wisp is usually priorized if your active Pokemon have high Attack, while Toxic does not look like having a particularly high priority instead.

- Damage over accuracy
While the AI appears to be able to calculate when it can OHKO your Pokemon, it does not factor in the accuracy of the moves. As result it often will go for a high damage supereffective Focus Blast, rather than secure a KO with a Psychic. While you can't realistically always plan ahead of the move selection, a 4x weakness is a nearly guaranteed move selection, and the AI ignoring the possibility of missing high damage moves can definitely cause some lucky misses for you.

- Fake Out (and Fling flinches)
If a Pokemon has Fake Out, it will use it the first turn the Pokemon is out, unless something prevents it from happening (Ghost Immunity, Psychic Terrain, but the ability Inner Focus does not prevent the AI from trying). There are also a few Ambipom, Infernape and Weavile sets that have Fake Out and Fling with an item that causes Flinch, on top of U-Turn or Thief, resulting in one of your Pokemon essentially being forced to skip two turns (or more), while watching the enemy steal your item and get away with it.

- Priority on low HP
If one of your Pokemon is at very low HP, expecially due to a Focus Sash or Sturdy, and the AI has access to a priority move, it will generally use it the turn after. Consider using Protect, swapping, and if playing Doubles plan your other Pokemon's action in order to not give give a completely free KO.

- Stall sets
Some sets run a completely stall oriented moveset: usually a damage over time move (Toxic, Leech Seed, Sandstorm, Hail, Wrap) as well as a mix of Stockpile, Substitute, Protect, Double Team, Leftovers, Recover, and similar moves or items. There's a large number of AI Pokemon with this kind of moveset, with many variations, and their general behaviour is to alternate the damage over time move to Protect or Substitute, and setting up their defensive or evasion boosts every now and then, occasionally even chaining several Protect uses in a row.

- Sending in Pokemon with supereffective coverage
When you beat an enemy Pokemon or a swap is forced and the AI has still multiple Pokemon in the back, the AI will priorize sending in the Pokemon with the highest damage potential. When a new Pokemon is sent out you are usually facing 2 possibilities: it is the Pokemon with the highest damage potential against your current active Pokemon, or if it does not have any super effective attack, it means the AI has no other super effective to use.

Part 2: Moves and abilities the AI does and does not recognize

- Encore
If a AI Pokemon is Encored on a status, setting or move you are immune to, it will generally either switch the following turn, or use the move once, fail and switch out turn after if still possible. If a Pokemon is Encored in a setup move, however, it might keep using it up to 3 times and possibly swap out after 2 or 3 uses if Encore is still active.

- Taunt
The AI will generally not swap out if Taunted if the Pokemon has at least one usable move. Pokemon forced to Struggle because of Taunt will occasionally not swap out, and sometimes will just KO themselves.

- Substitute and status moves
While the AI will still try to status a Substitute every now and then, but it will usually stop after one failure, unless the specific set only carryes status moves.

- Misty Terrain
The AI does not properly recognize Misty Terrain. Moves that have both a damage and status component like Nuzzle, as well as occasionally Swagger and Flatter, do not trigger the immunity message and the AI can attempts to use them despite the immunity to the status provided by the terrain, giving you free turns and even free boosts. Fliers/Levitaters are often the ones not recognizing the presence of Misty Terrain and attempt to use status moves despite the immunity message, expecially most Rotom variants who carry one or two status moves. Note that this does not happen for Rest: the AI will never attempt to use Rest if affected by Misty Terrain (as well as by Electric Terrain).

- Lightning Rod, Storm Surge and Levitate
Differently from Misty Terrain, the AI does recognize those abilities, and will stop using moves that trigger them, or swap their Pokemon if they don't have any usable move left due to those abilities. In fact, it can use those against you by occasionally swapping their Pokemon with one of them that can absorb or redirect your last attack.

Part 3: Doubles and Multi Battle special interactions

- Spread moves
Differently from Singles, the AI loves to priorize spread moves in Doubles. If one of your Pokemon is weak to a spread move and single target ones will not pick a OHKO, generally the AI will go for the spread option. Occasionally not caring if this hits his own partner.

- Explosion
The AI use of Explosion is generally inconsistent, and while the AI will often commit to it first turn, sometimes it will wait or not use it at all. Being matched with a partner which will be immune to it, the Pokemon or its partner being low enough that it'll potentially die the same turn, and not having any super effective option to use usually trigger the usage. Do note that one Lickilicky set carries a Normalium-Z with Explosion, and the AI will always use it before Explosion.

- Earthquake, Sludge Wave, Discharge and Surf
The AI likes using Earthquake anytime the other active Pokemon is a Flying-type or Levitater. If this combination happens (in fact, several trainers have rosters that cause it to happen) the Earthquake sequence is nearly guaranteed. This is also somewhat applicable to Surf matched with Water Absorb, Storm Drain, and Discharge with Ground-type Pokemon or Lightning Rod and Volt Absorb, the rare Sludge Wave sets when matched with Steel-types, and Explosion while matched with a Ghost-type. The rare Telepathy Pokemon also contribute to activate this mechanic.

- Wide Guard
The AI is not coded to play around Wide Guard. It does not use it reactively if it is available (rather, often uses it even if you are not using any spread attack), and does not react to you using Wide Guard either. Combined the AI giving priority to spread moves in Doubles, and several AI sets only running spread moves, Wide Guard can completely deny certain sets if used correctly.

- Protect with Spread moves
Even though Protect usage is not consistent and generally not predictable, there's a few situations in Doubles where the AI decides to Protect, usually when matched with a spread attack user. If the situation is favorable to using a spread move, you can often expect the other Pokemon to Protect that turn, but it rarely attempts to double Protect in those situations. Despite this, there's an interesting interaction that causes the AI to Protect even if it's immune to the partner's spread move due to typing or ability.
 
Last edited:
part 2: Guide

GENERAL POINTERS FOR SAFE RUNS

In order to have a successful run in the Battle Tree, it is important to consider all the mechanics and AI sets and make sure you are not completely walled by common occurrences. Due to the volatile generation of teams, you might get lucky and never run into your counter, or eventually run into it and get your nice looking streak interrupted. It is obviously impossible to have an answer to everything considering there's hundreds of different sets and virtually infinite combinations, but there are a few general basis your team composition needs to cover in order to produce reliable post-legend streaks. Since Trainers after beating Red/Blue do not change anymore, once you are past battle 50 the game becomes a endurance challenge that puts a stress on what your team can answer to and what it cannot.

- Have a backup plan
This should always be the first thing to consider: do not have only one strategy. Your Rain based fast and high damage team might look strong, but it's a matter of time until you run into a slower Weather setter or a Pokemon with Cloud Nine and you get swept away because your fast Pokemon are now slower. Expecially for Doubles, you need to have a reliable answer in the event that your strategy is denied on the first turn or turned against you due to a combination of enemy abilities or movesets. A few examples would be having a Lightning Rod user for Rain teams in order to deny enemy Thunder users, Misty Terrain or Aromatheraphy access if your team is weak to status, a staller if your team is fast and need to answer a Trick Room going up, a Fire- or Ground-type to answer opposer Sunny teams, etc. Always consider that your plan can go wrong and you might need to buy time. Expecially when it comes to Weather or Trick Room, since there's AI Trainers focused on every Weather and both Trick Room specialists and high speed dedicated Trainers, which you will inevitably run into at some point.

- Stats matter
The AI has access to fully EVd Pokemon, which eventually will have 6x31 IV. Once you are past the first 20 battles the Trainers will start using actually competitive Pokemon, eventually including plenty of Legendary Pokemon and Mega Evolutions, as well as a number of both cheesy and brutal sets that require to be dealt with quickly. You will need to use actually strong Pokemon that are also properly trained, and unfortunately your story team will likely have to be replaced for this purpose. Put your time into engineering a proper composition, breed and train the Pokemon you want to use, and choose the right items, because else you will inevitably fall to the sheer higher power of the AI sets. If you plan of using a sub-optimal Pokemon, make sure your team composition actually benefits from its presence and is well equipped to deal with opposing threats.

- Check the AI Trainer data
Another important thing to consider is that the full data about the AI sets is available, and you should use it to your advantage. Going in blind, considering the amount of different sets plus the existance of several unexpected sets can quickly lead to a disaster. Some examples include a Speed invested Alolan Marowak, a evasion focused Zapdos set, OHKO move users, the list is very long and varied. When you start a new round, check the Trainer name, look for it on one of the many available resources, note which sets for each Pokemon it has available, and plan accordingly. If the Trainer can have several sets of a specific Pokemon, make sure to not exclude either when picking your moves until the set is revealed by a move or the equipped item activating.

- Accuracy over power
It is a typical error coming from competitive Pokemon PvP to consider moves like Fire Blast, Hydro Pump, Stone Edge, which are often taken due to them securing more OHKOs than their weaker but accurate counterparts, good as well for facilities. Unfortunately, it is not the case. When going for long runs, one must remember that misses do happen. Missing a move in wrong moment can turn in a loss, and once you start to accumulate battles, the rolls eventually will be unfavorable. While competitive formats are usually "best of X", and unlucky match generally won't matter seeing you still have the chance to fall back and try again, the Battle Tree like other facilities does not allow the rematch and you always have to start from 0 after a loss. Unless you really don't have another option, try to only run 100% accurate moves. Do not put yourself in situations where a miss equals Game Over, and consider that the AI has both evasion moves and the occasional Lax Incense / Bright Powder equipped, further punishing low accuracy.

- Dealing with stallers
One of the biggest threats if left unchecked on the Battle Tree are staller sets. They might not look particularly threatening at start, but once they have accumulated several evasion stacks and are dropping your Pokemon's HP with Toxic or Sandstorm while you cannot hit them other than with a lucky roll, it surely becomes a frustrating way to lose. There's several ways to deal with stallers, one of which being istantly targetting them down, but it's not reliable since several of them run Bright Powder or Lax Incense thus could get lucky misses, or there might be a more dangerous threat to deal with in Doubles, on top of most staller sets actually being on bulky Pokemon. It is recommended to include somewhere Taunt (in fact, several staller sets have no damage move and will just Struggle to death once Taunted), or status immunity/cure, run a Steel type which is both immune to Toxic and Sandstorm, having a perfect accuracy move to deal with Double Team spammers (Smart Strike, Aerial Ace are both TMs, Z-Moves can also work in a pinch but might not be available when you need them), Haze or Clear Smog to clear boosts, and similar alternatives where your movesets allow.

- PP UP your Pokemon
If you decide to run your own staller (very common in singles, and viable to have in doubles), absolutely PP UP it. The AI only has base PPs, and in the possibility of a stall war, you want to have the edge. In fact, PP-stalling the AI is a common strategy for dealing with some threats. While maximizing the PP of fragile sweepers is generally unnecessary as they rarely get to live more than 3 turns, it's recommended to still PP UP them, or at least important moves with 5 or 10 PPs that need to be used more than once per fight, in order to deal with the possibility of Pressure Pokemon adding up, which is likely to happen against post-legend Trainers as lot of Legendaries that will show up do have Pressure as possible or only ability.


BASIC TIPS FOR TEAMBUILDING

This section will provide some important tips for building a team composition depending on which mode you are playing. While they are not mandatory guideline for success, following them will provide you a possibly painless experience for your first runs, allowing you to practice the mechanics better and eventually be able to build your own compositions.

- Singles
There are a few notable aspects important to singles, which are linked to the limited pool of Pokemon available: you will not able to have a perfect coverage or always outspeed & OHKO everything, so the composition must be built with contingency plans.

Hyperaggressivity can get relatively far when using very strong Pokemon like Mega Salamence or Garchomp, but will eventually be prone to RNG past 40 battles, where the AI uses both strong competitive sets (including Scarfed high BST Pokemon) which could outdamage you, and very strong walls which might end up statusing you to death.

The key aspects of a safe singles composition should be a strong setup based Pokemon (preferably with a realistic way to boost its Speed, something that Dragon Dancers excell at), at least a safe switch in for this Pokemon who can handle threats that the main sweeper cannot (usually relying on typing cores like Steel/Dragon/Fairy who can swap from the weaknesses of their sweepers to the immunity/resists of the bulky Pokemon) and then either another strong offensive threat which complements the main one, or a second defensive Pokemon who complements the other one.

Access to Substitute and reliable recovery on the main damage threat is a big plus, as it allows to screen from potential status, OHKO moves and pure RNG shenenigans with Quick Claw or evasion. For this reason the main sweeper is often picked with a single or two attacks, and a combination of Substitute, boosting and recovery.

The idea is to find a setup opportunity for one of the 3 Pokemon on the lead of the opponent, and once your Pokemon is ready, proceed to 1v3 where in the worse scenario your remaining forces can pick up the last weakened enemy. As you cannot know what the backline of your opponent is, have to rely on the lead it presents for maximum safety.

Some examples of successfully Pokemon and some cores for singles:

Damage threats:

Recommended set:
Dragonite @ Lum Berry
Nature: Adamant
Ability: Multiscale
EVs: 252 Speed / 252 Attack
Moves:
- Dragon Dance
- Roost
- Outrage / Dragon Claw
- Earthquake / Fire Punch
With its Multiscale, Dragonite is virtually impossible to OHKO aside from a OHKO move or Mold Breaker. This provides with a solid opportunity to set up with Dragon Dance on neutered enemies, have a safety net against most backlines, and once enough boosts are accumulated STAB Outrage can wipe most non Fairy-types off the game. The coverage of choice is usually Earthquake or Fire Punch to hit Steel types. It is possible to run Dragon Claw over Outrage if you don't want to be locked, but it comes at a significant damage output cost. The Speed EVs can be tailored to your necessity about outspeeding something you consider threatening, but 252 is a good compromise as Multiscale takes care of the survivability anyway.


Salamence @ Salamencite
Nature: Jolly
Ability: Intimidate -> Aerilate
EVs: 252 Speed / 252 Attack
Moves:
- Dragon Dance
- Roost
- Return
- Substitute
One of the staples of Singles facilities, Salamence is similar to Dragonite in concept, but provides additional Intimidate support and base Speed at the cost of coverage. Salamence is generally used as mono attacker, and you generally look to set up Salamence in a position where it can beat the entire enemy team on its own. Compared to Dragonite, 252 Speed is recommended in order to outspeed some significant threats at +1 that otherwise would be missed without full Speed investment. It runs mono attack in Return (or Frustration, if you prefer). The job of Salamence is to abuse slower weak phisical attackers or status users in order to stack up Dragon Dances behind the safety of Substitute, until it's ready to decimate the enemy with Aerilate boosted attacks.


Scizor @ Scizorite
Nature: Adamant
Ability: Technician -> Technician
EVs: 252 HP / 252 Attack / 4 Speed
Moves:
- Sword Dance
- Roost
- Bullet Punch
- Substitute / X-Scissor / Bug Bite
A bulkier and harder to KO Pokemon, relying on priority and faster setupping via Sword Dance, while also being immune to Sandstorm and Toxic which are common enough to be a threat. The set rounds up with a Bug STAB (Bug Bite gets Technician boost and removes enemy berries but is slightly less powerful than X-Scissor) or Substitute. Once more, it is possible to invest into Speed to outspeed opposer Scizor and some other priority users, but generally unneeded compared to simply having more bulk. Note that enemy Bruxish and Tsareena might block your Bullet Punch with one of their abilities if you are not using a second attack, but are weak to Bug-type attacks.


Glalie @ Leftovers
Nature: Timid
Ability: Moody
EVs: 136 or more Speed / 172 or more HP / rest in Special Attack
Moves:
- Protect
- Substitute
- Taunt
- Frost Breath
What would looks like a usually terrible Pokemon has a very important niche as setup sweeper on Tree singles, and that is provided by its ability Moody. Provided that it can stay on a neutered lead, Glalie can just alternate Substitute, Protect and Taunt in order to accumulate boosts while preventing the AI from statusing or setupping back. The moveset used for Glalie for the Battle Tree is Timid natured with 136 or more speed EVs, which allows it to outspeed a significant part of the tree up to minor scarfers at +1 Speed, with 172 HP EVs which give a full HP Glalie 16 turns to stay behind Substitute and Protect against a faster strong opponent while recovering with Leftovers and fish for the necessary boosts before it runs out of HP to Substitute again. The remaining EVs are invested into Special Attack to provide better damage rolls. The only attack used is Frost Breath, and that's because, being it a guaranteed critical hit, it can safely ignore any defensive boost that the AI might have set up, hopefully having also got some accuracy boosts from Moody as well. Note that Smeargle can tecnically do a very similar job if preferred albeith requiring more boosts to sweep, but having the option to use Spiky Shield, Toxic Shield or Spore instead of Taunt.


Bulky backlines:

Suicune @ Leftovers
Nature: Bold
Ability: Pressure
EVs: 252 HP / 252 Defense
Moves:
- Calm Mind
- Scald
- Rest
- Icy Wind
Another staple of the Singles, Suicune with its ability Pressure, great defensive type and stats is one of the premiere stallers for this mode. It also merges very well with Salamence Scizor and Dragonite, as it resists their 4x weaknesses while also being able to automatically unthraw from potential freezes with Scald. The Nature and EVs (expecially Speed) can be personalized based on what your team is made by, but otherwise a generic Bold with 252 HP and Defense EV is good enough to stomach most non super effective phisical hits. The typical set will be running Calm Mind, Scald and Rest with Leftovers, looking to setup several Calm Mind boosts on weaker special attackers or statusers while at same time depleting their PPs and healing up with Rest. The last move slot is usually dedicated to Icy Wind, allowing Suicune to reduce the speed of something that the lead might have issues setting up on, allowing them to sweep when Suicune dies or can swap out. Scald can also provide clutch Burns but is not something that should be really relied on.


Tapu Fini @ Choice Scarf
Nature: Modest
Ability: Misty Surge
EVs: 196 or more Speed / 252 Special Attack / rest in HP
Moves:
- Moonblast
- Surf
- Ice Beam
- Grass Knot
Tapu Fini can fill a somewhat comparable role to Suicune as switch in for Ice attacks, while also neutering any Dragon attack from faster AI threats, but due to Misty Terrain it has no access to a reliable recovery so cannot perform the stalling job Suicune does. On other hand, it can successfully use its STAB attacks combined with full Special Attack investment and its natural bulk to absorb hits and retaliate while providing a safety cushion against status for who comes after it, allowing safer setting up. Tapu Fini is generally paired with Aegislash as Aegislash appreciates the Misty terrain since King Shield doesn't block status moves. For this purpose, with a Modest nature, enough Speed EVs to outspeed almost the entire unscarfed tree (196) and the rest in HP, equipped with a Choice Scarf, Fini can use respectably strong Moonblast and Surf in order to weaken or flat out KO a threatening enemy for the lead or a backline that the lead cannot harm. The last 2 moves are used for any coverage the team could benefit from, but Ice Beam and Grass Knot are usually the best as they hit a few 4x weaknesses like many opposer Dragon types and Water/Ground types OHKOing or close to OHKOing them right away. Remind however that if you are also using a Dragon-type, Misty Terrain will also weaken your own Dragon attacks by 50% against grounded targets while it is active once Tapu Fini is sent in.


Gliscor @ Toxic Orb
Nature: Careful
Ability: Poison Heal
EVs: 252 HP / 252 Special Defense
Moves:
- Earthquake
- Toxic
- Protect
- Substitute
Gliscor is a Pokemon oriented to stall out enemies that can't KO it, depleting the PPs of their important moves or wasting Trick Room or Weather turns: its typing provides him with 2 complete immunities (as well as being immune to the threatening and common Thunder Wave), several resists and only 2 weaknesses, 2x to Water and 4x to Ice, and it has a notably strong phisical bulk on top of a unusually high Speed for a defensive Pokemon. Gliscor runs a very stall oriented set with the combination of Poison Heal + Toxic Orb (also providing him with what is basically status immunity after one turn), and then running Protect, Substitute, Toxic and Earthquake. Gliscor's 4x weakness to Ice attracts the Ice attacks that Salamence and Dragonite hate so much, but the higher speed tier usually allows him to Keep substitutes going alternating with Protect, while Poison Heal allows to keep healing up at same time. Earthquake and Toxic round up the moveset allowing him to both deal with several Electric-types which threatens Water-types like Suicune or Fini, or out-stall enemies that it can successfully wall and might still be threatening to the rest of the composition even after their Ice attack PPs have been depleted.


Aegislash @ Leftovers / Weakness Policy / Ghostium-Z
Nature: Brave (possibly with 0 speed IV)
Ability: Stance Change
EVs: 252 HP / 252 Attack
Moves:
- King Shield
- Sword Dance
- Shadow Sneak
- Sacred Sword
Aegislash is another Pokemon who can at same time work as staller and pseudo-setupper, due to the combination of an amazing ability, typing and the moveset to go with it. Since Aegislash runs Leftovers, it usually is never paired with another Leftover user, however it can run a Z crystal or even Weakness Policy to abuse the solid bulk it has. The typical set for singles will be running a Brave nature with as close to 0 Speed IVs as possible in order to always underspeed the enemy, as the last thing you are looking for is being hit in Sword form. 252 Attack EVs combined with usually 252 in HP are excellent generic EVs to use. Toghether with King Shield which is necessary to swap between forms, Shadow Sneak as main stab allows to pick up a clutch KO on a Sturdy/Sash user that your lead might have fallen to, and Sacred Sword as main coverage, while not being boosted by STAB, is able to ignore defensive boosts and also provide perfect coverage with the Ghost attack, making Aegislash an excellent answer to AI sets that like to boost their Defense. King Shield also allows to get very important Attack drops on the enemy, which can further enhance a Substitute user or Aegislash himself to safe set up later on.


Chansey @ Eviolite
Nature: Timid
Ability: Natural Cure
EVs: 252 Speed / 252 Defense
Moves:
- Soft Boiled
- Seismic Toss
- Substitute
- Minimize
Chansey is usually known in competitive Pokemon scenarios for its insane special bulk, and that is as well true when it comes to facilities: it can shake off large majority of the special attackers that don't carry strong Fighting attacks, while PP stalling them with Soft Boiled or slowly chipping them with Seismic Toss. A Timid nature with full or near full Speed invesment is usually chosen in order for Chansey to be able to set Substitute on status users, with the remaining 252 EVs in Defense in order to patch the very mediocre phisical bulk, and allow Chansey to take some phisical hits as well. Natural Cure as ability allows Chansey to swap out of Status if necessary, and Eviolite is the obvious item choice. Stacking Minimize on weak enemies can effectively allow Chansey to either beat a team on its own due to the dodges combined with Substitute, or at very least deplete threatening attacks' PP while recovering with Soft Boiled. Due to the nearly unexistant offenses of Chansey, Seismic Toss is really the only attack that can do actual damage to enemies, being a flat 50 HP damage per use.


Sample cores:

Mega-Salamence + Aegislash + Chansey
The three Pokemon benefit from being able to swap into each other immunities, and while Salamence's Intimidate and Aegislash's King Shield can progressively neuter phisical attackers, Chansey can instead take care of special ones. The core relies on providing either Salamence or Aegislash an opportunity to set up safely and then proceed to sweep, and Aegislash and Chansey provide a safety cushion if a significant series of bad RNG forces Salamence out.

Mega-Salamence/Dragonite + Suicune + Gliscor
Suicune provides the safety swap in against Ice attacks, and Suicune and Gliscor complement their stalling abilities toward special and phisical offenders in order to provide the opportunity of setting up Salamence. Suicune having access to its own setup and Speed control allows it to act as backup when Salamence cannot have the option of getting up Dragon Dances without risking to be taken out by RNG.


- Doubles
Doubles present a completely different situation from Singles. Due to the presence of plenty of spread moves, the threat of Taunt and Roar/Whirlwind, the way more threatening Trick Room compositions, and the possibility of being doubletargetted causing way more One Turn KOs than in Singles, setting up is not really a viable option, as more often than not something will negate the setup or just KO your sweeper(s) before they can do anything. With 2 Pokemon active per team, it is safer to rely on conditions that benefit your entire team comp or most of it, such as Weather, Terrains, Trick Room itself, assuming you got a safe way to guarantee their presence. Abilities that set Weather, and Tapus who get to set Terrains by just being sent in are very common in every successful composition, and for those who run Trick Room or Tailwind, they almost always side the setter with a Fake Out user in order to negate 1 enemy Pokemon and reduce the potential OHKO turn 1 possibilities to a bare minimum. With this in mind, always consider that there's also AI trainers that run Weather and Trick Room based compositions, so make sure that despite any type of synergy you decide to run, you also have a backup plan in case the enemy manages to set up their own condition. Being it stallers, priority, something that can use their own condition against them, make sure the teams you are running is not centered on one condition and falls apart the moment that one is negated.
A special mention goes to two moves and an ability that are significantly stronger in Doubles and should always be considered:
- Intimidate: with this ability affecting both enemy Pokemon at same time and the option to keep swapping your Intimidate in and out, it's one of the best abilities to run in your composition. However, be mindful that the enemy also can have it (in fact certain trainers have access to multiple Intimidaters), that Defiant, Contrary and Competitive exist and might need to be dealt with, and that critical hits ignore the Attack penality, so don't get overconfident on a enemy that you Intimidated multiple times and completely ignore it.
- Protect: the AI loves to double target in Doubles. Protect should be run if not on all Pokemon, on nearly all of them, simply because a properly timed Protect can essentially buy yourself a free turn for your other active Pokemon to get a KO, do chip damage, stall out a condition, etc. At same time, also always take note if a enemy Pokemon has access to Protect, because you do not want to get yourself in the same position, where you double target into a Protect giving the enemy a free turn for his other Pokemon.
- Wide Guard: when coming to Doubles, the AI actually often priorizes spread moves, and it has been proved consistently that the AI does not recognize or act around Wide Guard. Since Wide Guard does not get the success penality for repeated uses like Protect, you can often completely negate one of the enemies by just using Wide Guard several times, expecially as several post-40 AI sets only have spread moves (Surf/Blizzard and Earthquake/RockSlide coverage are both very common). Obviously also in this case be mindful of the few AI sets that do run Wide Guard.

- Multis
There is not much practice or investigation done on Multis, but with the fact that both Trainers in play in each team only run 2 Pokemon, it is simple to realize that the most efficent strategy is to nuke down both the Pokemon of a single Trainer first, in order to turn the match in what is essentially a 2 vs 1. If playing with a human partner, it's best to just run 4 very strong sweepers who have synergy (Weather and Terrain based is generally the most reliable) and focus down whatever enemy does not run Protect. Similar guidelines to Doubles apply, on not relying completely on a condition to succeed, but less so as once the match is turned into 2 trainers vs 1, it becomes very hard to lose. If you are using a scouted AI partner, try to scout a Trainer that is using 2 Pokemon with 4 offensive moves, because else there's a chance they will waste turns setting up or using status giving the opponent a potential free KO on them. Once you find a reliable AI trainer to go with, pick 2 Pokemon that complement it well enough and aren't susceptible to be hit by its spread moves (for example, if the AI has Garchomp with Earthquake, make sure you have Flying-types or Levitaters), or that buff them (Mega Charizard Y for example works well with fire type based Trainers).



ULTRASUN/ULTRAMOON DIFFERENCES

The second set of titles did not bring a lot of changes to the Tree: most of it is identical.
The following changes are present, on top of the availability of Tutors for BPs and the other Mega Stones at the vendor.
- Normal Battles do not have a Banned Pokemon list anymore, and do not delevel your Pokemon, meaning you can run it down with your level 100 Ubers if you so please. (Super Battles are still identical, though).
- A few accidentall oversights in AI sets got corrected, like some Sunny Day sets now having Heat Rock instead of Smooth Rock, or Incineroar-3 having a Fire move added to use Z-Firium with.
- Electivire-3 is now a Physical based set rather than Special
- A few minor changes in some movesets (entire list available here: https://pastebin.com/jt9TQEdP)
- All sets that previously had less than 4 moves except Serperior-4 now have 4 moves.
- Lilie is now available as default AI partner (She has Ribombee-2 and Comfey-3, mainly supportive sets)
- A new extra trainer can show up every 10 battle: Kukui.
- BPs are now also obtainable via Mantine Surf and can be used with the various Move Tutors at the beaches as well as buying items.
- One of the four Move Tutor is in the Battle Tree area, while the others are at the Surfing spots of the other 3 islands. They teach the following moves:
Snore (4 BP), Heal Bell (8 BP), Electroweb (8 BP), Defog (8 BP), Low Kick (8 BP), Uproar (8 BP), Bind (4 BP), Helping Hand (8 BP), Shock Wave (4 BP), Block (8 BP), Last Resort (12 BP), Worry Seed (8 BP), Covet (4 BP), Bug Bite (4 BP), Snatch (8 BP), Recycle (8 BP)

Iron Tail (8 BP), Spite (8 BP), After You (8 BP), Giga Drain (8 BP), Synthesis (8 BP), Ally Switch (8 BP), Signal Beam (8 BP), Gravity (8 BP), Stealth Rock (8 BP), Iron Defense (8 BP), Telekinesis (8 BP), Magnet Rise (8 BP), Bounce (8 BP), Role Play (8 BP), Fire Punch (8 BP), Water Pulse (4 BP)

Iron Head (8 BP), Aqua Tail (12 BP), Pain Split (8 BP), Tailwind (8 BP), Thunder Punch (8 BP), Endeavor (16 BP), Focus Punch (16 BP), Icy Wind (12 BP), Zen Headbutt (8 BP), Seed Bomb (12 BP), Laser Focus (12 BP), Trick (8 BP), Drill Run (8 BP), Magic Coat (8 BP), Ice Punch (8 BP), Wonder Room (8 BP)

Liquidation (16 BP), Gastro Acid (8 BP), Foul Play (12 BP), Super Fang (12 BP), Outrage (16 BP), Sky Attack (16 BP), Throat Chop (16 BP), Stomping Tantrum (12 BP), Skill Swap (8 BP), Earth Power (12 BP), Gunk Shot (16 BP), Dual Chop (12 BP), Drain Punch (8 BP), Heat Wave (12 BP), Hyper Voice (12 BP), Superpower (16 BP), Knock Off (12 BP) and Dragon Pulse (12 BP).

- There's now a Bewear to play with and scare you by trying to hug you in the Battle Tree area.
 
Last edited:

shiny finder

forever searching
is a Community Leaderis a Programmeris a Smogon Discord Contributoris a Site Content Manager Alumnusis a Top Social Media Contributor Alumnusis a Senior Staff Member Alumnusis a Community Contributor Alumnusis a Contributor Alumnus
Chatot Wrangler
Wow, very thorough! I don't really dabble in the Battle Tree (so others can QC this better than me), but after reading through your guide I have a few suggestions: Perhaps you could add an example team(s) that showcases the tips you describe. You do mention the Aegis/Chansey/Mence core for singles and the immunities they have, so perhaps you could expand on/showcase that. In other words, an "in-depth" example that would give readers a starting point for designing their own teams for the formats. You also mention "Florges-2", so a link to where I can find out what that means could be a nice addition.

As a general question, is there something I should look for when Scouting an AI partner? Is it just synergy with my own team, or do some AIs really play well with exploiting all of those AI mechanics you mention? Any I should avoid? It's probably a hard question to answer, and maybe not even answerable, but is something that might be nice to add if you have any insight on that.

Lastly, in terms of things I've seen in similar guides, you could add in the auto-level mechanics (i.e. below 50 stay that way), and the BP you can earn in the streaks.

As someone who's not an expert on the Battle Tree, those are just some things that I would be interesting in seeing/knowing. Apologies if I overlooked those points, and nice work!
 
Wow, very thorough! I don't really dabble in the Battle Tree (so others can QC this better than me), but after reading through your guide I have a few suggestions: Perhaps you could add an example team(s) that showcases the tips you describe. You do mention the Aegis/Chansey/Mence core for singles and the immunities they have, so perhaps you could expand on/showcase that. In other words, an "in-depth" example that would give readers a starting point for designing their own teams for the formats. You also mention "Florges-2", so a link to where I can find out what that means could be a nice addition.

As a general question, is there something I should look for when Scouting an AI partner? Is it just synergy with my own team, or do some AIs really play well with exploiting all of those AI mechanics you mention? Any I should avoid? It's probably a hard question to answer, and maybe not even answerable, but is something that might be nice to add if you have any insight on that.

Lastly, in terms of things I've seen in similar guides, you could add in the auto-level mechanics (i.e. below 50 stay that way), and the BP you can earn in the streaks.
I was thinking of adding a "novice proof teams" part somewhere, mentioning a handful of singles core and typical doubles setups going over the specifics of the synergy used, but I was not sure if it would be appropriate for a generic guide/walkthrough.
If you think it's a good idea, i'll quickly gather some more informations on other good beginnerfriendly combos from our friendly nerds dedicated runners and dedicate a section to it in the Basic Tips chapter, together with some pointers to what sort AI partners to look recruit.

As for the "Florges-2" and similar mentions: this sort of naming are generally used in regard of the order in which the set appears in the specific battle facility.
However, the resources we use with this collection are player-made excel sheets or external sites, I am not sure if it's fine to point out said resources on smogon guide. The creators don't mind, as I had already asked when starting the guide, but I believe Jellicent mentioned that linking to external sites was not necessarly a good plan.
Specifically, the resources we generally use are all at the bottom of the first post of our dedicated thread (expecially the trainer data sheet and the calculator made by turskain):
http://www.smogon.com/forums/threads/battle-tree-discussion-and-records.3587215/
So if any of those would be fine to show on the guide, I'd be happy to oblige.

And for the last part, definitely easy to add to the first chapter, in fact i'll do that right now.
 

atsync

Where the "intelligence" of TRAINERS is put to the test!
is a Pokemon Researcheris a Contributor to Smogon
The IVs for the AI are all at 11 for battles 1 to 20, 21 for battles 21 to 40, and 31 onwards from there, for any AI Pokemon generated.
This isn't correct.

I tested this a bit a while ago and even retested the IVs after reading your guide. From mock battles with Punk Guy Evander (battle 1 of the streak in super singles), I found that his Muk1 speed-tied with my level 23 Alakazam with 57 Speed, indicating IVs=18/19 rather than 11. It's probably 19 based on how it was in the Maison.

I posted a bit about IVs a few months ago here.
 
This isn't correct.

I tested this a bit a while ago and even retested the IVs after reading your guide. From mock battles with Punk Guy Evander (battle 1 of the streak in super singles), I found that his Muk1 speed-tied with my level 23 Alakazam with 57 Speed, indicating IVs=18/19 rather than 11. It's probably 19 based on how it was in the Maison.

I posted a bit about IVs a few months ago here.
Thank you for the clarification, i'll add this quickly.
And while i'm at it i'll let the others know :)


I rephrased that paragraph as

The IVs for the AI vary from the specific trainer, progressively increasing as you progress through the facility. While it is not perfectly omogeneous, the enemy IVs will start at 19 in Super Battles, then trainers will eventually hit 23, 27 and finally 31 IV once you reach the sets that appear only past battle 40.

I don't think there's much point in explaining the whole "set 1 to 50 have X, set 51 to Y have Z" unless actually going to link a database of the trainers. Plus probably excessive data for a walkthrough.
 
Last edited:

shiny finder

forever searching
is a Community Leaderis a Programmeris a Smogon Discord Contributoris a Site Content Manager Alumnusis a Top Social Media Contributor Alumnusis a Senior Staff Member Alumnusis a Community Contributor Alumnusis a Contributor Alumnus
Chatot Wrangler
I was thinking of adding a "novice proof teams" part somewhere, mentioning a handful of singles core and typical doubles setups going over the specifics of the synergy used, but I was not sure if it would be appropriate for a generic guide/walkthrough.
If you think it's a good idea, i'll quickly gather some more informations on other good beginnerfriendly combos from our friendly nerds dedicated runners and dedicate a section to it in the Basic Tips chapter, together with some pointers to what sort AI partners to look recruit.

As for the "Florges-2" and similar mentions: this sort of naming are generally used in regard of the order in which the set appears in the specific battle facility.
However, the resources we use with this collection are player-made excel sheets or external sites, I am not sure if it's fine to point out said resources on smogon guide. The creators don't mind, as I had already asked when starting the guide, but I believe Jellicent mentioned that linking to external sites was not necessarly a good plan.
Specifically, the resources we generally use are all at the bottom of the first post of our dedicated thread (expecially the trainer data sheet and the calculator made by turskain):
http://www.smogon.com/forums/threads/battle-tree-discussion-and-records.3587215/
So if any of those would be fine to show on the guide, I'd be happy to oblige.

And for the last part, definitely easy to add to the first chapter, in fact i'll do that right now.
Thanks for taking what I said into consideration :) Yeah, I think a practical example or two (1 singles, 1 doubles perhaps) would really help cement what you're talking about. It doesn't have to be super detailed (i.e. to the point where you're describing play-by-play strategies); anything you think really highlights those tips you gave (Having a backup plan or synergy for example). I don't think you really need to even include a handful of cores (though please do if you feel you must); a good example or 2 would suffice.

Our guides occasionally have links to other sites/resources (but I can see why that should be avoided if at all possible), so if you're unsure I'd clarify with him. With that said, it seems like you could just link to our dedicated thread since it has those resources, which might be a good idea anyway. It would provide a way for people to discuss, seek resources, or just get involved in the forums (which is always a good thing!). You could even throw in our thread as a sort of conclusion to your guide.
 

Lumari

empty spaces
is a Site Content Manageris a Top Social Media Contributoris a Member of Senior Staffis a Community Contributoris a Top Contributoris a Top Smogon Media Contributoris an Administrator Alumnus
TFP Leader
I haven't actually played the Tree but I played the Maison when that was a thing / co-wrote the guide on that one, so I assume I'm somewhat qualified to comment on the non-Tree-specific stuff here / double check stuff compared to Maison mechanics; just gonna comb through this and highlight some stuff that stood out to me, feel free to take into account / edit whatever you wish.
In similar fashion to other "first set of titles" of previous gens, there's no Tutors available for BPs, but several evolutionary items, held combat items and also every Megastone for Pokemon present in Alolan Dex are available for buying.
Would like to see some tables / listings of all the items and their prices so we can include them in the guide after upload.
However if you Scout a trainer, he will always have the abilities it had at the moment when used in Multi battle.
You mention this a few times, but at least at this point in the guide I have no idea what it is, please clarify.
The only "cheat" (if so we can call it) that has been consistently proved itself true is that the AI knows when one of its moves will 1HKO your Pokemon and will generally go for it, but that's as far as it gets.
Would rephrase this as "knowing your Pokemon's effective stats" (meaning raw stats + boosts from Choice items etc), since it also knows when it'll outspeed / underspeed them.
Most AI sets that are based on setup commonly tend to use the setup move 2 to 3 times before going on offensive.
Dragon Dance and Quiver Dance users expecially seemb to default to attempt to priorize reaching +2 Atk/Spatk unless they can pickup a 1hko before it, and even then they generally still try to hit the +2 (which means, when Intimidated, they will often DDance 3 times). Curse users also look coded to attempt to Curse 2-3 times in similar fashion.
In the Maison, this used to be:
- they'll boost if they have a Speed-boosting move and don't outspeed your Pokemon;
- they'll boost if they have a power-boosting move and don't OHKO your Pokemon;
- they'll attack if have a guaranteed OHKO and either outspeed your Pokemon or don't have a Speed-boosting move.

With indeed some scattered cases where weak Pokemon that already had a number of boosts would go for the 2HKO instead of the OHKO, but the above was pretty definite for a lower number of boosts. Don't know if any proper research has been done on Stockpile / other defensive boosts, but at least for offensive boosting moves, "stop boosting if you both can OHKO and, if applicable, outspeed" was pretty universal in the Maison. I don't know if any further research was ever done on "they can potentially use offensive boosting moves like DD and QD until all affected stats are maxed out, but they'll always stop using defensive boosting moves like CM and Curse after six uses even if some of the affected stats were lowered during setup" but it's probably not worth mentioning regardless so whatever.
Tailwind nearly always is set anytime is not up (occasionally even while Trick Room is active) and Icy Wind is generally spammed by Pokemon who have it, expecially in Doubles where the AI loves spread moves. Bulldoze also tends to be priorized, often even at the cost of the partner's speed in Doubles.
In the Maison, this only applied if the user was slower than your Pokemon; if it was faster it'd just treat it like any low-power attacking move. If this is still the case specify pls.
Thunder Wave, Nuzzle and Hypnosis are very high on the priority list for the AI, expecially Paralysis since it both acts as status and speed control. If a set runs those, you can be expecting it will be attempting to status hax you first unless your active Pokemon is immune to it.
Not really true in the Maison since it was still subject to the AI prioritising guaranteed OHKOs (e.g. I was never really worried about Politoed4 using Hypnosis over Focus Blast when I was using Greninja or Weavile in Doubles).
When you beat an enemy Pokemon or force a swap and the AI has still multiple to send in, the AI will priorize sending in the Pokemon with the highest damage potential.
If this is still the case, specify that this is based solely on the move's Base Power + type effectiveness, not the Pokemon's actual offensive stats; see e.g. Carbink4 with Explosion often getting sent out second even though an actual strong Pokemon was still in the back.
A special mention for No Guard: the ability looks good on paper, until you realize the AI also runs OHKO moves in some sets, plus several "full inaccurate moves" sets, meaning that your Machamp or Mega-Pidgeot might actually have their own ability turned against them. If you are planning your team comp around a No-Guard Pokemon, make sure to consider the downside of not being able to evade.
Are No Guard users actually good? In the Maison, when they surfaced it was usually just as ""low-ladder"" ideas ("missing sucks -> No Guard means no misses -> let's run No Guard", without considering the Pokemon in question as a complete package); I vaguely remember Wide Guard Machamp surfacing once as a glue mon on a reasonable Doubles team, but overall point stands, I don't know if the No Guard users are any good in the Tree but if they're not then a special mention for No Guard means you're effectively encouraging the use of unviable Pokemon and I'd rather see it axed.
- Answer to stallers
Might make sense to expressly separate this into a singles and doubles section, because in the Maison the ways to deal with them in singles and doubles were vastly different; in doubles the main counterplay was just "kill them quickly", adding a Pokemon immune to passive damage was often a good idea too but Steel-types were already near mandatory for separate reasons, if you could toss a random Taunt on one of your Pokemon that worked too but it was by no means something to build an entire team around, and that was basically it; whereas in singles it was often feasible and optimal to just PP stall evasion users and similar sets entirely, not sure how viable that still is in the Tree though with the massive powercreep.

Similarly, idk how to feel about some of the moves you mentioned, since (as always going off Maison experience here) Aerial Ace and cousins were just actively bad (too weak to be run as main attacking moves, not enough moveslots available to toss on as filler), similar deal with Clear Smog (which I've only seen run viably on Doubles Gastrodon and does not strike me as a singles move at all, neither does Haze); just make sure you don't end up encouraging the use of unviable stuff in your efforts to be comprehensive.

Another typically abused and a bit more advanced suicide lead is Scarfed Durant, with the plan of using Entrainment to put Truant onto the enemy, allowing your sweeper to alternate setup and protect, and then just clean up in same fashion.
Elaborate a bit on Durant teams so that people don't end up running stuff like Durant + BP Blaziken or expect to make 1k with teams like idk Durant / Garchomp / Aegislash.

Also, about the moveset / trainer lists since Jellicent discussed that with me too, I do think we should avoid linking to offsite content since we'd have no control over it staying up / we'd be "advertising" other people's sites and, in a way, stealing their content, but I don't have an issue with sheets / listings maintained by Smogon users, i.e., turskain's sheet is fine to link until we can re-host it here (which will be a lot of very tedious work so not on my immediate to-do list). Iffier on the calculator since it's just an edited version of Honko's + it's basically redundant with the sheets, but I'd like to see the sheets linked in some way.

Some other stuff that you could consider adding is a Pokemon threatlist (reckon there are a couple more foes to be highlighted than in previous gens now that the AI runs both Mega Charizards and lol Swords Dance Mega Mawile), sample teams at the very least too (individual Pokemon writeups like we did in the Maison guide could be extreme) but I see some other people have already suggested as much, and a resource section where you could link the sheet + the discussion thread.
 

Lumari

empty spaces
is a Site Content Manageris a Top Social Media Contributoris a Member of Senior Staffis a Community Contributoris a Top Contributoris a Top Smogon Media Contributoris an Administrator Alumnus
TFP Leader
also tagging some dudes from the thread, if you guys wish to read over this and point out whatever edits need doing that would be wonderful NoCheese turskain ReptoAbysmal GG Unit Level 51 Josh C.

if there's anyone else that you guys feel would have valuable input feel free to tag them too because I haven't been keeping up with the thread other than a few cursory looks and don't know of many "new" strong contributors :(
 
Hikers Vivek and Stellan are both listed in the Sand specialists and, while they clearly show prefernce for rock/ground/steel types, take special note that they include most of the Trick Room users available, which don't fit that theme at all; I'd include them in the list of TR users for that reason.

I also hesitate to include "swapping out for no reason" in the list of unpredictable behaviors since the anecdotes people provide don't really bolster the claim. AI decisions in doubles are made on the basis where the poke on the player's right makes its decision first, and the poke on the left, opposite the player's lead, makes its decision second; the previous turn if applicable influences those decision but, most importantly, the AI selects for each poke independently from the other. People who don't really acknowledge this fact can easily misinterpret something like switching as baseless. The closest thing people have found to a truly random, pointless AI decision, by myself and others, is when the AI targets its ally with Will-O-Wisp, and the ally has no attack or ability that is even affected by burns.

On the subject of switching, I didn't see that the AI does tend to switch targets into predicted NVE attacks, which it didn't do throughout Gen VI (where it switched in order to nullify or absorb the previous attack.) Now, the AI will switch fairy types into dark moves, for instance. You may find it relevant enough to include.

As for using NVE or suboptimal moves, I thought we'd established during the Maison days that the AI did this when no OHKOs were readily available, leading instead to random attempts to inflict damage. I would elaborate on the point further than simply chalking it up to "sometimes the AI does random things" especially when there are people who switch into said randomness and immediately believe the AI read their command and reacted to it. I'll concede if I'm just in the minority on this mindset now.

I assume Worldie has already selected a dedicated proofreader, because many things jump out at me which need polishing. Great job all the same.
 
Last edited:
Well building a "threat list" is kind of a issue, since there's a insanely big variation of "what's a threat and what's not" depending on the specific team.

ReptoAbysmal I don't miss to point every change I make to the topic in the Discord, so others can point me possible flaws. I believe there's general agreement on the points, outside of possibility of better wording / clarity obviously, but that's what the QC is for anyway :)
However, I have not really witnessed intentional swaps to absorb attacks or attempt to get NVE effects, at least, not in doubles. I need to inquiry a bit GG over this since I believe he's the only "high streaker" in singles I can readily contact.

I'll get to work on the extras, either ways, and improve the wording.
Will specifically see if I can rack up a mention of the "threatening sets" that isnt 3 pages long.
 
Yeah you do see switches to a bulky resist from time to time in addition to the more prevalent instances of the AI switching in something with an ability that nullifies your previous move used (should also note that this can happen even if you have a different Pokémon on the field than you did the turn before). I don't think it's consistent enough to have any takeaway besides "if you have multiple equally accurate moves that can KO something, use one different than the turn before because there may be a switch."

For a concise way to discuss threatening sets in Singles, I think you can say you need either something that hits hard and fast (which in most cases is going to come from a Mega or Z move) or a crippler in the lead position due to the diverse array of naturally bulky stat boosters (Zard X, Mega Gyarados, Tyranitar). Then a lot of seemingly annoying defensive sets can be dealt with simply by using Substitute or Rest, which have much more utility than some random coverage move that isn't going to OHKO bulkier threats even if they're weak to it (e.g. you can use Thunderbolt but it's not gonna prevent Mega Gyarados from DDing on you or Walrein from firing off a OHKO move or two).

That last part kinda ties into the whole accuracy vs. power thing. Most sweepers designed for competitive play rely on prior residual damage and/or inaccurate moves to pick up KOs that just aren't going to be possible with the 100% accurate variants. Combine that with Focus Sash/Sturdy, which are common enough that running into 2 or 3 such opponents in a battle isn't really 'hax,' and you're not going to get very far with glass cannon sweepers. Defensively, you're up against opponents that are free to gamble with the powerful, inaccurate moves, so simply having a 'check' to a set or having a Pokemon that resists a particular attacking type is often not going to be enough to win 100+ battles in a row (crit Head Smash still does a number on Rock resists, something can resist Ice but still get frozen/hit by Sheer Cold, etc).
 
Last edited:
I've read a good chunk of the guide so far, Worldie has also run a couple ideas by us in the Discord server as well; looking pretty solid but thanks for tagging to remind me to come back to this. I'll read over it again at some point and see if anything could be added or elaborated on.
 
While gathering the data for the threat list, I have done a bit of experiment and confronted with others, and it does appear that the apparently unpredictable swaps generally have the AI attempt to swap a pokemon in a NVE or immuned move.

I have reworded the paragraph as it follows:

Another event that, expecially in Battle Tree, has caused often confusion and/or frustration from the player is the AI doing apparently unpredictable things. In the long run, you will occasionally witness weird behaviours like double or triple Protects, swapping a Pokemon out for no real reason, using a resisted or even immuned move instead of going for the neutral or supereffective alternative, killing his own partner with spread moves, and the list goes on.
Every move selection from the AI is on a priority list, but no move actually ever has "100% chance" to be executed. When not able to OHKO, the AI could simply choose one of the moves available to it, not necessarly the highest damage or the best target like a human would.
Furthermore, possible incomprehensible swaps can come since the AI can attempt to swap out a Pokemon that just got hit with another one in his team which can resist or be immune to the move just used.
Does this look better?
 
There are other instances which will often create confusion or frustration for the player, those being the semblance of apparently random behavior by the AI. As your streak continues, you will occasionally if not inevitably witness such odd behavior as consecutive uses of Protect, needlessly swapping Pokemon, using resisted or nullified moves while clearly superior alternatives exist, and so on. Typically, the AI prioritizes moves based on power and effect. When able to inflict a KO, its choices are easily predicted; however, without an imminent KO, they cannot be taken for granted. It's these instances you'll notice the AI selecting suboptimal moves which merely inflict any kind of damage or status, unlike the choices a person might make. In doubles, they may also select an unexpected target.

Swapping is generally done when a benched pokemon can resist or nullify a specific attack which hit the target the previous turn. While this was often done in earlier generations, it was exclusively for type immunity through absorption or otherwise, or to circumvent move-locking. Now, the AI may switch merely to resist the attack which hit it. This includes being hit by its own teammate's spread move.
Please don't take this the wrong way, but your paragraph was so jarring to read that I felt like working on it. No doubt, the QC fellows will come up with something much different.
 
Last edited:
I'm not absolutely taking it the wrong way, worry not, I am happy to receive feedback from others. I do indeed like your wording much better.

Differently from most, I like team work ;)
 
That's a relief! I rewrote my response at least four different times in attempts to sound as non judgmental as possible. I'm happy to help in any way I can.
 
Did replace the paragraph with your wording, and changed the name to "Apparently unpredictable AI behaviour" which fits it better since we do have a realistic explanation to most of it.

I was thinking of moving the 2nd paragraph to a separate part down into the Ai tendencies, but since the "swap into a NVE or immunity" is not consistent (the AI might or might not do it, doesn't seemb to have a particular priority over simply attacking) I don't think it should be separated on its own.

Still unsure on what can trigger the AI to double/triple protect or use status moves into immunities (though the latter might just be the AI having issue recognizing certain sudo-immunities like Misty Terrain, Comatose, Shields Up, etc.)
 
I've carefully gone through the entire database, and identified what I believe are the pokemon worth of mentioning as "threatening sets".

Now, asking specifically to the big runners Josh C. GG Unit ReptoAbysmal Smuckem , which of the following archethypes are worth including in the guilde, and did I miss any relevant set?

Second, most of the 1 and 2 sets disappear after battle 30 and usually can just be bruteforced anyway. Are they actually worth including in said list?

Here's a very crude list:

FLINCH HAX
Threat: can flinch you several times in a row preventing any action. King Rock users often commit to Fake Out > Fling > potentially Thief or U-Turn away
Aerodactly-1 (Scarf rock slide)
Ambipom-1 (fakeout flinch)
Braviary-1 (fast king rock)
Greninja-3 (fast king rock)
Infernape-1 (fakeout flinch)
Togekiss-3 (Para+flinch)
Weavile-1 (fakeout flinch)


STATUS HAX
Threat: Generally run Paralysis, Sleep and/or Confusion, and often spam it until both of your Pokemon are afflicted.
Suggested counterplay: status immunity, Lum Berry, or kill first.
Aggron-3
Ampharos-13
Audino-1
Aurorus-1
Azelf-2
Crobat-4
Dragonite-1
Druddigon-1
Electrode-1
Ferrothorn-1
Gengar-1 (fast hypnosis)
Klinklang-1
Lapras-1
Magmortar-1
Mesprit-2
Milotic-1
Mimikyu-1
Musharna-1
Nidoqueen-1
A-Ninetales-1
Ninetales-12
Porygon2-4
Probopass-12
Raichu-1
Regice-4
Regirock-3
Rotom-Fan-23
Rotom-Frost-3
Rotom-Heat-23
Rotom-Mow-4
Rotom-Wash-23
Sableye-1
Shiinotic-234
Togedemaru-14
Umbreon-13
Vespiquen-23


METAL BURST/REVERSAL/COUNTER
Threat: often running Focus Sash or Sturdy in the set, will generally retaliate your attack with a massive counterattack for double damage, or in the case of Endeavour users, with Quick Attack the turn after.
Suggested counterplay: do not commit big massive hits for 1hko, chip first in order to get the KO with no retaliation.
Aggron-14
Bastiodon-4
Bisharp-4
Cobalion-2
Durant-1
Gliscor-4
Mawile-2
Rhyperior-3 (also has Horn Drill)
Sableye-4
Sandslash-A-1
Sawk-2
Sceptile-1 (Endeavour+Quick Attack)
Snorlax-1
Staraptor-1 (Endeavour+Quick Attack)


STALLERS
Threat: will attempt to draw the match by using defensive moves, while letting your Pokemon die to chip damage, status, or simply not be able to hit due to accuracy drop or evasion stacks.
Suggested counterplay: Taunt, status immunity, guaranteed hit moves, Z-moves, focus down first.
Amoonguss-1
Araquanid-13
Articuno-1 (Hail/Fly)
Avalugg-13
Blissey-1
Breelom-4
Carbink-2
Cradily-34
Cresselia-23
Dhelmise-4
Dusknoir-3
Garchomp-1
Gastrodon-3
Glaceon-4 (A-veil)
Gogoat-3
Lilligant-234
Ludicolo-13
Mandibuzz-14
Milotic-24
Mimikyu-23
A-Muk-1
Muk-1
Pelipper-2
Politoed-1
Regigigas-1234
Rotom-Wash-1
Sandslash-1
Serperior-3
Shuckle-34
Spiritomb-4 (Sub+PainSplit)
Tauros-4
Tentacruel-4
Throh-1
Thundurus-1
Torterra-4
Toxapex-4
Trevenant-2
Unfeazant-4
Vaporeon-23
Venusaur-1
Vespiquen-1
Walrein-1
Weezing-4
Whimsicott-3
Zapdos-2


SETUPPERS
Threat: will generally commit to use a strong setup move 2-3 times, then potentially sweep if left alone for too many turns.
Suggested counterplay: Encore, Taunt, focus down first
Armaldo-1
Avalugg-2
Barbaracle-1 (Bright Powder)
Carracosta-1
Charizard-4 (Mega X)
Cobalion-3 (Swagger/Psychup)
Conkeldurr-3
Feraligatr-24
Ferrothorn-4
Gastrodon-4 (Amnesia+Curse)
Gothitelle-4 (Flatter/Psychup)
Gyarados-34(4=Mega)
Hariyama-4 (Belly Drum + priority)
Magmortar-2 (Belly Drum + priority)
Minior-4
Oricorio-Sensu
Scrafty-13 (Sub+Focus Punch)
Serperior-4 (Specs+only Leaf storm)
Snorlax-2 (Belly Drum)
Spiritomb-3 (Swagger+Psychup+Priority)
Suicune-13
Swampert-4
Tyranitar-23
Volcarona-34


OHKO
Threat: access to 1hko moves
suggested counterplay: Substitute, Sashes, Sturdy, Focus down immediately
Articuno-2
Banette-3 (Mega Prankster Destiny Bond)
Dugtrio-2 (Fissure/Toxic/Sandstorm/Sub, Brightpowder)
Froslass-4 (Sash+Destiny Bond)
Glalie-3
Landorus-1
Mamoswine-3
Mismagius-4 (Perish Song + Mean Look)
Pinsir-4 (scarf guillotine)
Tyrantrum-3
Walrein-4 (sheer cold, fissure, sleeptalk, bright powder)


FAST HIGH DAMAGE
Threat: very fast Pokemon with access to very high damage, potentially able to sweep on their own if your composition has no answers
Suggested counterplay: access to strong priority, reliable Trick room or crippling, stalling.
Aerodactly-4 (Band)
Beedrill-34 (Mega)
Charizard-3 (Mega Y)
Entei-3 (Scarf eruption)
Garchomp-3 (Scarf)
Gengar-4 (Mega, Destiny Bond)
Golisopod-4 (triple priority, Life Orb)
Metagross-4 (mega)
Salazzle-4 (Z-overheat)
Slacking-3 (Band)
Terrakion-2 (Scarf)
Thyplosion-3 (Scarf eruption)
 
I really don't consider myself a "big runner" by any means, but thanks for valuing my insight.

Honestly, enough is unchanged since ORAS aside from some loss of tutored moves that I don't think the threat list has changed too much from The Dutch Plumberjack article. It was short and concise.

That said, I would gloss over the addition of Megas and some Z-move users, particularly mentioning pokemon such as Mawile and Charizard who will always be Mega past battle 30 if another ally hasn't already activated a stone.

Also, in order to allow the article to apply to USUM, perhaps you could curtail writing for a bit, to see what kinds of changes the tutors bring to the enemy sets. Something may make a startling jump in potency, but it takes a lot to become another Walrein4.
 

Lumari

empty spaces
is a Site Content Manageris a Top Social Media Contributoris a Member of Senior Staffis a Community Contributoris a Top Contributoris a Top Smogon Media Contributoris an Administrator Alumnus
TFP Leader
Yeah, for total clarity, I never intended for a threatlist to be comprehensive, just a couple broad categories like OHKO move users / evasion users, with one particularly threatening example for each. When I wrote it for the Maison guide, the intent was never to cover everything, just to give a general idea of a couple particularly scary foes you'd be facing.
 
What about mentioning a couple of the "big offenders" of each category, mentioning the entire pokemon set (moveset, stat distribution, item) and both what the threat is and how/why to play around it?
 

Users Who Are Viewing This Thread (Users: 1, Guests: 0)

Top