Reporting a new record today of
121 wins in Battle Tower Multi. Another long post inbound!
Intro
Multi Mode is my favourite mode of battling, no question. I love the unpredictable nature of it, and the fact that it’s much more of a challenge than other modes because it often feels more like 2v1; if you get a bad partner, you’re screwed and you essentially have to carry the can yourself. Because of this, a lot of people don’t have the patience for it and recorded streaks with an NPC partner I’ve seen on Smogon rarely get much further than 100 wins. The true key to success in this mode comes from playing around your AI partner’s piss-poor tactics and (more often than not) doing twice the work yourself.
It took me two attempts to get to 100 wins, but my second was a lot smoother as I’ll detail below. (As a side note, I was expecting to get the Gold Shield for this, but it turns out Gen 3 is prejudiced and that’s only for Singles).
Basics
As mentioned, the AI plays quite poorly (and seemingly much more so when it’s on your side). However, there’s a reasoning to the decisions it makes. If you’ve read the
article on the Battle Subway and in particular the section about AI tendencies, much of it still holds true here. Knowing how the AI operates is essential because if you can reliably intuit how your partner is likely to behave you can work more or less in sync with them to get the most optimum results.
So, the key points to keep in mind:
-The AI will go for a kill when it can. This is the most basic point. If there is a Pokemon with low health they are capable of KOing, then nine times out of ten they will act accordingly.
-In a double battle, the AI generally tries to avoid damaging their partner, but won't hesitate to do so if it would result in a KO against the enemy (even if it also KOs the partner). This includes Explosion and Selfdestruct: I’ve witnessed a Metagross blow up and take down its partner Ursaring while failing to KO either of my side’s Pokemon.
-Conversely, the AI is aware of basic strategies relating to abilities/immunities that play in the partner’s favour. So they will use Toxic on a partner with Guts (even if the partner has very low health), use Teeter Dance if the partner has Own Tempo, or use fire attacks on a partner with Flash Fire.
-The Subway article mentions that the AI seems to prefer using moves that hit both opponents, such as Surf, even if it’s not necessarily the most effective move. I haven’t always found this to be the case. There have been instances where my partner could have used Surf for a double super-effective hit (such as against a Tyranitar with low health and its partner Regirock) but opted to use a similar-powered move that would KO Tyranitar, such as Thunderbolt. Interesting.
-The AI seems to prioritise the usage of status moves. I’m not sure if this is just specific to your partner. From my experience, it seems that whenever I chose a partner who had any status moves they’d just spam them ineffectively. This is especially important with Counter/Mirror Coat: both allies and opponents will often use them even if they have no prospect of surviving a move.
My Team
Since I knew I needed a powerful team that could rapidly KO lots of opponents, I went with a tried-and-tested combination.
Tauros@Choice Band (Intimidate)
-Double-edge
-Earthquake
-
-
EVs: 252 Atk / 252 Speed / 6 HP (Adamant)
Because I’m only running two moves, it’s important to have partners who can handle the two Pokemon I can’t – Misdreavus and Gengar. This is also why Tauros always goes up first; sending it out last is begging to be Banded into a useless move. Hidden Power Ghost would have been nice, but I’m not an expert at RNG so I let it go. Besides, it’s too specific; even most Steel-types usually take a reasonable chunk of damage from Double-Edge. Double-Edge is favoured here over Return: the recoil can sting, but the increased base power scores KOs on various bulkier Pokemon Return often doesn’t, or leaves them with a shred of health so that my partner will be baited into picking them off.
Tauros is probably the best lead in this format. Intimidate is a great ability, and allows Tauros to survive most physical attacks; partnering it with another Intimidate user makes any physical attackers the opponent leads with a total joke: Gyarados and Salamence are greatly favoured when I can get them. Machamp and Hariyama can still occasionally be difficult to take down, but they’re much less threatening with -1 or -2 Attack.
Latios@Lum Berry (Levitate)
-Psychic
-Ice Beam
-Thunderbolt
-Dragon Claw
EVs: 252 Sp.Atk / 252 Speed / 6 HP (Modest)
I debated using Latias, but Latios’ extra power won out in the end. There isn’t much of a better choice for an item: Lum Berry has better utility than Leftovers in Doubles. I initially ran Calm Mind, but multi is not the format to boost; it’s too fast, and most opponents Latios struggles with can be 2KO’d anyway, or simply piled on with my partner.
The objective with my team is always to get at least one KO on turn 1. If this is accomplished the match becomes very hard to lose. If both foes present a challenge to my side, I’ll generally try to KO both Pokemon on one side so it’s 2v1 in my favour. Note that Gen 3 switching mechanics means that it is possible to KO both side’s Pokemon in one turn (if you’re lucky and/or predict well, of course).
Partner Criteria
So here’s the most important bit. Choosing the right partner in the Battle Salon is highly important. Up until around the 35th battle I would typically partner with an apprentice; if you manage to create one, early on they’ll be a decent partner due to having fully-evolved Pokemon – after about 4 rounds, however, they fall behind the NPCs in terms of strength and should be ignored. I have two; a Beauty with Mawile, Lanturn, and Ninetales (Earthquake bait, the lot of them) and a Bug Catcher with Shedinja, Dustox, and Pinsir. All apprentice Pokemon have 85 EVs in each stat and a random nature, and do not usually have great movesets.
After 49 or 56 wins (I’m not sure which), the NPCs start bringing legendaries. This is when it gets really interesting. Generally when picking a partner I’ll go for someone with a strong special attacker as their lead, preferably immune to Earthquake. Zapdos, Aerodactyl, Gengar, Dragonite, Latias, Flygon, Skarmory, Gyarados, and Salamence all make good partners for Tauros (Gyara and Salamence especially for the sweet double-Intimidate): Moltres and Articuno are somewhat more risky but doable in a pinch. For the partner’s backup slot, strong physical attackers are useful to complement Latios, but another powerful special attacker is good too. If a potential partner has two Levitators, it’s practically a done deal. Anything weak to Earthquake is an immediate no-no. In fact, anything that takes damage at all from Earthquake is pretty much undesirable, with some exceptions like Starmie, Suicune, and Regice.
If I can’t find a strong offensive partner (which happens fairly regularly) then I’ll go with someone with bulky Pokemon that can draw fire from my guys and act as a meatshield whilst I try and wreck the opponent’s team. Suicune is particularly excellent at this, since it’s so bulky it can soak up a ton of hits and generally get at least one kill.
However, solely defensive partners are a BIG no-no. Doubles is not the format for defensive play and chances are your partner won’t achieve anything useful before being KO’d. Obviously, if both of my Pokemon faint, the match becomes 2v1 against the partner; in that instance, I generally just turn my game off because the chances of them winning alone are close to zero. A team with anything less than 8 offensive moves is sub-par; this can make choosing someone with any of the legendary birds dangerous because of their abysmal movesets. I used a spreadsheet of Frontier Pokemon to help determine what my prospective partners were carrying but this wasn’t always useful because many trainers in the later rounds carry multiple versions of the same Pokemon (being told your partner has a Dragonite with Earthquake, for instance, doesn’t really help).
Picking a good partner means that you’ll generally have an easier time of things, but even with the best partner possible they’ll still trip you up a lot of the time. While it’s dangerous to assume you can predict what your partner will do, keeping the AI tendencies I mentioned earlier in mind means that it’s possible to at least make an informed guess. All the usual standards of CPU behaviour seem to apply for your partner, though I have often wondered if there’s extra coding in place to make your partner play like an idiot. Obviously, if your partner has a Starmie with Ice Beam up against a Flygon and Altaria, the odds of them playing badly are much less, but it’s still dangerous to assume they’ll take care of things for you.
Even with the relatively good spread of partners available once I hit 70 wins, it’s still not guaranteed that you’ll get a good partner, and several times I found myself in a room full of NPCs with EQ-weak leads. To that end, I mixed records with myself so that I’d always have at least one solid option.
School Kid Orienne
Latias@Lum Berry (Levitate)
Psychic
Ice Beam
Thunderbolt
Surf
252 SA/Speed, 6 HP (Modest)
Aerodactyl@Choice Band (Pressure)
Double-edge
Rock Slide
Earthquake
Aerial Ace
252 Att/Speed, 6 HP (Adamant)
“If it ain’t broke…” was the thinking here. While this team is solid, there’s still some variance. Ideally Latias will be in position 1, as this combos well with Tauros; Aerodactyl tends to avoid using Earthquake when partnered with Tauros and this can sometimes be a hindrance. Additionally, since Latias and Tauros have the same Speed stat I have to keep this in mind when attacking, but this hasn’t cost me a match yet.
Also, it’s amusing that the game tries very ineptly to pretend that they both have the same original Trainer when they don’t (note the different Trainer IDs in the attached pics).
Threats
Quick Claw users are still majorly threatening, particularly those with OHKO moves or Selfdestruct/Explosion; random Focus Band activations are also grating. I’ve learned to avoid attacking Swampert and Blastoise at full health due to the likelihood of them carrying Counter.
Pokemon that boost can be hard to stop if they manage to get going, particularly bulky titans like Snorlax, Metagross, and Registeel. Registeel has an annoying tendency to come out second when Tauros has already selected Double-Edge. I largely try to avoid switching, so it’s rare that I can pivot Tauros back in; I have to rely on my partners ganging up once everything else is dead. There's very little you can do to stop boosters short of taking them down as quickly as possible: simply knocking out all of the other Pokemon on the field and ganging up on the remainer works well, too.
Losses
My first loss
My first loss at 108 wins (this was prior to mixing records) would have been totally avoidable if I’d just played more carefully. I’d teamed up with Cooltrainer Gillian, who was using Zapdos (Thunderbolt, Drill Peck, Substitute, Thunder Wave) and Dragonite (Double-Edge, Aerial Ace, Earthquake, Dragon Dance) which I’d been pretty impressed with so far.
Battle starts against Xatu and Forretress. Tauros took down Xatu with Double-Edge, Zapdos used Thunderbolt on the Mr Mime that came out to replace it, taking it to about 60%. Forretress used Double-Edge on Tauros and got it down to around 45%. Next turn, Tauros dropped Mr Mime, Zapdos Thunder Waved Forretress, and Forretress Double-Edged Tauros again and killed it. Latias comes out, and Forretress uses Explosion to kill Latias and Zapdos, getting a critical hit on the latter. One Pokemon left on each side, but the opponent’s last Pokemon… was Glalie. Carrie’s Dragonite, of course, had no chance. Not my finest hour at all, and I’m gutted to have screwed up what was looking like a good streak.
Second loss
Was partnered with Orienne and went up against Tyranitar and Swampert. Tyranitar got a Quick Claw’d critical hit Crunch on Tauros, dropping it, while Latias used Ice Beam on Tyranitar and Swampert Curse’d. I had a feeling it was already over at this point, but I duly sent out Latios and used Dragon Claw to get Tyranitar to about 30%. Latias used Psychic on Swampert for about 50%; it Cursed again, but Tyranitar used Crunch on Latios again for about 90%. Figuring I may as well get in one kill, I finished off Swampert with Psychic; Steelix came out to replace it and got frozen by Latias’ Ice Beam. Tyranitar promptly KO’d Latios and Latias in turn. Aerodactyl EQ’d for a double KO, but the PokeManiac I was facing sent out… Meganium, which got a boost off the Ancientpower it dropped Aero with (adding insult to injury). Can’t fault the AI here, it played about as well as I would have. Still, I’m quite proud of this run and I think I’ll be trying again in due course.
If you read all that, thanks and I hope you enjoyed it. Any feedback/comments welcome!
