Yeah, that's an unfortunate bug, but it makes sense - at least there's an
attempted thought process behind the bad decisions there, even if it's an erroneous one.
The most irritating AI in the series for me is
definitely the one in Sword and Shield, which very frequently bases its decisions on
total RNG (outside of the Tower, I know - that's always an intentional step up and does not count).
It just turns every turn into a dice roll - even if you know they have a coverage move, they could go
forever without actually using it. Since most of my team in the main campaign was made of notoriously weak Pokémon, especially from the first to fourth Gyms or so
and I was further limiting my own options by trying to resist the Experience Share and catch up my lowest-leveled Pokémon, there were many times when I
should have lost, but too often, the AI used its best option just once and then immediately switched to spamming useless alternatives instead of finishing my Pokémon off. There were times when I thought a battle was hopeless and was just waiting for it to be over so I could prepare properly/start over/try again with a different strategy, and it would have been faster if I did... but instead of allowing me even that even passably engaging form of gameplay and decision making, I still won because the AI
stalled for so long instead of finishing me off, which was just tedious and boring.
Gym Leaders at least had decent AI, but nearly every random route Trainer (at least as far as the fourth Gym - dunno if they got better or if I just stopped paying attention) and even some important rival battles were very clearly nothing but dice rolls, and it was just so much less fun.
And I don't even just mean that Sword and Shield were "too easy" all around - several of the aforementioned Gym battles were actually pretty memorable and pleasantly surprising, and I think three of them even beat me on my first try.
In my defense, two of those losses were me trying to get away with switch training my weakest Pokémon instead of taking them seriously as a boss, but one of them actually did just catch me off guard, and I appreciated that!
This is a ramble I wrote to a friend of mine when I was analyzing SwSh during my first playthrough
yes, this is the kind of thing I do all the time... probably to no one's surprise, and I think it conveys how I felt while going through the early-game slog of Galar - I believe I had just finished Route 3 and was about to enter Galar Mine when I wrote this?
Early-Game Balance
"Okay, so the way the damage formula works, attacks don't just do more damage based on higher stats.
They also do more damage at higher levels - even if the Pokémon happen to have exactly the same stats despite the level difference.
This is meant to compensate for the way stats increase: like, if I'm level 35, my Attack is about 35% of what it would be at level 100, and if you're level 35, your Defense and HP are about 35% of what they'll be at level 100... but that means I'm doing 35% of the damage I'll eventually do, while your defenses are only around 12.25% of what they'll eventually be, because Defense and HP are effectively multiplied together.
That would mean a level 1 Pokémon using a move against another level 1 Pokémon would be (roughly) one hundred times as strong (1% over 1% squared) compared to a level 100 Pokémon using the same move against another level 100 Pokémon (1 over 1 squared).
To compensate for that, level is another factor in the damage formula; as a consequence... give or take some amount (I think level 100 Pokémon actually turn out to do more damage to each other than level 50 Pokémon? not sure exactly what causes that but I'm giving a highly simplified version here), a level 15 Pokémon fighting a level 15 enemy ought to be doing about the same amount of damage as a level 50 Pokémon fighting a level 50 enemy... with the same move.
The problem here-- a level 50 Pokémon fighting another level 50 Pokémon is going to have access to stronger moves than, say, Tackle and some 40 BP STAB that usually relies on its worse offensive stat.
For most (offensively oriented) competitive Pokémon, even a 70-80 BP move is below average for your primary attacking option - imagine how long battles would take if you had to settle for your early-game moveset instead!
And as mentioned, it's not quite evenly scaled - I think level 100s do slightly more to each other than level 50s, which also means level 15s do slightly less to each other than level 50s... on top of having to rely on these weaker moves. amendment: I'm not actually positive if I was correct on this part - someone let me know if this is wrong!
So battles get slower. This part is true in every game!
But... generally speaking, there are solutions in place:
- Moves like Leer and sometimes Howl are widely distributed among early-game Pokémon
- Some games give early-game Pokémon better moves than others (SM, as a pretty extreme example, gave Z-Moves to turn your average Normal attacks into a 100 BP move at a minimum to tear through a slow fight, starting when your own team was no stronger than level twelve... not that that was necessarily the best solution, but the option was there!)
- And in general, early-game Pokémon tend to be proportionally frailer than their evolutions - in-world, this is just intuitive because most of them are gaining physical size and naturally end up bulkier, but it serves a gameplay purpose, too!
In Sword and Shield:
- Growl is on everything; Leer is on almost nothing (at least that I've seen)... my Wooloo even started with the effectively redundant Growl and Defense Curl
- Sand Attack is also very common, further reducing average damage per turn
- I actually have one Pokémon with Hone Claws, which negates both of these... but she's a special attacker, and her only physical move worth using is a 40 BP Normal-type move without STAB, so it adds up to very little anyway
- Next to no Pokémon seem to have coverage moves this early (this is actually also an indirect problem with the starters all being monotype: in SM, for example, every starter got a decently useful move of a second type early on to reflect its eventual evolution, and then USUM went on to add a third!) another amendment: I realized later that this is also a consequence of the way TRs were handled - by locking all of the usual viable, interesting moves and good coverage options behind Raids and giving you only bottom-of-the-barrel moves as rewards for exploring in the actual campaign, Sword and Shield limited your move options pretty heavily, especially compared to other games from Gen V and later with reusable TMs...
- One of the first Pokémon you can encounter has Fluffy
- Skwovet, the single most common encounter, has 75 base HP and 55 Defense - compare Bunnelby, which has 38 for both
In past games, threatening coverage and strong STAB moves are used only by opponents who also have the AI to use them well, which actually makes it harder and not easier, but it also means a) if an enemy is able to take you out, the battle is over quickly and b) beating these enemies is a matter of working around those moves rather than praying they use the wrong ones (see Ilima in SM and USUM, as well as the Trainer's School teacher in USUM - they were specifically designed to make you give attention to your non-starter team members, and if you didn't, they would not hold back). Meanwhile, in SwSh, most of the NPCs seem like they're really inconsistent about using these even when they have them, which mostly just wastes your time and makes it both way easier to win when it shouldn't be and moderately harder to plan ahead in any meaningful way?
But yeah, the average battle with my Wooloo has her not at all threatened (she takes 50% of her health at most)... but they were also taking like five or six turns per Pokémon for a while, so it's just slow without being challenging. Which is a combination that doesn't work so well. XP"
But the most egregious part is that
every Max Raid boss uses pure RNG to select its moves. As if getting the right NPC partners wasn't luck-based enough (and their own AI is also pretty tragic...), the more "difficult" five star raids stack so many advantages that you can't possibly win consistently in single player, then "balance" it by choosing suboptimal moves 75% of the time and absolutely wrecking you (or, more accurately, ignoring you and wrecking the AI partners with no counterplay) the other 25%.
What the heck happened to the amazing boss design of Totem Pokémon? Those were asymmetrical fights just like this - they had advantages that you as a player could not possibly match, and they also had clear disadvantages relative to you as a player - but
they made the most of their mechanics, made the best decisions possible and actively made an effort to challenge you, and they were still perfectly fair and possible to beat consistently given the right approach. Meanwhile, Max Raid bosses are just downright
broken, and their core mechanics don't even attempt to give you a fair chance or leave room for any innovative strategies ("hey, wouldn't it be fun if we made it possible for them to nullify
every effect except raw damage even after players succeed in setting them up, and then we gave them temporary total immunities to all status moves on top of that, and then we made even strategic choices of damaging moves barely matter because most of your attacks are hitting barriers instead of their HP?") - if they were controlled by a competent AI, you'd pretty much
never be able to win. But the thing is that their moves are selected so randomly that they just throw away those advantages - instead of being properly balanced and giving players a chance the
right way, they once again have the issue of stalling for no reason and leaving openings completely at random so you can just pray your way to an unearned and unstrategic victory.
Let me put this in perspective... in a single-player five star raid battle, as many as
seven moves happen every turn (your team's one per Pokémon and the boss's three attacks), and you only control
one of them. That one move is
never going to be enough to outweigh the other random selections in impacting the outcome of the battle.
This is the feature that's supposed to appeal to players who want to be challenged and want opportunities to test their skill. A+ game design.
There's a point where the AI being totally RNG-based is more frustrating than actually losing a fight would be - compare Sun and Moon's bosses, which will beat you right away if they actually can (see the early-game boss Ilima's Smeargle, which isn't hard to beat at all if you've caught other Pokémon, but it
always takes advantage of its coverage moves if possible and punishes you if you try to solo him with just your starter). Even as an
extremely early-game boss, being beaten like that in no way unfair or unfun - that's a wake-up call or a mercy kill or whatever else you want to call it. It makes it clear that you're doing something wrong and makes it clear what you need to do to prepare, and then you can try again once you've figured it out. The game doesn't have to be
hard - it just has to make your decisions matter. It should react to them, punish them and reward them, not just ignore them the way Max Raid battles do.
OH RIGHT-- speaking of RNG! You know what the absolute worst part of Max Raids is? The fact that you only get to attempt
one Poké Ball, and it's
after you finish the battle. If Pokémon breaks free, then rather than just somehow making the battle harder, or giving it another turn to use a move or recover from sleep, or wasting your resources, or inhibiting you in any meaningful or interesting way... it just
ends the battle and makes it all for nothing.
I genuinely do not understand how this entire feature got to be so popular, but I just can't stand it. Every time I
try to get back into Shield - and it's really not for lack of wanting! some of the Wild Area events have motivated me to pick it up again if they have cool rewards or G-Maxes that I like - it just becomes so actively frustrating that I find myself resenting it and want to stop.
And it doesn't help that this and the completely gutted Battle Tower are the only forms of repeatable postgame content in the game... This is why I haven't picked up Shield in months.