it was a recent change that had 3/4 required votes to overturn it in the council phase. If its winrate drops, we'll revert it. Pretty simple.
the palkia suggestion would also result in never getting both draco and rend ever
-
Welcome to Smogon! Take a moment to read the Introduction to Smogon for a run-down on everything Smogon, and make sure you take some time to read the global rules.
-
Congrats to the winners of the 2023 Smog Awards!
Format Discussion Scarlet/Violet Random Battle Sets
- Thread starter A Cake Wearing A Hat
- Start date
would it be better if grafaiai could perhaps get protect on its move set? Idk I feel it would be good in stalling out turns better. not too sure though
It already does.
Last edited:
Doubt you'll be able to observe anything significant since it's already a perfectly balanced pokémon, and it will be very hard to isolate this change from other meta changes that may affect Barraskewda's performance, and from usual ±0.25-0.50% variance that seems to happen every month. I want to believe the onus for justifying a change after a tryout phase should be that it's shown to improve the win rate, but, when a pokemon's win rate is already slightly above 50% in randbats... it doesn't really make sense, I guess (since that would mean it becomes unbalanced, and thus, require a level drop, and then we get into that vicious circle ad eternam).
One justification that I'm supposing, upon thinking about it, is how getting a flinch might allow, in some cases, a 2HKO without being revenged kill. I also forgot about hazards last night (was late!), which will make waterfall 2HKO with similar odds to liquidation in several cases.
I assume that Barraskewda having such poor bulk with high speed (thus making it more likely to benefit from a flinch) is also why this change isn't being tried on other slower or bulkier liquidation users.
However, if what I mentioned is the justification, and it's seen to improve or cause no change (thus keeping the change as I understand), I feel like Arceus-Fire would become another pokemon for which Waterfall over Liquidation could be justified. Mostly because in Arceus-Fire's case, you will use liquidation mostly against Pokemon that have STAB rock or ground type moves.
That's unfortunate. Although upon looking at the coverage thingy I was using, I noticed a lot of the Pokemon Thunderbolt/Thunder would be useful against aren't in SV yet, and that those that are in don't exactly threaten Palkia in any meaningful way (Gyarados, Pelipper, Golduck, Vaporeon) or will be 2hko'd/3hko'd by both rend and t-bolt.
We did this mostly because, while revamping Gen 7 (which we're doing), we realized "oh we have waterfall on mega sharpedo for the flinches right why did we stop doing that". So we're seeing if we stopped doing that because it's bad.
Also, changes in STAB move can definitely strongly impact a Pokemon, even if it's just a small difference and there aren't many stats for it. Calyrex (horseless) had Psychic removed (in favor of more Psyshock) a bit ago, and in just a month we noticed an almost 2% winrate drop for that, despite it only appearing 1/3 of the time compared to other pokemon. I trust that we'll be able to sniff out any effect on Barraskewda over enough time.
Makes sense I guess since Barra and M-Sharpedo are kinda similar as fast wallbreakers.
That said I wouldn't compare STAB moves that target different defense stats (physical vs special) as the opponent faced matters there (unlike, say, liquidation vs waterfall). I'm hoping we see something conclusive in waterfall vs liquidation's case even though I won't hold my breath.
oh I guess I never really saw it
Regieleki is still fairly useful with tera electric but it's a massive massive downgrade from tera ice. If we could calculate it, I'm fairly confident winrate difference between the two would be vast. Tera electric eleki is way underleveled because it's balanced based on eleki's (presumably high) winrate, which is almost certainly that high because of tera ice. it cannot be stated how big of a deal bolt beam coverage is for this thing.
You are limited to one Tera Blast user per team, because Tera Blast is usually a dead slot otherwise. While Regieleki is kind of the exception to this, because its movepool is so bad that 80BP Normal coverage is actually OK, we can’t make an exception for it to the wider Tera Blast coding without breaking stuff. Most Tera Blast users hate having a wasted moveslot on Tera Blast because another Tera Blast user has a better matchup so you have to Tera them instead. And just imagine getting a team of 6 Tera Blast users…
So this is sadly unavoidable. All Tera Blast users have to have another role and moveset in case they generate after a Tera Blast user. If they didn’t then the game would crash.
I guess there's no way to have the computer decide what mon needs tera blast the most. well i appreciate the response
i think you should give camperupt the chance to roll will o wisp over yawn since he only has 4 moves in his pool and i think having the chance to burn would be cool, if it's possible i think you could also give it fire blast over lava plume if will o is an option
Ran into a Shadow Tag Goth today, but it was level 88 IIRC. If Shadow Tag is only legal in Ubers, why did Goth still have a level like it was low-tier? I mean obviously it's not like Goth needs any more help being shit but I'm curious where the distinction between higher tiers and lower tiers for the sake of leveling is demarcated for bans of moves/abilities. I don't remember what the case is/was with Last Respects and Moody.
Levels in random battles are determined by level balancing via winrates, not by tiers. Pokemon with low winrate get their level increased and vice versa. Gothitelle started at level 82 in the format, but it was very bad at that level and continually got buffed. Even now at level 88, it still tends to be below average, e.g. its winrate was 49.05% in July.
Last Respects users were very strong at higher levels, so they have been significantly nerfed (now, Houndstone is 73, Basculegion is 68, Basculegion-F is 70). Scovillain/Glalie were also very low level compared to their tiering when Moody still existed, though they are much higher level now that Moody has been removed. Seems like Shadow Tag Gothitelle just isn't that good in randbats.
Check out the link, which explains level balancing in greater detail, if you are interested.
Last edited:
This again just shows how broken the team generator is. a curse setup, unaware mon in the lead against a FULL physical team. oh, did u think veluza did a crit once? naah, not in 10 turns. was an easy surrender. the second screen is self explaned - 3 screeners.. while grimmsnarl COULD have the bulk up set or i believe carbinx has a leftis /AW set but im not sure on that. this is completely ridicoulus. and yesterday i had a nasty plot, dark pulse, flash cannon, focus Miss set on lucario. why would a setup mon (with tera fighting) a LOW accu move as its main source of dmg, when it literally has auro sphere?? thats why it has nasty plot, to make up for the lack of dmg. oh and lucario was my only special attacker that game, missed to focus misses and got killed. rest of the team was again physical amd got walled by enemy bulk up sweepers again. "yea ThAtS juSt ThE gAmE, iTs hAx" - no, these factors u can control in random should be controllable f.e. change low acc moves to the flamethrower, auro sphere etc. i mean u already do it to all thunderbolt mons, no one has thunder but its the same case. ill gather more screenshots to show how broken this meta is but until then, fix these. ty
Attachments
Oh, thx for the bright insights you gave with your answer. i understand that my message is written with a grain of salt but i think everyone understands human emotions and can still give a normal answer, since didnt only write full of tilt but with actual shown prove in form of screens and extra explenation. so i kindly ask again for a review and a comment on the following topics:
problem: generating full physical / special teams without 0 counter to enemy set up mons.
solution: having a regulation that at least 1 special/physical move has to generate in a team.
problem: team generating THREE screeners.
solutions: i read somewhere in this thread, that there is actually a check for this in your code, so it should generate another set of a mon if there is one with same purpose. maybe check on this code again cause this happens quite a lot. (edit: there is also still a anti hazard generating problem)
problem: lucario special move set / setup sweeper.
item: lifeorb, tera: fighting, moves: nasty plot, dark pulse, flash cannon, focus miss.
solution: a setup sweeper needs to rely on safe dmg, or the whole setup is obsolete. aura sphere instead of focus miss is at least 51% better to a max of 99% and if there are cases where a risky focus miss is better, then the aura sphere moveset still has more positives than negative aspects.
problem: low accu moves bring more, unneccesary randomness to a ranked, competitive mode
solution: yes, its a random mode. but there are skill factors a player can control and show skill in form of decission making and move selection. example: player 1 has a strong sweeper mon as its last. player 2 has only 1 mon to counter it from the start, but he has to rely on focus blast / hurricane / fireblast etc to hit. he misses and he loses. did player 1 deserve the win even though he made the right decission throughout the whole game where he has a 3v1 scenario? he should be rewarded with a win at this point, not rely on hax to win when he was better the whole game.
and there is a saying from the hearthstone game, that a big pro player said: randomness favors the weaker, less skilled player more than the higher skilled.. just putting that out there.
thx for considering, reading and i kindly ask for a reasonable answer on these topics.
Last edited:
You're really, really overestimating the damage output of lucario. Focus blast does more damage than aura sphere on average even when you take the misses into consideration. And even then: 1st, you already have a 100% acc. STAB with flash cannon; 2nd: with aura sphere, you WILL miss on a lot of necessary KOes you need to land immediately given how frail lucario is, especially when its lack of bulk is made worse by the life orb. I'd rather OHKO 70% of the time, than never. Having a 180BP nuke on a set up set, on which you already have several 100% acc move, is way stronger than you seem to think. Only running low BP moves on a sweeper that doesn't have great offensive stats, even with a life orb, means you will not be able to threaten a lot of the bulkier mons in the format, which mons, however, will threaten to KO you back
Oh and without mentionning that special Lucario does have aura sphere on its movepool in gen9 rands. It's a 50/50 split between the two moves
i can totally see ur argument on this and understand it, but you have given my argument more points for that. special lucario in it self is a weaker, gimmicky set, just straight fact. but why make 2 weak gimmicky sets? just leave it with the aura set and the 2 physical (orb or band). there is no reason for adding double negative rng factor to it. and you cant hit everything neutral with dark pulse and flash cannon. and for lack of dmg, yes, thats why it has nasty plot. still not amazing but it deals enough dmg. focus miss isbjust a new layer of randomness the game doesnt need
lucario has 110atk and 115 spA, being mixed is also a benefit since it might throw your opponent in the wrong gameplan. The simple existence of focus blast makes the set way more threatening too, since not many mons are able to take 1 focus blast at +2, let alone 2. The list of things that can check the special aura sphere set that can't check the focus blast one is notable, even at neutral. Bottomline: focus blast is a necessity and the relatively meh accuracy is completely ok given the absolute sheer difference in power. 60BP is really, really, really huge
a mixed set would also be interesting, yes.
but where i maybe put too much faith in aura sphere, there you also put too much in focus blast. it has a reason its called focus miss. all that dmg is useless if you dont hit. for example, what does more damage over 2 turns after a nasty plot: aura + aura or (high luck) 2 focus blast (winner focus blast), 1 focus blast, cause 1 miss (winner aura s.) or both focus blast miss (ofc, aura wins). there is literal math behind it that statistically, you deal more dmg with aura and you rather deal safe dmg once before fainting than a chance to never hit and lose the game. cause u might have a prio move that is strong enough to finish the job (not uncommon).
no, you deal more damage in average with focus blast, actually. Aura sphere is only 80BP, not 90BP. That's also why we always run tbolt over thunder, because 90 to 110 BP is small enough of a gap that tbolt does more in average, but not necessarily aura sphere over focus blast, because, once again, you'd rather OHKO than never KO more often than not.
This has been discussed previously and rejected. Full physical/special teams are rare, and whether something is a physical or special attacker (or mixed?) is not something that is tracked by the code.
Set generation tries to prevent multiple screeners by removing screens from the pool if the team already has screens and the set's movepool has 6 or more moves. This does not apply to Grimmsnarl and Carbink's screens sets since they only contain 5 moves. We discussed what to do in that situation, and decided that duplicate dual screens were superior to having a single screen.
Similarly, you can get double rocks if a set with only 4 moves, like Camerupt, is generated after the first rocker. There's nothing we can do about it.
Random Battles operates on the principle of optimal sets (principle 3). Sometimes, this means using lower accuracy moves, because they are overall better than their weaker, more accurate counterpart. Fire Blast is on average a better move than Flamethrower, so most Pokemon that can get both opt to use Fire Blast. This is not always the case; for example, Chi-Yu's Fast Attacker (choiced) set uses Flamethrower because it already has a stronger Fire move in Overheat and values Flamethrower's accuracy more.
Odds management is itself a skill and part of the game. In general, we will not deliberately make sets worse in order to reduce RNG.
Last edited:
thx for ur detailed answer! so for problem 1 its like: "haha, you got haxed, deal with the auto loss."
for problem 2: debatable but again, it generated 3 screeners, not just 2 ^^ it could have given at least hypno a descent, semi offensive build with leftis or AW to be a strong special wall. it has a smaller move pool but i have seen worse troll sets on random battles.
problem 3: it has a ranked ladder, its supposed to be a competitive enviroment. you are completely right with your argument if we consider "normal" battles where people build their teams. but again, this is a random mode. this mode stacks randomness, on randomness, on hax, on low accu moves and on more randomness. its in no way competitive but it trys so hard to be. why else did many high ranked players leave when gen 9 random b. dropped. it was a complete clown fiesta and ofc it got better over time but it still lacks a lot to be as good as it could be and im sure the people behind it try their best. tldr, when you add so many layers of randomness but you have things a player cab control, then consider giving the player at least the option - like a checkmark box (do you want to generate mons with low accu moves or the safer option) - just a thought but you know what i mean.
Hypno generated first. Team generation cannot go backwards, on a functional level. Therefore, you can get three screeners if you get Hypno and then Grimmsnarl and then Carbink with Hypno first. We cannot under any circumstances fix this.
We will not remove focus blast from the format. Stop asking.
Essentially yeah, but it's more a code limitation than anything else. Getting e.g. 5 special attackers and a Chansey with Seismic Toss (physical!) isn't gonna help you against the opposing Blissey.
This is a miniscule chance, but mons generate in order. Hypno generated first, so when Carbink and Grimmsnarl generated after Hypno, they tried to have their screens removed but couldn't due to only having 5 moves. If Carbink had generated first, Hypno would have indeed not had any screens moves. There is no retrospective checking during generation to avoid endless code cycling, and server strain. This is not something that can be changed.
First of all, as someone who's involved in the tours scene a lot, you don't do yourself a favour by saying statements like "why else did many high ranked players leave because of gen 9 rands" when that... did not happen. Gen 9 is by far the biggest and most played format in showdown's history, with the most active ladder, and a remarkably higher activity ladder than gen 8 had. Gen 9 is an objectively popular format based on activity, and also an objectively high skill format based on ladder analysis. If high ranked players did leave, they weren't prominent enough for it to have been noticeable, which means they weren't really high ranked.
Aside from that, randbats are made technocratically, and via statistics. We track how any changes we make in set design impact that Pokémon's winrate, and if changes are positive or negative they stay or are reverted respectively. We don't need to overhaul not just the code but the actual inner functioning of Pokémon Showdown! (which would never get past the core Showdown! team, you're asking for the home page of the site to be changed here) in order to give players the option to have movesets that are statistically proven worse than what they would otherwise get. It doesn't make sense.
We don't just throw these inaccurate moves on mons for no reason, and try hard to ensure that wherever possible Pokémon have at least one reliable STAB move. But a format of solely 100% accurate moves isn't really functional.