Yeah, this is an interesting prospect. I think there is some advantage to be gained for good strats people already use like Weakness Policy sets that wait to Tera, especially for teams that aren’t using something like Dauntless Shield to pop Mirror Herbs early. I think there’s a lot of potential for Normalium Z in particular to utilize this, but rather awkwardly, the best users of Normalium Z are the ones least suited to waiting a few turns, and holders of Normalium Z can’t Terastallize. I think having one HO mon like Intrepid Sword Mega Heracross or Download Blacephalon and one support mon like Friend Guard/Power Spot Type: Null would be the best way to try that.So I was having a nice night getting ready to wind down, until I was nerdsniped by this thread claiming Metronome is the prime way to reverse engineer Showdown RNG seeds and predict every random outcome of a battle.
(Also before that, I found out Metronome is in a team tour which was also interesting and kind of led to finding this but wasn't really related directly.)
Notably there are some varied takes on how it would affect the Metronome Battle meta, from having no effect to breaking everything. Like most dichotomies, the true answer is neither yet lies somewhere between the two, though you can probably figure for yourself which side it falls more towards just by thinking about what Metronome gameplay is. But personally I wanted to see it for myself because it sounded neat.
So based on the instructions in the video, I played around for a while locally to get a feel for how the calls work, and you can see the fruits of my efforts with a proof of concept replay on main here kind of predicting stuff like this: https://replay.pokemonshowdown.com/gen9metronomebattle-2271400207
(Fun fact: There are 581 Metronome eligible moves as of the current end of Gen 9.)
View attachment 699643
View attachment 699648View attachment 699649View attachment 699650View attachment 699652View attachment 699653View attachment 699645View attachment 699654View attachment 699646
View attachment 699661
Also I just want to show Azumarill getting the clutch win in this demonstration, even though I didn't press tera to change the outcome at all.
Putting together my rambling thoughts overall:
It's a fun party trick when it works. I don't think it really affects much for this format in practice but in theory maybe it does something.
Firstly it takes 2 turns to get enough info to work with, which is sometimes enough for the match to be decided anyway. I specifically set up the optimal conditions of having no speed ties and every mon actually getting an action for this to work out too.
The added effects of speed ties on random calls are very undocumented, which could also be said for many other interactions in the doubles format as I discovered on the fly, but speed ties especially seem to add a ton of frame advancements everywhere, which I believe includes when Intrepid Sword procs, which causes many edge cases that I just didn't want to bother with accounting for. Also mons can change speeds mid match and you have to account for that too.
In this test battle I was lucky to have a very simple T1/T2 to calc calls off of simple attacks and buffs/debuffs. Immediately after on T3 spread moves kind of messed me up, and I didn't really know what RNs are tied to proc chances or other things that could affect future outcome orders (though target choice seems pretty clear).
Maybe if you put in a lot of time in you could figure more patterns out, but even I don't think it's worth it, because there's just not that much to do other than show off and mess with people on ladder, if you have the time and correctly get the calls from the first 2 turns right. Ultimately the outcome is still mostly decided from the start of the battle.
Technically you do have one possible influence on the battle of pressing tera or Normalium Z or mega evolution at one point, but to actually take advantage of it you need to give up on pressing that button until you know it can matter, which by turn 2 may be too late. So there's already an explicit tradeoff you make by going this route to affect the battle later, by giving up a turn 1/2 tera to not affect the battle earlier, and running Normalium Z/a mega stone arguably doesn't make up for the item slot on its own either.
You also need to be able to predict future calls accurately too and the amount of future future calls that will be affected which I also would not want to bother with. Ultimately I just feel like the potential return on the outcome of a single game is too low for the effort put in, compared to just playing more games in the same time. But if there is ever a point that someone does manage to actually take advantage of this information to influence the outcome of an arbitrary metronome battle in practice, then I would respect the effort at least.
Even if you have a bot simulate the battle from that seed, you also can't account for the enemy's exact set or when they will press whatever button they have unless they already did it, and if we ever reach the point of Metronome 20XX where everyone is perfectly predicting each other then neither player will have pressed their button anyway. This was actually kind of similar to a thought experiment I had in Gen 8 with dynamax once.
So as of now I would say the "counterplay" is basically just having any speed ties involved, or Imposter, or something like Quick Claw or King's Rock that adds a unrevealed random call until it procs to throw people off, or just not bothering at all.
tl;dr I want to sleep. But I also wanted to figure out what this all means in practice. In conclusion I think this is neat and funny but probably won't practically affect anything, feel free to prove me wrong though with another example or anything I'm missing. Hopefully now you have a more informed idea of what this all means for Metronome Battles, because I don't know where to start with anything else. Maybe Showdown will do something about it and render this all irrelevant anyway before it even matters again. Who knows? These are interesting times. Even just Metronome (the move) getting a spotlight for something is nice and noteworthy enough for me.
Also the smogon dex description seems to be up, looking good.
EDIT: Expanded on the possible point that you could have to give up the impact of an earlier turn 1/2 tera just to potentially get something out of a tera later if you want to have an informed impact on the battle's outcome with this exploit, which is pretty timely to bring up regarding the recent question.
If you were going to employ this RNG manipulation strategy, I think it you might as well program a bot to do it automatically, since Tera/Z timing is otherwise the only thing a bot doesn’t know when to do.
I’d never considered using this technique in Metronome battle, but I had thought about it before. The real formats to try this kind of thing in are formats with randomized team generation. If you’re player 1, then your team was generated first, so with six randomly chosen species of data, it seems pretty feasible to reverse engineer the battle’s seed and not only be able to predict RNG outcomes for the rest of the battle but also have full knowledge of your opponent’s sets. A format like Hackmons Cup practically gives you 6*4 metronome calls worth of RNG on top of the species, ability, item, EV, and IV randomization. With that much info and knowledge of your opponent’s lead species, it might even be feasible to guess the seed as player 2.
