RBY OU - Patching 1/256 + Recoveries Misses

[Heroic Troller]

Inspired by Counter being patched to avoid desyncs, i would like to bring about a discussion about misses. For long time Rby has been tiered following the rule of staying the most cartridge alike as possible. Sleep and freeze clauses were the exceptions and life went on, until now. I feel like the pandora's box has been finally open up after last decision: to patch a clear unintended bug rather than not, making a step closer to "the Rby that the devs intended originally". For those unaware every move in Rby has a one in 256 chances to miss, no matter his accuracy. Showdown's Rby will never be catridge's, and the choice to preserve Counter and patch just further pushes the case, so why not do another step in the right direction? In the direction of competitiveness, while things like Psychic drop's effect comes into play and arguably adds layer to the game, the miss is just an annoying side effect, unpredictable, uncountable, unintended, so do the right thing, push the game metagame forward and fix what is broken.

The same applies for the recoveries misses, they serve no purpose, despite being a hair better than 1/256s. No one sane would check every turn if Recover can hit, especially when your focus is on the battle itself. Rby can be improved, so why not doing it? Don't keep the status quo for the sake of it if the tier can be better, more enjoyable, more competitive.

 

[dhelmise]

Counter was not patched to avoid desyncs, it was patched to fail whenever desyncs would occur. This rule would be applicable for other edge cases (which i think they already are? tagging Plague von Karma to confirm) This is effectively what the recovery moves and 1/256 misses already are. You're comparing a game-breaking bug to a 1/256 chance to miss, and I feel like this thread shows where the line sits, and implementing these "fixes" would be crossing the line into something that isn't any better than just a pet mod. You could use the desync clause mod as a reason for this, then use this as a reason to fix draining moves missing versus substitutes in japanese rby, then use other connections to fix the present but in gsc, then fix stat overflow in oras/sm/ss, etc, and it starts to blur the clear line where we should stop. I don't really think this is a wise idea; however, I obviously don't play RBY so I'm speaking exclusively from a development standpoint.

 

[Hipmonlee]

The 'we must strictly follow cart' ship has long since sailed. If this is about that then all I can say is whenever we have had a vote on this stuff it ends up being like 90% fuck the carts.

If you are actually serious about patching misses, then I guess my response is: I dont want to. They are rare as hell, and they give people something to play for even in the direst of circumstances.

And this goes double for Recover fails. They are funny and force interesting counterplay.

 

[Plague von Karma]

I don't see how Pandora's Box was opened at all. I see the Desync Clause Mod as a way of allowing PS to avoid implementation headaches when it comes to such a complicated, game-breaking glitch. The Counter patch was one of three, the others being the Thaw and Psywave Desyncs, the latter of which has been implemented. There's also more that haven't been addressed, such as the Bide Desync and Fire Spin Glitch...which actually affects all partial trapping and is really tied to Mirror Move but whatever, fuck I guess we can call it the Mirror Move Trapping Desync? Anyway, this is more in line with things like the Acid Rain Glitch in DPP where the game can crash just because you happen to have a Cherrim on your team and Pursuit decided to get funky.

I very much dislike the idea of going down a slippery slope to begin patching RBY's core mechanics, this is very different to what I've outlined above. So far, the only things that have been patched are outright game-breaking desyncs that would otherwise cause battles to end. We are going from that to patching out minor inconveniences which are in no way game-breaking. What you are asking is that we stop playing RBY and instead move towards an idealized RBY. There are things we just have to accept as part and parcel of our game. I don't see why we should make this gigantic leap and suddenly enable other asinine "fixes" in the process. Do you want draining moves to fail against Substitute in RBY as they should? You know that was a bug that became a feature in GSC onwards, right? How about fixing Focus Energy so it's like NetBattle, or allowing critical hits to reach 100%? Maybe we should patch Substitute Self-KO? This opens up a massive can of worms and really isn't the hill you want to die on. You will miss with Body Slam sometimes, you will have Sludge poison a little more often than it should, you will click Soft-Boiled with Chansey at 64% because you didn't look at the screen properly, you will never thaw from being frozen...these are part of the generation you chose to play. I just don't see why we should change the current status quo. Every generation has some bullshit mechanic that makes everyone scream the house down.

At the same time, you could say that this can of worms is actually full of those cool jelly ones that I really like. The fixes I outlined could be considered a good thing, and I'd believe you if you said they are. However, we then need to think: where do you draw the line from there if this goes forward? It's fairly difficult to discern what is and isn't "as the devs intended", as you say, especially when the "paraslam" change as we call it was officially stated to be a fully intended mechanic, yet people refer to it as a glitch. You could argue that Stadium is where everything is "fixed", especially because of how different it is, but that's not what you want, is it? This is a completely subjective debate, a stark deviation from a tiering policy that is meant to reduce this as much as possible. That said, I'm happy to debate this subject, I find it very interesting.

If you want this mythical perfect RBY, how about banning Substitute in Stadium and playing that? I haven't heard anyone except like one guy say a single positive thing about that move's changes, hell I'd gladly try to make that happen. Worried about losing Hyper Beam? Stadium Zero NatDex, let's get it lads. Fuck, LGPE still has an active playerbase if you REALLY want to escape the ruling nightmares this generation gives you. Of course, chances are none of this would replace "standard RBY", but does this not satisfy your queries?

I am on this site to play RBY, not some "Neo-RBY" Pet Mod where we selectively remove mechanics that inconvenience you maybe once every 30 games. If you want that, however, go for it! That said, if the community ups and says "ayo let's patch this", then by all means go for it.

--

In regards to Recovery Failure, there's already been an approved PS suggestion to add a visual indicator that gives you the HP number where it'll fail. I don't see why we need to patch this when a quality of life improvement will largely solve the problem. Plays like Alakazam doing 2 STosses + Psychic in the mirror to potentially cause the glitch to occur are perfectly fine to preserve, in my opinion, especially with this upcoming QoL feature.

I know someone is going to bring up the Illegal DV "patch" that happened a while back, so I want to counter this bit now. This wasn't really a patch, but a modification to Tradeback Clause which allows you to get Max DVs from there. I quote;

The only confirmed way to have perfect DVs for every RBY Pokemon would be to transfer them from GSC, which means that we would technically allow to use GSC DVs, but not GSC Tradebacks. We decided to go with that option mainly because it sounds very inconvenient to make players choose a DV spread in their RBY teambuilder.

Is it silly? Absolutely, but this ain't the thread for that argument...

--

EDIT: Thought about this a bit more, I think I fixated too much on the patch point. So really, the conversation should be about whether we should extend patching into generic glitches and quirks at all. As I sort of went over in this post, it's a very murky and difficult topic to discuss. What would we define as glitches? There's quite a few clear-cut ones that I mentioned, sure, but if it wasn't for the Stadium 2 statement regarding the "paraslam" mechanic I'm pretty sure people would have called it a glitch and we'd have had this thread sooner. To put it simply: it's very, very hard to truly determine what's a glitch and what isn't. You'd think Slash is bugged in how frequently it crits for most Pokemon in RBY, but no, that's intended.

Contrary to what people like to say, Smogon's tiering policy is built on precedent, which makes any tiering decision a potential catalyst for new ones. Remember, this is why this thread even exists, so we should be extra careful when deciding this. Could you, say, infer that Gen 6's Mega mechanics were bugged because of Gen 7 changing it to that "you get that speed etc when you mega" mechanic? There is no right or wrong answer to this, because there is no official statement, and GF is known for retroactively fixing glitches every release. Hell, could you patch Sky Drop in BW Doubles OU so you can replace the Gravity ban with Gravity Sleep Clause? Again, this can be seen as enticing, but should we step into this speculative territory at all when we just don't know? What are you gonna do, get out the source code from the gigaleak and evaluate the code or something?

It just seems a bit extreme to go to such lengths for what should, ideally, be simulating the core games. It's in the word "Simulator". So far, we have only patched to bypass PS's upper-limit: literal game-breaking glitches that you cannot work around. What do you do? Would you say the desync patches were wrong? Should we implement desyncs or have the game end rather than change them to move failure? There are many routes down this topic.

 

[Diophantine]

Just get rid of the 1/256 misses. Fuck missing Explosion. Fuck all this "slippery slope" talk. It's objectively making the game better and doesn't even have meta nuance like any of the counter-examples people are coming up with. It's the smallest of changes that has no downside in being made.

 

[eden]

RBY hasn't been cartridge accurate for ages, freeze clause/sleep clause/desyncs being patched out.

adhering to precedent is a ship that sailed long ago viz a viz rby

there is inconsistency in even application of clauses. I'm pretty sure freeze clause/sleep clause are from stadium? I'm equally sure stadium is not what we simulate.

we very much cherry pick whatever we perceive as making it better, why not do it again?

the only relevant question here should be as to whether the competitive player base for RBY welcomes patching this out or not

i loathe the slippery slope/absurdity line of argument against changes that are beneficial, because most of the arguments that pop up in relation to that end up being very hypothetical what ifs that no one's interested in.

this isn't a core gameplay mechanic/nor does it add much depth in my opinion, but I'd love to hear the opinion of people who are more competent than myself as to what concrete positives it adds

 

[Hipmonlee]

You can play Pokemon Violet already if you want to.

 

[ABR]

The 'we must strictly follow cart' ship has long since sailed. If this is about that then all I can say is whenever we have had a vote on this stuff it ends up being like 90% fuck the carts.

If you are actually serious about patching misses, then I guess my response is: I dont want to. They are rare as hell, and they give people something to play for even in the direst of circumstances.

And this goes double for Recover fails. They are funny and force interesting counterplay.

I’m glad the RBY council has members saying shit like “this uncompetitive thing is funny.”

Anyway we already selectively clause/patch the fuck out of RBY so an objective competitive improvement that is minor yet extremely helpful should definitely be done. Also the clear best RBY’er ever wants it.

 

[Bughouse]

So let me get this straight...

Because in later generations GF changed the way move accuracy is coded, that makes the fact that RBY has 1/256 misses uncompetitive?
The hell is this incoherent argument?

 

[Plague von Karma]

So let me get this straight...

Because in later generations GF changed the way move accuracy is coded, that makes the fact that RBY has 1/256 misses uncompetitive?
The hell is this incoherent argument?

Adding onto this, I forgot to mention that fixing the accuracy notation is also an approved PS suggestion, if that helps at all. As in, not fixing the accuracy itself, but giving accurate presentation of the odds, which could be seen as the real root of the problem.

Also, hang this table on your wall or something. Left is Stadium's stated accuracy, middle is RBY's true accuracy, right is that accuracy as a percentage. We've even been pushing the true percentage notation on the Strategy Dex in C&C.

Credit goes to Pokemon Battle Historia btw, good place for statistical content.

 

[Hipmonlee]

I’m glad the RBY council has members saying shit like “this uncompetitive thing is funny.”

Anyway we already selectively clause/patch the fuck out of RBY so an objective competitive improvement that is minor yet extremely helpful should definitely be done. Also the clear best RBY’er ever wants it.

Oh cool, you have given me the opportunity to mention that explosion misses are also fucking hilarious.

Sorry for not taking this very seriously but this thread reeks of bad faith.

[edit] - see later post.

 

[Zarel]

If "what the devs intended" is a good justification, let's make Ghost-types counter Psychic-types. That's significantly clearer dev intent (the official strategy guide as well as the anime itself advises you to use Haunter to beat Sabrina), as opposed to the miss chance which is only attested by other games changing how accuracy works. (The anime even shows pokemon dodging 100% accurate moves without using evasion-raising moves, so I think dev intent is hard to establish there.)

I propose we implement this by making Ghost immune to Psychic moves, and also making Night Shade deal double damage to Psychic-types. I think this would make RBY less dominated by Psychic-types (Psychic-types are clearly uncompetitive in RBY), and maybe Mewtwo would no longer be unbeatable! Maybe Mew could even be unbanned!

 

[Excal]

I don't see how Pandora's Box was opened at all. I see the Desync Clause Mod as a way of allowing PS to avoid implementation headaches when it comes to such a complicated, game-breaking glitch. The Counter patch was one of three, the others being the Thaw and Psywave Desyncs, the latter of which has been implemented. There's also more that haven't been addressed, such as the Bide Desync and Fire Spin Glitch...which actually affects all partial trapping and is really tied to Mirror Move but whatever, fuck I guess we can call it the Mirror Move Trapping Desync? Anyway, this is more in line with things like the Acid Rain Glitch in DPP where the game can crash just because you happen to have a Cherrim on your team and Pursuit decided to get funky.

I disagree with the comparison in magnitude wrt the counter desync and this. I guess in a sense the desync glitch can literally break a battle, but at the end of the day the RBY community chose to make a patch rather than to flat out ban counter which would have been more in line with "keeping pandora's box closed". I think Troller has a valid point to some extent.

I think a lot of Troller's post hinges on the non-invasive practical improvements that his suggestion could make. Personally, I think there's no downside and this is more of a case of having to jump through hoops to make a slight change to improve the game we play. These small situations caused by a glitch can actively disrupt a battle by taking the game out of the player's hands in a particularly bothersome manner. I don't think it's a slippery slope considering it's such a small modification that changes pretty much nothing fundamentally about playing the game other than protecting players from potential outlandish situations. I support it (I also agree w/ eden's post), but that's just me.

 

[Plague von Karma]

I disagree with the comparison in magnitude wrt the counter desync and this. I guess in a sense the desync glitch can literally break a battle, but at the end of the day the RBY community chose to make a patch rather than to flat out ban counter which would have been more in line with "keeping pandora's box closed". I think Troller has a valid point to some extent.

I agree that Troller has a point, thus why I've taken this seriously. It's an interesting discussion.

I think the difference is bigger than it looks, though; it's particularly with the Thaw Desync. The current precedent is "patch all desyncs", and this came with the introduction of Desync Clause Mod on PS. This is the only "patch" RBY has truly received (so "patching the fuck out of RBY", in ABR's words, is technically incorrect). If you were to avoid patching desyncs and instead ban what causes them, you would have to ban either all the moves that can freeze or all the moves that can burn. That's just one situation, both patchless outs resulting in almost the complete removal of an offensive type. Want to ban Bide and Mirror Move next? Seriously, what alternative was there, and why are we focusing on one part of a much bigger picture? 1/256 is a perceived glitch that doesn't break the game in any way, and the alternative to patching it is to leave it alone to do its statistically improbable thing.

 

[ABR]

Requesting the removal of sleep/freeze clause if this does not go through, thanks.

- breezy

(serious edit: those are far more invasive, cart-straying, and meta-changing than the 256 stuff!!!)

 

[Amaranth]

I've been DMed by someone asking to patch RBY freeze into 'modern' freeze because it's "more competitive". Maybe yes, but I want to play RBY. If I wanted to play newgens I would play newgens.

Clauses are not just an RBY thing - let's not act like Sleep Clause gives precedent to patch out random shit we dislike because otherwise we'd be patching Scald's burn rate down in ORAS or whatever other change you want to think of in any other tier with sleep clause. I don't really see why if we're all collectively happy with Sleep Clause in every other gen we'd be unhappy with Sleep+Freeze clause in RBY - they both limit status to make the game playable, and we've all agreed that they're necessary but we don't want to change anything else.
If you're gonna use the "b-but sleep clause" argument then let's follow it all the way and do like BW - you can make very many arguments about sleep being uncompetitive anyway, just look at all the dicerolls involved with hitting all the inaccurate sleep moves + early wakes (and sometimes late wakes too) singlehandedly deciding games. But we can't exactly go banning Ice Beam and Blizzard, so either way something will have to be claused out there - and that's fine!! Literally every tier does this or has done this at some point in the past!!! RBY is not special in this regard.

Where RBY is special and unique is the Counter patch thing, but even that was justified by the Acid Rain precedent. Desync patches still do NOT open the gates to patching everything we like - if they did, DPP could also be patching Jirachi's Iron Heads to be less invasive, but they don't, because they are "a real tier" that has to follow real smogon rules. RBY for some reason thinks they can do whatever the hell they want. And then we wonder why the broader smogon community thinks we're insane lol.

There is no valid precedent to justify patching out 256es. We are not playing true cartridge RBY but we are trying our hardest to stick as close to it as possible while remaining competitive - much like DPP, which "patched out" sleep via Sleep Clause and Acid Rain. We are not special and, outside of the status clauses, we don't get to patch out hax. No tier gets to patch out hax, why on earth should we.

I'll accept patching 256es the moment DPP patches No Guard to only be a +20% accuracy boost and nerfs Iron Head's flinch rate to 15%. Until then, this is all nonsense

 

[Hipmonlee]

Oh cool, you have given me the opportunity to mention that explosion misses are also fucking hilarious.

Sorry for not taking this very seriously but this thread reeks of bad faith.

Ok, Excal asked me to edit this. And fair enough, I was just escalating shit here. But editing was getting a little awkward, so here's a reply instead.

Firstly: I dont really care much about the concept of fixing uncompetitiveness. It's a game, the point is to have fun. Part of the fun of RBY is its batshitness, and 256 misses and recover fails are definitely part of that. Honestly I would be happy to go back to a cart legal sleep and freeze clause, but, as I alluded to earlier, I lost that argument years and years ago. Sorta in the "before you were born" kinda era for some of you. But ultimately people only play this game because it is Pokemon. There's a cost to making the game less and less like the cart. Sometimes the cost is worth it and sometimes it isnt. Misses are just not important enough for me to want to pay that cost.

Secondly: I have a certain amount of time I am able to spend on questions like this, and this thread hadnt really passed the hurdle of convincing me that it was made in good faith. So I dropped a dismissive reply while I was working. If this thread was made out of a genuine desire to improve the metagame, rather than a desire to undermine the counter ban, then I'm sorry.

Thirdly: ABR's post called me out for my shitty arguments while also arguing that we should do it just cause Troller says so? This is obviously bad faith.

 

[Triangles]

As a spectator, ridiculous 256 misses and players fucking up Chansy recovers is by far the best part of watching RBY.

 

[Plague von Karma]

Since some people clearly just don't get it, I'm going to dedicate this post to explaining Desync Clause Mod's existence. It's not hard to understand: there is a single patch comprising every desync.

The first thing that needs to be understood is that RBY has been patched "once". This patch was Desync Clause Mod: events that lead to desyncs are changed to move failure. This currently comprises Counter, Psywave, and thawing. I wouldn't be surprised if this also leads to Mirror Move + Partial Trapping KOs and Bide's weird damage building, but that's something for another time. As aforementioned, Illegal DVs were a simplification to Tradeback Clause that makes it say "Tradeback moves from GSC are not allowed." rather than "Tradeback moves and Pokemon from GSC are not allowed". In that case, calling it a "patch" was poor wording.

Before you say "Counter came first, we never thought about thawing", no, that's disingenuous. You can see it being discussed here, here, and here. You will notice that from the last post, this vote was made with the end goal of eventually patching more desyncs. That led to the creation of this singular patch when I finally got around to implementing Counter and Psywave into the mod. In other circles outside of this thread, there was a lot of discussion regarding the Thaw Desync as an argument for patching Counter as well, with the argument that later down the line all desyncs should be patched. As we now know, on the tiering decision itself, you will also see that this became a reality, because it was the correct decision to make.

So, now that you know not to simply call this "the Counter patch", let me explain why there is no viable alternative to Desync Clase Mod. To avoid this mod, you would need to make cases for banning the following;

• Bide
• Counter
• Mirror Move
  • You could also ban partial trapping, but the ramifications of this would be much larger, in my opinion. That would include Bind, Clamp, Fire Spin, and Wrap. Note that "Fire Spin glitch" is a bit of a misnomer for how this desync operates.
• Psywave
• Every move that can burn (Ember, Fire Blast, Fire Punch, and Flamethrower)
• Probably a few more things, I think there was something with Light Screen / Reflect?

By doing this, you are removing every single Fire-type move except Fire Spin, unless you choose to ban partial trapping, in which you literally erase Fire-type attacking moves to avoid patches. I don't see this as a reasonable alternative to patching the game: you're making every RBY Fire-type worse than they already are to own the libs. Recall that we tier Pokemon, not moves. Outside of desyncs, these moves are completely harmless, otherwise they would have been banned 2 decades ago. Do you really want to see Ember on a banlist when the alternative to making a super long list is right there? Part of what has been the core of our tiering policy is not making a long list of rules to aid accessibility, and banning every Desync move adds 8 new lines - with possibly more to come by the way, this game shits the bed every other year - many of which won't make sense to outsiders. On that note, it makes more sense to have a consistent sweeping patch rather than picking and choosing what desyncs to patch, don't you think?

Desync Clause Mod isn't just a thing for Smogon's formats either, it's a simulator convenience that has been extended to recent additions like Nintendo Cup 1997 through inheritance. Ergo, we can see this as a unique case akin to Sleep and Freeze Clause. Sleep and Freeze Clause were originally implemented because of Stadium having them which, while odd reasoning now, was fine back then. We've kept them because the status quo didn't need to be changed and, again, it was a simulator convenience: not having games end in absurd scenarios like switching in a Pokemon on Sleep Powder during PP stall to cheese is a good thing. Here, I see Desync Clause Mod in the same light: not having games randomly end because your Chansey is a bit too eager to Counter that Snorlax is a good thing.

This generation already requires reading a guide before even entering a game. With what I've said here in mind, I believe that Desync Clause Mod is clearly the superior route to take, and I don't see how you can even remotely equate this with a 1/256 or recovery move failure patch.

On a side note, if you want all desync moves banned, you'll also have to campaign for a Beat Up ban in GSC, but I'm sure you'll all back up and go "yo, wait dude, that's different" for some stupid reason.

7/5/2021 Edit:
I forgot a minor point regarding this after a conversation with a friend on Discord. Some players have proposed patching Dig/Fly in the past, here's a video showing the Semi-Invulnerability Glitch for context. I have never seen people actually defend the argument after a single point for the same reason we have adamantly opposed this PR thread: it's ridiculous and a stark deviation from the cartridge. I don't see why 1/256 or Recovery Failure are treated any differently outside of people hating them more.

 

[power]

Requesting the removal of sleep/freeze clause if this does not go through, thanks.

- breezy

(serious edit: those are far more invasive, cart-straying, and meta-changing than the 256 stuff!!!)

This is absolutely correct. Sleep clause shouldn't exist, especially in modern generations; it's incredibly invasive and can't be replicated on cart. Moves that exclusively induce sleep should be outright banned, just like baton pass.

The fact that we're in gen 8 and still using such an outrageous cart-straying pet mod is ridiculous.

edit:a word

 

[MetalGro$$]

I agree with pretty much everything TIN wrote. Cartridge-wise, this is just nitpicking that can never be justified. Playing and spectator wise, your last hope of you(r teammate) pulling out the unlikeliest of wins is gone.

I know 'uncompetitiveness' is the whole rage now, but Pokemon is inherently related with odds, ESPECIALLY in RBY. Proposing this in RBY basically goes against everything RBY is. I, not the best player ever, have personally enjoyed the very slight edge I've gotten in games from recognizing when I was on the recovery percentage number, shoutout TonyFlygon, but even that won't be necessary anymore as PvK pointed out that the alert you're on a Recovery glitch HP number is set to be implemented. It can be found here, for if you didnt read PvK's long post: https://www.smogon.com/forums/threads/visual-indicator-for-rby-recover-failure.3677391/#post-8737684.

Also, whatever happened to the whole principle of leaving old gens alone? There's probably a way to force yet another meta shift in BW instead of messing with RBY!

 

[Hipmonlee]

Firstly: I dont really care much about the concept of fixing uncompetitiveness.

Ok, this bit apparently caused some consternation so here's an explanation of how I think about these things. This is based on a conversation I had in the OGC chat, which I figure might be better to spread a bit wider. I was hoping to find like an old post of mine where I explain this, but that took too long so I am just gonna write it up again.

So apparently the popular definition is that things are uncompetitive when they reduce the odds that a stronger player will beat a weaker player. And the reasoning goes that anything introducing variance will be necessarily beneficial for the weaker player, and should be removed whenever possible. Personally I think this is a gross oversimplification and I think, particularly in the case of rby, it can be utterly self-defeating.

For instance, consider a scenario where we have three players in the finals of a tournament, and they are doing a round robin to see who wins. And in that finals you have a great player, a good play and a player who is a bit rubbish... like, they're just a bad rbyer...

So lets say the great player wins their set against the good player 60% of the time, and they win their set against the bad player 90% of the time. And the good player beats the bad player 60% of the time.

My stupid script I threw together gives me the following odds of victory:
Great player wins: 0.6459899129746836
Good player wins: 0.2988404865506329
Bad player wins: 0.05516960047468354

So then consider a rule change is made, that effectively halves the odds the bad player beats either better player, and gives a slight advantage to the great player against the good player. So now the great player has a 95% chance of beating the bad player, and a 62% chance of beating the good player, but the good player has an 80% chance of beating the bad player.

Now I get:
Great player wins: 0.6395206898482558
Good player wins: 0.3468839240718965
Bad player wins: 0.013595386079847695

So, even though this change always increases the odds of the better player winning any individual set, it actually overall reduces the odds that the best player wins. This is because part of the differentiation between the two high level players was their success rates against lower level players, and this change has actually reduced that.

Right now this might seem like a fairly abstract hypothetical but I think its pretty close to what you might see in rby. And in fact, I think it could be even worse than this in practice in rby.

You might not know this about rby, but there isnt actually much variation in the teams people use. And because of that it runs the risk of becoming extremely formulaic.

So lets say I am in an rby tournament against a generic good new gen player who dabbles in rby. In general I expect them to know the basic ideas, like: paralysed Chansey can win freeze wars against unparalysed Chansey, but it has to watch out for Snorlax repeatedly doubling in against it. I expect them to be slightly inaccurate, but mostly that is just about minimising risks. But if I bring a fairly standard 6 and none of those risks end up costing them, and the battle follows the standard lines, then it's fairly likely that this battle is going to come down to some really basic stuff like whose Tauros wins the speed tie.

But! The reason I am great at rby is because I know what to do when the battle doesnt follow the basic formula. If my Chansey is paralysed, and I freeze their Snorlax on the first turn it comes in, then---barring miracles---I've won. I haxed them, but we take those. On the other hand, if they have the paralysed Chansey and I get my lax frozen, then I have enough weird lines in my repertoire that I can probably still make it close. This is the sort of shit that differentiates great players from good ones. It's not things like switching in Exeggutor on the turn your opponent uses Reflect. Everyone knows how to do that.

Imagine we removed freeze from the game entirely. It might mean that the great players and the good players all massively increase their odds of beating bad players. But if it makes the game more likely to follow the standard lines more closely it could actually reduce the odds the great players beat the good ones.

Luck events create unpredictable scenarios. In rby this is crucial to the functioning of the game. Without luck events every battle would basically be the same. Luck events give you the richness that allows for differentiation between great players and good ones.

In my opinion what we really should be pursuing is better differentiation at the highest levels of play. And I dont think that reducing luck elements from rby is going to help with that at all.

However! Nobody is asking us to remove all of the luck from rby. The question here for me is whether these specific examples will reduce or increase the skill cap.

I think the recover glitch is an obvious no. That one leads to some really weird situations that you really need to think to adapt to. One player trying to find ways to get their pokemon chipped with a small enough amount of damage to let them recover again, while the other is trying to prevent that. This is precisely the sort of thing that we should be trying to encourage in rby.

256 misses, well, it's really hard to tell... But what isnt hard to tell is that it requires patching the game, which is a definite cost. So weighing up a weird "it may be better but it isnt very clear", against a definite cost, its a pretty easy no for me.

There you go.

For those of you who were thinking that despite being involved in pokemon tiering for nearly 20 years I dont consider it something that requires serious analysis, I guess you dont really know me very well. So:

Hi,
My name is Kit.
I love overanalysing every fucking thing.
But I genuinely believe Pokemon is one of the most conceptually beautiful competitive games ever devised. So I especially love overanalysing that.
Check out my rby prediction guide.

 

[Plague von Karma]

I was asked by EB0LA to share a new point regarding the "what should we patch in RBY" debate since it's a fairly interesting one that I haven't thought about. I'm personally for just implementing them, as they are fairly rare occurrences, but some people are bringing up the idea of patching and I think it's worth opening a discussion on the topic, as this can be seen as a line between desyncs and simple glitches in that one side's input is completely removed until their Pokemon faints.

There are various glitches involving Freeze that result in player input being removed, particularly ones involving multi-turn attacks. I am educated on two specific ones: Faster Hyper Beam + Freeze in one turn, and Haze + Sleep/Freeze + any Multi-turn move like Rage or Thrash. These glitches occur because of the moves being used prior to the status. They remove player input on use, and when it's interrupted in some way, particularly with a freeze in the equation, this happens. Any multi-turn move is applicable in regards to the Haze glitch: think Sky Attack, Hyper Beam, and Bide.

In these situations, you can't attack, switch, or forfeit. When I say player input is removed, I literally mean it is completely removed until the affected Pokemon faints.

Here is footage of the one I tested today;

Skip to 4:00 for when the funniness begins.
It should be noted that you have to get frozen before you recharge with Hyper Beam. SadisticMystic also has footage of this happening outside of a Link Battle here.

FOMG and Hipmonlee asked me if this happens on any turn, you can see this happening on a slow Hyper Beam here, s/o to SadisticMystic.

And here's the "Haze Glitch" as it's commonly referred to, iirc this isn't implemented on PS;

I showed this with sleep and it's the only glitch of its kind where this happens. Otherwise, freeze is the only offender.

The most relevant area you'll see this in RBY OU is prolly someone gambling on an Ice Beam w/Snorlax or Chansey when Tauros prematurely Hyper Beams or something. If it works though, you can quite literally do anything you want. Switch in your Recover uses, get some HP back, maybe set up some Slowbro Amnesias, etc. Probably the best punish you can possibly get for such a thing, but it's definitely the riskiest...

Anyway, that's all I have to share. The Hyper Beam interaction is on PS as shown here.

 

[august]

Since this thread has devolved into arbitrary discussion about patching things that we don't like, visual showdown aids for mechanics, and tables of move accuracies (??), I would like to address some points about using statistical simulation as a means to argue for or against certain causes.

Ok, this bit apparently caused some consternation so here's an explanation of how I think about these things. This is based on a conversation I had in the OGC chat, which I figure might be better to spread a bit wider. I was hoping to find like an old post of mine where I explain this, but that took too long so I am just gonna write it up again.

So apparently the popular definition is that things are uncompetitive when they reduce the odds that a stronger player will beat a weaker player. And the reasoning goes that anything introducing variance will be necessarily beneficial for the weaker player, and should be removed whenever possible. Personally I think this is a gross oversimplification and I think, particularly in the case of rby, it can be utterly self-defeating.

For instance, consider a scenario where we have three players in the finals of a tournament, and they are doing a round robin to see who wins. And in that finals you have a great player, a good play and a player who is a bit rubbish... like, they're just a bad rbyer...

So lets say the great player wins their set against the good player 60% of the time, and they win their set against the bad player 90% of the time. And the good player beats the bad player 60% of the time.

My stupid script I threw together gives me the following odds of victory:
Great player wins: 0.6459899129746836
Good player wins: 0.2988404865506329
Bad player wins: 0.05516960047468354

So then consider a rule change is made, that effectively halves the odds the bad player beats either better player, and gives a slight advantage to the great player against the good player. So now the great player has a 95% chance of beating the bad player, and a 62% chance of beating the good player, but the good player has an 80% chance of beating the bad player.

Now I get:
Great player wins: 0.6395206898482558
Good player wins: 0.3468839240718965
Bad player wins: 0.013595386079847695

So, even though this change always increases the odds of the better player winning any individual set, it actually overall reduces the odds that the best player wins. This is because part of the differentiation between the two high level players was their success rates against lower level players, and this change has actually reduced that.

the first thing that I always do when I am presented with results from a simulation is to see if I can recreate the results myself. since there was no code provided, I decided to implement Hip's scenarios myself in the R programming language. high level languages are particularly useful for sanity checking results, but come at the cost of computational speed. the code that i used to implement these scenarios can be found here: https://pastebin.com/sneDddHA (warning, it is not pretty or fast but as they say in Babe, that'll do, pig). i will say that i did receive slightly different results than hip and i believe that my code is correct, but everyone is human and its possible that I made a mistake somewhere.

the next thing that i want to mention: 1 simulation result does not mean anything. you should always be attempting to invoke the central limit theorem and looking at the sample means rather than putting your eggs in the basket of a single simulation run. the following histograms show the sample averages over 1000 round robin tournament simulations, repeated 1000 times. there is still considerable monte carlo error in these histograms.

Scenario 1 histogram:

Hide

Scenario 2 histogram

Hide

code:

Hide

https://pastebin.com/Adtw6jWb

the end conclusion under these 2 scenarios is the same though - the "great" player does win the tournament less of the time despite their winrates being increased against both classes of players. things like this will often lead to the question: why is that? let's take a look at the changes from scenario 1 to scenario 2.
In scenario 2:
the great players winrate vs the good player goes from .6 to .62 (3.3% increase)
the great players winrate vs the bad player goes from .9 to .95 (5.5% increase)
the good players winrate vs the bad player goes from .6 to .8 (33.3% increase (!!!!))
now, it is definitely possible that these are accurate. however, in order to be careful you should always check how sensitive your simulation is to certain inputs changing. this is referred to as a sensitivity analysis. in order to perform a sensitivity analysis, we want to make a scenario where we feed in that probability that governs the good players winrate vs the bad player, and look at how it changes the winrate of the "great" player in the tournament.

Hide

the code for the sensitivity analysis can be found here:

Hide

https://pastebin.com/ir7cEb4Z

from the simulation results, we see that this winrate is quite sensitive to the probability, and in fact only surpasses the scenario 1 threshold for values near p = ~.77, which is still over a 28% increase, compared to the 3.3% and 5.5% increases for the other winrates, and should require justification.

if i were to look into building my own simuation for the sake of arguing here, i'd probably look at something like follows.
lets assume that an average rby game lasts around 100 turns. we can say that of these 100 turns, the better player will spend 75% of them attacking, which is probably a bit high anyway. that means about 75 attacks will be fired off in each game. on average, every 3.4 games will contain a 1/256 miss for the better player against the worse player. we can say that anywhere from 25% to 75% of these misses will negatively impact the game and result in a loss for the better player. this will alter the winrate by anywhere from 7% (.25*(1/3.4)) to 22% (.75*(1/3.4)), both of which are (and this is true) below the 33% threshold that was assumed here, though it is worth mentioning that it is also substantially higher than the other winrate change percentages. this is also the worst case scenario, as it assumes that better players are not positively impacted by a worse player being on the negative end of a 1/256 miss.

i heavily encourage the use of statistical simulation to help make things understandable and help people to visualize your arguments. statistics is a beautiful field that takes information, turns a few cranks, and creates information that is digestible for the masses. however, it is important to be able to justify the assumptions that youve made, and when you can't, you should be able to quantify how much certain assumptions matter to the results that are being presented, otherwise you are leading people right into the classical trap of "there are lies, there are blatant lies, and there are statistics".

to paraphrase the great statistician jim clark: always remember, giving statistical tools to pokemon players (ecologists) is like giving a gun to a baby

 

[Hipmonlee]

I really thought about removing the statistical part entirely. I just wanted to get the concept into people's heads that when you say 'increase the odds the better player wins' that it may not be enough if you dont consider who the better or weaker players are, and just assume the increase will be the same at all levels. Hence one example is enough. And it felt like a fun little paradox.

Code is attached, you can see what I mean when I called it stupid.