Resource RBY & Tradebacks Bug Report Thread

Thanks to Plague von Karma and the programmers, eighth Pika Cup pokémon that were unfairly banned are now allowed in the simulator. Even Muk and Kangaskhan are in, yay!

Currently there are only four pokémon that are still banned, I'll mention them below:

1. Hitmonlee levels 20-29
2. Hitmonchan levels 20-29
3. Omanyte levels 20-29
4. Kabuto levels 20-29


The sim has improved a lot since 2018, well done chums!
 

Sceptross

The words were voted out by a landslide
is a Community Contributor
Hello,

When testing the viability of Substitute Chansey in RBY OU, I noticed that Counter fails if the user is behind a substitute, as can be seen in this replay, in turn 2:

https://replay.pokemonshowdown.com/gen1ou-1414342202-uadgsall698l8qkkflvmsexcvuaer0qpw

However, this is not cart accurate. An attack absorbed by a Substitute can be countered for the amount of damage it would have done to the user had it not had a Substitute. This can be seen in action in the following video:


This is a behaviour that may have a big impact in the tier. A Chansey with Substitute and Counter that is not paralysed potentially shuts down what are arguably the three most common Snorlax sets (Reflect+Ice Beam, Reflect+Selfdestruct, Reflect+Hyper Beam), since IceLax can no longer fish for freezes other than the turn Chansey switches in, and the normal moves can't be used at all or you risk getting KO'd by Counter. Obviously, in the current state of things, this is not the case, since Counter fails and Snorlax can just go for Body Slam and break the sub.

The current way of Chansey dealing with these sets are potentially less consistent, since they are prone to critical hits, freezes, or being boomed on, none of which Substitute+Counter Chansey cares about (SD is a little bit iffer, but the Chansey user can play mindgames by threatening a switch to Rhydon or a frozen/asleep Pokémon). Therefore, I'd like to request a fix to this behaviour.
 
Last edited:
i was asked to add this to the thread, @ haze being incorrectly implemented on showdown

“in this replay,
1 gengar hazed +6 hypno, but the stat boosts werent removed as hypno’s psychic ohkoed gar, then did 86% to lax (as a player commented in the chat)
2 haze actually is supposed to reset para/burn stat changes including reapplications, in addition to removing opp status, however it didnt do this either as snorlax outsped the no longer parad hypno turn after haze
also the para didnt visibly disappear til after lax bslam lol, yeah idk…
https://replay.pokemonshowdown.com/gen1tradebacksou-1414078613

obv it is of very minimal relevance but haze can see very occasional use in ubers (vaporeon which some ppl put in d rank), and perhaps in tradebacks (eg on tentacruel)
 
also, the fly and dig bans are not implemented in tradebacks ou on showdown.

(sry for double post, i j noticed this today watching a replay)
 

Plague von Karma

Banned deucer.

Hipmonlee

Have a nice day
is a Community Contributoris a Senior Staff Member Alumnusis a Smogon Discord Contributor Alumnusis a Tiering Contributor Alumnusis a Top Contributor Alumnusis a Battle Simulator Moderator Alumnusis a Four-Time Past WCoP Champion
Damn, this really wasn't in known bugs? It is something that's been known about but if it's not documented as such then that's an issue in and of itself that's a really good catch. Hopefully it can be fixed at some point bc current sim mechanics unfairly favour the PT user.
 

Amaranth

is a Site Content Manageris a Forum Moderatoris a Community Contributoris a Top Tiering Contributoris a Top Contributoris a Dedicated Tournament Hostis a Tournament Director Alumnusis a Social Media Contributor Alumnusis a Community Leader Alumnusis a Past SPL Champion
Moderator
I have no handy way to test this on cartridge but as far as I know Substitute should NOT block secondary effects from moves (eg. Body Slam para chance, Ice Beam freeze) if it breaks, and it currently does not behave like that (I played two games testing this 1 2 and unless i failed a 30% check 40+ times in a row it's not at all implemented correctly)

This is a pretty important mechanic to RBY UU so I would like to see it addressed quickly

EDIT: False alarm, Sceptross tested on cartridge and the mechanics guide I was referencing is evidently incorrect
 
Last edited:

Sceptross

The words were voted out by a landslide
is a Community Contributor
According to the following replay (turn 38), PS implements partial trapping moves in a way that if the trapper faints as it is trapping the opponent (i.e. via Poison damage) the trapping effect is still in play in the following turn:
https://replay.pokemonshowdown.com/gen1uu-1477464728-ddjz7d9iub2ln88dk9a0b0btjgoz7drpw

This is not cart accurate, as can be seen in the following video. The partial trapping condition must be removed as soon as the trapper faints. This bug has big implications in several RBY lower tiers, namely UU and NU.

 

phoopes

I did it again
is a Site Content Manager Alumnusis a Senior Staff Member Alumnusis a Community Contributor Alumnusis a Contributor Alumnusis a Battle Simulator Moderator Alumnus
Gen 1 - When a trapping move (in this case wrap) is used via metronome it only lasts for one turn. Please see replay, turn 99: https://replay.pokemonshowdown.com/gen1ou-1472946994-7bmsgqbdtpr98t2rig3xq7mf9o7ykxepw
When looking at this replay, it seems you were paralyzed during what would have been the 2nd turn of Wrap. Can you try to replicate this scenario so we can see what would happen if you weren't paralyzed? I'm assuming you were able to select Metronome again, which based off the quick research I did, would be incorrect. I'll add it to the list once you confirm, though.

In Gen 1 the move Bide should include damage from burn/poison when it's 'unleashed', however this doesn't happen (turn 29): https://replay.pokemonshowdown.com/gen1ou-1478732327
According to the research I did, if a Bide user isn't directly attacked while Biding, Bide should fail. So it looks to me like this isn't a bug. If someone with more definite knowledge could confirm though, that would be helpful.

Also, Bide doesn't damage substitutes - turn 20: https://replay.pokemonshowdown.com/gen1ou-1478861386
This bug was already reported here.

According to the following replay (turn 38), PS implements partial trapping moves in a way that if the trapper faints as it is trapping the opponent (i.e. via Poison damage) the trapping effect is still in play in the following turn:
https://replay.pokemonshowdown.com/gen1uu-1477464728-ddjz7d9iub2ln88dk9a0b0btjgoz7drpw

This is not cart accurate, as can be seen in the following video. The partial trapping condition must be removed as soon as the trapper faints. This bug has big implications in several RBY lower tiers, namely UU and NU.

Unless I'm misreading (so correct me if I'm wrong), seems the bug was already reported here. I linked your post in the OP though because of the cart video, very helpful! Thanks!

(Copied and pasted from Suggestions as suggested)
[Note: The person who told me about is mechanic is Enigami, he knows more than me and might be able to fill in the exact mechanics that'd need to be known to implement properly]

In Gen 1, to prevent the games from desyncing on a link battle, the game pre-generates numbers for the first 10 RNG rolls. However, the game only generates values from 0-252 rather than 0-255 for these rolls. This effects every RNG related mechanic before the first 10 RNG rolls are used up, including accuracy, secondary effect chance, crit chance, damage rolls, and even speed ties. The bug here of course being that PS does not simulate this, all rolls are 0-255.

Examples of effects on gameplay:
-Just about every effect becomes more likely, since typically RNG rolls look for X number or lower to take into effect. This also means you cannot 1/256 miss in the first 10 RNG rolls, as that happens when you roll a 255 for accuracy.
-Jynx's Lovely Kiss has a 191/256 (~74.6%) chance to land normally, but if in the first 10 RNG rolls, it has a 191/253 (~75.5%) chance instead. Blizzard has 230/253 (~90.9%) accuracy and a 26/253 (~10.3%) chance to freeze if it hits.
-If Alakazam uses Psychic, the game does a total of 4 RNG rolls for: Accuracy, Damage, Critical hit, Special drop. Add another for full paralyses and speed ties if they apply.
-Player 1 wins speed ties if the roll is 0-127, player 2 wins if the roll is 128-255. If the maximum roll is 252 instead the odds change to 128/253 (~50.6%) for player 1 and 126/253 (~49.4%) for player 2.
Note: All of this occurs only within the first 10 RNG rolls, after these rolls all rolls become the standard 0-255.

A few things need to be checked to implement properly such as: Is the player this increases the speed tie chance for consistent? (ie: Is it always the host?) What order do RNG rolls occur? (If there are 2 remaining RNG rolls and Alakazam uses Psychic, which RNG rolls get changed odds?) And just general verification of the mechanic itself.

edit: Looking back on the chat, Enigami seems to say the order is: Accuracy, Critical Hit, Secondary Effect, Damage roll; and that under certain circumstances damage roll may use 2 rolls. Still worth verifying all this on cart to ensure proper implementation.
Added
 

Plague von Karma

Banned deucer.
Incredibly minor one here.
1654436172049.png

The movesets here are incompatible in NC97 specifically because it uses Red and Green learnsets. Stress-tested as I'm writing a script for something stupid.
  • Nidoking can only learn Double Kick via an L43 Nidorino, while it can only learn Thrash at L23 as a Nidoking. These moves can never be obtained together until Yellow, where Nidoran-M learns Double Kick early enough.
  • Butterfree can only learn Harden via an L7 Metapod in Yellow, and String Shot + Tackle via a Caterpie. In Red and Blue, Metapod doesn't learn Harden upon evolution, leaving the moves incompatible. Kakuna had this oddity but it never got fixed in Yellow, funnily enough.
As implied, these incompatibilities were resolved in Yellow, which is why this problem has come up at all.

I'm not sure how these bugs happened on the sim but I have an inference from gen1jpn/rulesets.ts where this stuff is handled;
JavaScript:
line 27:  nidoking: {doublekick: 43},
I don't believe Nidoking needs this line at all, as Nidorino appears to handle it. As far as I can see, the validator thinks Nidoking learns it on its own at L43, which doesn't happen. At the same time though, not having it there seems like it'd default to L12? Am I overcomplicating things? That aside, there's no entry for Metapod at all, which should have Harden set to L1 only. I don't know if Yellow learnsets make validation weirder though.

I don't actually know if this is the problem but if it is then I can fix it in a flash.

EDIT: I did a Pull Request which got merged, will be live on the next hotpatch..
 
Last edited:

phoopes

I did it again
is a Site Content Manager Alumnusis a Senior Staff Member Alumnusis a Community Contributor Alumnusis a Contributor Alumnusis a Battle Simulator Moderator Alumnus
When looking at this replay, it seems you were paralyzed during what would have been the 2nd turn of Wrap. Can you try to replicate this scenario so we can see what would happen if you weren't paralyzed? I'm assuming you were able to select Metronome again, which based off the quick research I did, would be incorrect. I'll add it to the list once you confirm, though.
I was playing OU and this came up again surprisingly, it seems that when Metronome calls a partial trapping move the move is not executed to completion. Could be a bug, will need someone to test on cart but making a note of it anyway.

https://replay.pokemonshowdown.com/gen1ou-1629249919
 

Hipmonlee

Have a nice day
is a Community Contributoris a Senior Staff Member Alumnusis a Smogon Discord Contributor Alumnusis a Tiering Contributor Alumnusis a Top Contributor Alumnusis a Battle Simulator Moderator Alumnusis a Four-Time Past WCoP Champion
Yeah, currently waiting for showdown devs to voice some kind of support for fixing this. Though my window for working on it may have closed already...
 

Sabelette

from the river to the sea
is a Site Content Manageris a Community Contributoris a Top Contributoris a Forum Moderator Alumnus
Poison is currently bugged on sim.

Cart behavior - if you switch in and get poisoned by your opponent's move right after your switch (whether its Poison Sting, Toxic, etc) you take damage at the end of the turn and increment Toxic, so 6.25% immediately, 12.5% next turn (the first turn your switched in mon can act), etc.

Showdown behavior - you switch in, take poison, but take no damage immediately. Next turn, if you are Toxiced the UI says "next damage: 12%" (which is accurate) but you take 6.25%. The turn after it says you will take 18% but you take 12.5%, etc. In other words, UI is correct but the damage is not, it's delayed one turn.

I will double check this in a link battle to confirm and edit after.

Edit: Llinos and Frrf checked it, UI is bugged; in single player what I said is correct, in link battles you don't take damage when switching into Toxic. UI is bugged and should be fixed to to show the right increment (one below current) for this situation.
 

Users Who Are Viewing This Thread (Users: 1, Guests: 0)

Top