Stat Reduction Bug

[M Dragon]

So as Crystal_ recently discovered, in RBY and GSC stats reductions work in a different way than in the later gens.
The standard multipliers for stat reduction (and the implemented in every simulator) are:

-1: floor [stat * 2/3]
-2: floor [stat * 2/4]
-3: floor [stat * 2/5]
-4: floor [stat * 2/6]
-5: floor [stat * 2/7]
-6: floor [stat * 2/8]

However, in RBY and GSC, it is programmed in a different way:

-1: floor [stat * 66/100]
-2: floor [stat * 50/100]
-3: floor [stat * 40/100]
-4: floor [stat * 33/100]
-5: floor [stat * 28/100]
-6: floor [stat * 25/100]

As you can see, the multiplier is different at -1, -4 and -5.

What does this mean?
Snorlax speed is 158.
Skarmory speed after a curse using old formula is: floor [238 * 2 / 3] = 158 (I think in NB it was 159)
Skarmory speed after a curse using ingame GSC formula is: floor [238 * 66 / 100] = 157

That means that Snorlax should always be faster than Skarm after a Curse instead of a speed tie, which is huge.

How do you think metagame would have changed if implemented correctly in PO and NB?

 

[magic9mushroom]

Don't you mean "how the metagame will change once it is implemented correctly in sims"?

Makes CurseSkarm slightly less useful vs. DrumLax, and... that's about it. So slightly less CurseSkarms and more DrumLaxes.

(CurseSkarm's use vs. last-Poke CurseLax isn't affected, because Lax also loses speed.)

 

[M Dragon]

Makes CurseSkarm slightly less useful vs. DrumLax, and... that's about it

And that is huge
Skarm is no longer a reliable BDLax counter (even without LK or FB).
If the BDLax user sends something like Raikou/Zapdos in the turn Skarm rests to force it out, Skarm cannot stop BDLax anymore.
With the implemented mechanics Skarm was able to outspeed Lax even after a curse, allowing it to avoid the 3HKO.

 

[magic9mushroom]

And that is huge
Skarm is no longer a reliable BDLax counter (even without LK or FB).
If the BDLax user sends something like Raikou/Zapdos in the turn Skarm rests to force it out, Skarm cannot stop BDLax anymore.
With the implemented mechanics Skarm was able to outspeed Lax even after a curse, allowing it to avoid the 3HKO.

That's a load of tosh (even ignoring the fact that it's been implemented as a speedtie for ages now and there's nothing "reliable" about a speedtie).

X, come back!
Go! SKARMORY!
Enemy SNORLAX used BELLY DRUM!

Enemy SNORLAX used RETURN! (~47%)
SKARMORY used WHIRLWIND!
Enemy SNORLAX was blown away!

[let's give Snorlax the benefit of the doubt and assume Skarm is immediately forced out by whatever came in, and then Snorlax gets in again without taking damage]

X, come back!
Go! SKARMORY!

Scenario 1:

Enemy SNORLAX used BELLY DRUM!

SKARMORY used DRILL PECK!
Enemy SNORLAX fainted!

Scenario 2:

Enemy SNORLAX used REST!

SKARMORY used REST!

[Snorlax will wake up and use Belly Drum as Skarm comes in with 0 sleep turns left, Skarm will wake up and phaze as it gets hit bringing us right back to where we were at the second turn above]

Skarm doesn't actually need Curse to stop DrumLax unless it's last-Poke or you fail to send it in on Belly Drum. You're talking about if Lax is turbo-drumming with a Beller (and Skarm doesn't also have a Beller on its team, otherwise they can just mirror the Bell as well). But yeah okay Skarm has one less get-out-of-Drum-free card.

 

[Bedschibaer]

Are there any other matchups that are changed by this?
Because if not i don't think this will be a very huge change to the metagame itself. Skarm is no 100% counter to drumlax anyways and there are several ways to break it. It was a 50% chance to outspeed after curse, now it's a 100% chance. Weakened or sleeping skarm is breakable by lax, which can be shown by countless theorycrafted scenarios. The impact of this is more important for individual matches. If the metagame changes there will be slightly more drumlaxes, but due to the low number of gsc matches this is barely any impact on the metagame, i don't think skarm is sinking in usage or viability because of that. Snorlax rising - lmoa.

 

[Jorgen]

Skarm is tremendously difficult for Drumlax to break on his own, even without Curse. Don't make it seem like Skarm is a shaky check to Drumlax, I mean Lax can beat it but holy hell is it difficult.

That said, I've had several games where I was tantalizingly close: the opponent had to rely on Cursing after waking up from Rest to survive the 999-attack BS 3HKO, and then winning the speed tie to continue. I've lost every scenario to win that game-winning speed tie :(

 

[Bedschibaer]

It wasn't really my intention to make skarm look like a shaky check. It's no 100% counter because fire blast drumlax exists and lovely kiss drumlax exists and is common. And as i said, the common scenarios where lax breaks a skarm without that (aka where the huge-factor of this bug comes into play) are when it's weakened or asleep (or parad, if you want to add that specifically) beforehand.

 

[M Dragon]

That's a load of tosh (even ignoring the fact that it's been implemented as a speedtie for ages now and there's nothing "reliable" about a speedtie).

X, come back!
Go! SKARMORY!
Enemy SNORLAX used BELLY DRUM!

Enemy SNORLAX used RETURN! (~47%)
SKARMORY used WHIRLWIND!
Enemy SNORLAX was blown away!

[let's give Snorlax the benefit of the doubt and assume Skarm is immediately forced out by whatever came in, and then Snorlax gets in again without taking damage]

X, come back!
Go! SKARMORY!

Scenario 1:

Enemy SNORLAX used BELLY DRUM!

SKARMORY used DRILL PECK!
Enemy SNORLAX fainted!

Scenario 2:

Enemy SNORLAX used REST!

SKARMORY used REST!

[Snorlax will wake up and use Belly Drum as Skarm comes in with 0 sleep turns left, Skarm will wake up and phaze as it gets hit bringing us right back to where we were at the second turn above]

Skarm doesn't actually need Curse to stop DrumLax unless it's last-Poke or you fail to send it in on Belly Drum. You're talking about if Lax is turbo-drumming with a Beller (and Skarm doesn't also have a Beller on its team, otherwise they can just mirror the Bell as well). But yeah okay Skarm has one less get-out-of-Drum-free card.

Point of this thread is that it has always been implemented that -1 speed Skarm is always faster than Snorlax.
Recently the speed tie thing was fixed, but ingame Snorlax is always faster.

There is a huge difference from Skarm being able to Curse without being outsped and Lax always outspeeding, meaning that it can break Skarm.
With old mechanics, Curse Skarm was a 100% counter to non LK/FB BDLax (unless para hax or well timed crit). Now Lax will win if other mon of the team (like Marowak) forces Skarm to rest and you go to an electric the turn Skarm rests, or if Skarm is weak.
In the scenarios you posted it is BDLax weakening Skarm, and that is not the point.

 

[magic9mushroom]

It wasn't really my intention to make skarm look like a shaky check. It's no 100% counter because fire blast drumlax exists and lovely kiss drumlax exists and is common. And as i said, the common scenarios where lax breaks a skarm without that (aka where the huge-factor of this bug comes into play) are when it's weakened or asleep (or parad, if you want to add that specifically) beforehand.

If Skarm is pre-paralysed, I don't think this bug is all that relevant. It'd need to both Curse and Rest in front of DrumLax to be affected.

FB DrumLax and Rest LK DrumLax can get past Skarm (though FB DrumLax has to FB it before Drumming to achieve much with FB), but a reasonably common partner for Skarmory is Misdreavus, which hard-walls them. The Belly Drum/STAB/EQ/LK set is the only one which can get past both.

Point of this thread is that it has always been implemented that -1 speed Skarm is always faster than Snorlax.
Recently the speed tie thing was fixed, but ingame Snorlax is always faster.

I believe PO has had them as a speed tie for 4 years? I don't think that's "always been implemented that -1 speed Skarm is always faster than Snorlax".

There is a huge difference from Skarm being able to Curse without being outsped and Lax always outspeeding, meaning that it can break Skarm.
With old mechanics, Curse Skarm was a 100% counter to non LK/FB BDLax. Now Lax will win if other mon of the team (like Marowak) forces Skarm to rest and you go to an electric the turn Skarm rests, or if Skarm is weak.
In the scenarios you posted it is BDLax weakening Skarm, and that is not the point.

Double-Edge DrumLax can break a 2-turns-sleeping Skarm pulling these shenanigans even under NBS mechanics, as can Body Slam DrumLax (30% to para on the turn it Curses). Return DrumLax also eventually wins because it still 4HKOs vs. the +1 Skarm and can just wait for a crit. That's... every variation of DrumLax there is. And obviously all of these had a 50% chance to defeat it in that situation for the last several years.

Not to mention that there are other ways to remove Skarm than "attack it with Marowak", or that a lot of stall teams run Suicune alongside Skarmory and can thus still defend against the combination.

As I said, CurseSkarm is slightly more vulnerable to DrumLax, making Curse on Skarm a bit less useful and DrumLax very marginally more useful (it's a situational improvement in one specific moveset-variation of one relatively-uncommon matchup - Skarm's usage isn't that high to begin with and switching to Skarm has never been the best way of beating DrumLax - that's "attack it").

 

[Royal Flush]

Was this really unknown? I'm pretty sure Borat told me this few years ago when I was hosting a GSC tourney here (back there NBS was still used and had the wrong spd on -1 skarm). He didn't explain all the details, though.

 

[magic9mushroom]

Was this really unknown? I'm pretty sure Borat told me this few years ago when I was hosting a GSC tourney here (back there NBS was still used and had the wrong spd on -1 skarm). He didn't explain all the details, though.

It was thought that -1 Skarm was supposed to be the same speed as Lax. NBS had -1 Skarm 1 point faster than Lax, which is probably what Borat told you was wrong.

Now, we have discovered that -1 Skarm is in fact supposed to be 1 point slower than Lax.

So it was indeed unknown (or rather, known wrongly).

 

[MoxieInfinite]

does this also mean steelix and -1 skarm doesnt tie, and skarm will always be able to phaze? i guess that could be a difference, though very minor as the scenario is quite rare.

 

[magic9mushroom]

does this also mean steelix and -1 skarm doesnt tie, and skarm will always be able to phaze? i guess that could be a difference, though very minor as the scenario is quite rare.

Yes. But CurseSkarm is not exactly a sweeper threat outside of some very specific endgame scenarios.

 

[M Dragon]

If Skarm is pre-paralysed, I don't think this bug is all that relevant. It'd need to both Curse and Rest in front of DrumLax to be affected.
FB DrumLax and Rest LK DrumLax can get past Skarm (though FB DrumLax has to FB it before Drumming to achieve much with FB), but a reasonably common partner for Skarmory is Misdreavus, which hard-walls them. The Belly Drum/STAB/EQ/LK set is the only one which can get past both.

There is a reason why Skarm + Ghost / Rock combos were popular, since they were a good way to stop bdlax in stall teams. Now that BDlax can break that core more easily, that is huge.

If Skarm is pre-paralysed, I don't think this bug is all that relevant. It'd need to both Curse and Rest in front of DrumLax to be affected.
I believe PO has had them as a speed tie for 4 years? I don't think that's "always been implemented that -1 speed Skarm is always faster than Snorlax".

No, much less time.
But that is irrelevant, since we have been playing GSC for a very long time and most stall teams were built assuming that -1 skarm was always faster than BDLax. That is the point of this thread.

Double-Edge DrumLax can break a 2-turns-sleeping Skarm pulling these shenanigans even under NBS mechanics, as can Body Slam DrumLax (30% to para on the turn it Curses). Return DrumLax also eventually wins because it still 4HKOs vs. the +1 Skarm and can just wait for a crit. That's... every variation of DrumLax there is. And obviously all of these had a 50% chance to defeat it in that situation for the last several years.

Not to mention that there are other ways to remove Skarm than "attack it with Marowak", or that a lot of stall teams run Suicune alongside Skarmory and can thus still defend against the combination.

Yes, BDLax had a chance of beating Skarm without FBlast or LK, but most of the times Skarm was the winner.
Marowak was just an example, you can lure and weaken skarm with a lot of things. Point is that with "new" GSC mechanics, BDLax has a greater chance of breaking Skarm especially if another mon can force Skarm to rest, and that affects team building.

 

[magic9mushroom]

No, much less time.
But that is irrelevant, since we have been playing GSC for a very long time and most stall teams were built assuming that -1 skarm was always faster than BDLax. That is the point of this thread.

Then either they've already been redesigned in the past several years or they've been abandoned.

Yes, BDLax had a chance of beating Skarm without FBlast or LK, but most of the times Skarm was the winner.
Marowak was just an example, you can lure and weaken skarm with a lot of things. Point is that with "new" GSC mechanics, BDLax has a greater chance of breaking Skarm especially if another mon can force Skarm to rest, and that affects team building.

Skarm still stops DrumLax as long as it doesn't have to stop something else as well. You are making a mountain out of a molehill.

Not to mention that not all Skarms on stall teams are CurseSkarm and this has no effect on other Skarms.

 

[magic9mushroom]

I know there's also issues with http://www.smogon.com/forums/threads/stat-reduction-bug.3520407/ which is genuinely interesting; whilst there's arguements over how important it is, it is competitively relevant (in GSC at least)

Oh, I agree it's relevant, and people will probably use Curse slightly less on Skarm (tbqh half the reason it's so common is already nothing more or less than "the Smogon analysis recommends it"; Toxic covers several threats CurseSkarm can't and ThiefSkarm is just generally great) because of it, but my point is that it's not a big deal from the DrumLax user's point of view because DrumLax has to get over so many hurdles to even get to the point where this matters (assuming one's opponent is even running Skarm to begin with).

(Curse's main niche on Skarm, before and after this discovery is stopping last-Poke EQ CurseLax on a stall team that doesn't have Miltank, Umbreon or Misdreavus. And that's totally unchanged.)

 

[Disaster Area]

Yeah that's cool just so long as it's accepted to make a difference is all that's important from the coding point of view ;)
carry on

 

[Joim]

I'm implementing today that fix for gens 1 and 2. So well, prepare the meta.