Preventing Speed Creep

FlareBlitz

Relaxed nature. Loves to eat.
is a Tiering Contributor Alumnusis a Top Contributor Alumnusis a Past SPL Champion
Well, that's kinda what I meant. Like if you want Machamp to outrun Skarm you give it enough speed in the analysis to outrun 4 SPE skarm and then say "skarm frequently runs more speed, if you want to try and outrun faster variants take some points out of ATK because the HP EVs are necessary to avoid a KO from starmie".
 

Great Sage

Banned deucer.
I didn't misunderstand. I was trying to say "why are you setting 4 Speed as the standard", when setting 0 Speed as the standard accomplishes the same thing and is significantly less arbitrary.
 

FlareBlitz

Relaxed nature. Loves to eat.
is a Tiering Contributor Alumnusis a Top Contributor Alumnusis a Past SPL Champion
Oh, I see. Let me explain.

Our analysis for Pokemon A has 252 HP/252 DEF/4 SPE. This is based on Guideline 2 in my post, which states that the best place to put an extra point is SPE.
We're trying to make Pokemon B outrun Pokemon A. Suppose this Pokemon needs 140 EVs to outrun 0 SPE Pokemon A, and 144 to outrun the 4 SPE one. In this case, we should use 144 SPE because otherwise Pokemon B would not outrun the Pokemon A set in our analysis!
Basically, I'm saying that we should maintain consistency. If we say "140 EVs outruns this pokemon" and someone goes to that Pokemon's analysis and finds out it doesn't....yeah. But at the same time, this prevents speed creep because putting 148 SPE is unnecessary when we note that "hey pokemon A often runs more speed than this, take some EVs out of attack if you want". I hope that makes sense.
 
bumping this. my thoughts on the 4/0 speed debate are that if we make 0 speed the standard, in order not to deceive readers intentionally we'll have to add "but x pokemon may commonly use 4 speed or more" to everything. i feel like the 4 extra evs are dumped in speed more often than not, which makes it more practical for us to use that as the baseline (this also is a reason why it's not just an arbitrary number to pick; 252/252/4 is a "standard" spread while 252/248/8 is not).

opinions from different people, especially ones who battle competitively on a high level, will help here. (c&c people, link those people here). for those people who don't want to read 10 pages of nerds arguing about shit, here's the problem. it's kind of stupid to have analyses add 4/8/12 evs to speed of low speed to beat similar low speed pokemon (metagross/empoleon/magnezone/skarmory/tyranitar, kind of) since that will never end. it makes more sense to limit these pokemon to beating minimum eved variants of the others, but we can't figure out whether "minimum" should be defined as 0 evs (since that is the literal minimum) or 4 evs (since that is oftentimes the practical minimum).

anyways, regardless of how we resolve that issue, we have to figure out what to do with the "no speed creep" rule. obviously we should discourage speed creep whenever we see it in analyses. but i think we should also do something like what QC did to current on-site analyses, and eliminate speed creep from those as well.

here is a non conclusive list of sets i believe are inflicted by speed creep (i didn't spend much time doing this so i am definitely missing a lot).

lead metagross + lead empoleon + spiker skarmory + cb tyranitar
many of azumarill's sets
wish cm jirachi

maybe after we generate a moderately sized list (which would still be incomplete) we can get some people to go through these and start fixing. i would be willing to be on such a team, obviously.
 

eric the espeon

maybe I just misunderstood
is a Forum Moderator Alumnusis a Researcher Alumnusis a Top CAP Contributor Alumnusis a Tiering Contributor Alumnusis a Top Contributor Alumnus
It'd be easy for a group of 3-4 people to go through all the analyses fix all the on site "speed creep" sets in a few days once we decide the only real remaining issue (4 speed or 0 speed standard). We've done larger sweeps before, and this should not be a particularly hard one.

@Phantom_IV: Sorry, I assumed that "None" meant None, not few.
 

firecape

This is the end...
is a Forum Moderator Alumnusis a Contributor Alumnusis a Smogon Media Contributor Alumnus
Well, Ubers in particular suffers from pretty bad Speed creep due to the large amount of Pokemon with the same base Speed. For instance, some Kyogre sets run enough to outpace Tyranitar, but then some Ho-Oh sets run enough to outspeed that.

All of these analysis have it in some form or another:

Giratina-O

Groudon

Ho-Oh

Dialga

Kyogre (this ones wierd, some have 4 HP, some have 4 Spe)
 
This is a pretty good example of what the metagame for pokemon has become. With shoddy and the stats it provides, along with the sheer number of users and matches per month, this game changes FAST. Meanwhile, you have smogon effectively putting pins in all the sets that work which is a rather static condition, as the EV situation points out.

There's no real way to mitigate this through a series of articles except for a discussion about that as proposed in the OP rather than giving out discrete numbers (which would mean constant editting of the articles). I like this idea, to say it in a short way.
 

FlareBlitz

Relaxed nature. Loves to eat.
is a Tiering Contributor Alumnusis a Top Contributor Alumnusis a Past SPL Champion
A lot of the semi-old UU analyses have speed creep, including things like 16 Spe for Azumarill and 28 Spe for Tangrowth. I don't envy whoever volunteers to go through and fix all that.

The most important thing we can do though is prevent new instances of speed creep. This is actually proving really difficult, for me at least, since I kind of skip over 16 Spe as the same as 4 Spe (due to the way I build teams). I think a fool-proof, if somewhat intensive, way to address this problem is to have all contributors explain the purpose of Spe EVs if they're running any amount besides 0, 4, or 252. That way if a contributor sees something like "running 12 EVs to outrun pokemon who are trying to outrun 4 Spe variants" they can immediately crack down on it.
 
Hi, sorry for bumping this thread, since its been a few days since the last post but here goes. I've always felt that speed creep is stupid and a pretty lame gimmick to get a suprise on your opponent at the sacrifice of a lesser stat or taking 4 EVs from here, a few from here etc just so that you can outpace something by suprise. Dandy. I think for the most part it should just be ignored unless its some metagame altering EV spread or something along those lines. While I realize my post carries little to no merit on the base of me not being QC, having a 2010 join date, and only have been playing serious competitive mons for less than one year, I still would like to voice my opinion. I want to use an example of speed creep gone wild. (Hypothetical, of course)

Standard Skarmory runs 16 Speed to outpace Adamant Rhyperior (granted this set is outdated to a degree as far as EVs go but its an example none the less and it popped off the top of my head). So Rhyperior starts running Jolly and only runs enough EVs in Speed to outpace Skarmory, and places the rest in HP. Skarmory then runs 20 Spe. Rhyperior runs 4 more Spe taking 4 more from HP. Skarmory now runs 24 Spe to outpace that Rhyperior. Eventually Rhyperior is running 252 Spe Jolly and Skarmory is running somewhere around 80 Spe to outpace that that it now no longer serves its purpose to be as defensive a pokemon while Rhyperior's set has changed entirely and it is no longer the heavy hitter it used to be, thanks to speed creep. Clearly this is not beneficial to either pokemon, although theoretically via natural order of things, people would go back to the standard since they want Rhyperior to hit harder and Skarmory to be more defensive to take this newfound older threat on. Yes, this situtation is rather extreme and albeit not very likely to ever come to fruition with any pokemon x that is meant to outspeed pokemon y.

With Breloom recently becoming a big threat and very popular due to its ability to slay the ladder, the previous example may not be very far behind from happening. Breloom is in a pivotal speed tier-while it is not slow, it can be fast enough to outpace some important stuff that can ruin its day. It is given 244 Spe to outpace Defensive Suicune, Rotom-A and BaitTar. The latter of which we can realisticlly ignore since no Tyranitar set has any business staying in. Rotom-A really doesn't fear much from Breloom so we can ignore that guy for the most part as well. Suicune however is important. Defensive Suicune tends to run 0 Spe as its goal is to be as fast as possible. But what is to stop that "oh-so-pivotal" 4 EVs from being moved to speed? Nothing. The risk of that is literally 0. So now Breloom users face a choice. Run 4 more Speed EVs at the sacrifice of perfect Poison Heal (not a wise choice) or take out 4 from attack at, once again, 0 risk. Obviously that would be a wise choice. And so begins the speed creep. Eventually both pokemon stop worrying about speed creep because Suicune will eventually win out thanks to higher base speed and it caps at some stupid shit or reverts back to the old set.

I'm not saying its inevetable and stuff, or that it will be breaking the game. Yes I realize this is a stretch and all that jazz. Don't tell me it is. I know. But I think we should focus more on having sets do what they should do.

tl;dr

Sets should have a specific aim. People will always run +4 Spe or a +Spe nature where possible to outspeed shit just to give them an edge up, but that doesn't mean we have to include that in every analysis. Its a given fact. It happens but until the metagame begins to shift around those changes they should probably just be ignored. Scizor isn't starting a huge trend to run 12 EVs in Spe just to outspeed 8 Spe Scizor (probably because you would rather want lower speed because the only thing you can really damage it with is Superpower and you want to take advantage of the opponents Def drop but whatever). Sorry for the long post and bump. I'm ready to be attacked for all my outlandish statements and my Breloom example. So yeah. There are some good examples for speed creep (having Machamp outspeed Blissey etc) but most of it is just shuffling around EVs.

tl;dr of the tl;dr

Speed creep is natural and eventually it reverts itself. Ignore it imo.

Basically I agree with whistle and I like FlareBlitz's suggestions.
 
well those of you nerds who live on scms may have noticed some changes

posting here for paperwork purposes and cause rising_dusk wants to check to make sure i don't mess up

empoleon
fixed subpetaya evs -- 12 hp 252 satk 244 speed modest (post #49 in empoleon update)
changed lead to 84 speed evs to hit 177, 1 more than 176
deleted evs section and moved contents to set comments of the two agility sets
put in [AC] tags in the agility sets as a quick aesthetics fix

breloom
subseed speed creep with skarmory -- moved all evs into hp/defenses

celebi
tinkerbell needed four more evs because primary option was hp fire
offensive and offensive cm speed creep with pokemon that beat base 95+, but 32 hp is a lo number so the last 4 evs are left in speed cause 4 spdef or def on offensive sets looks stupid
perish song only needs 32 speed for jolly tyranitar

dragonite
mix set speed lowered 8 evs and put in satk instead

electivire
mix set speed lowered 4 evs and put in atk instead

gyarados
restalk speed evs moved to sdef
cb speed bumped down to 244, added mention of adding speed for celebi

heatran
changed torment speed by 1 to 244 (jolly tyranitar)
changed heatrap speed by 1 to 230 (adamant scizor)

jirachi
wish + cm speed set to 244 from 245

infernape
changed scarf evs from 24 atk / 252 satk / 232 speed rash to 64 atk / 252 satk / 192 speed naive and modified SC accordingly

metagross
lead (others 70s) set to 4 speed -- added empoleon/skarmory mention
cb (other 70s) set to 4 speed
mixed attacker (207 magnezone) explanation changed since that seems like the real reasoning... wtf 207 magnezone?

roserade
lowered subseed speed to 244 from 246

skarmory
removed 16 speed and put them in defense -- rhyperior is not a good argument

swampert
changed cb to 177 speed down from 181 and maxed attack, rest in hp -- added skarm mention

tyranitar
choice band (skarm) down to 177 speed -- added skarm mention
boah (blissey and machamp) down to 147 speed -- added bliss/machamp mention
mixed (scizor/base 70s) -- added "scizor killers" mention

vaporeon
specs only needs 44 evs to beat 0 speed skarm, not 48

zapdos
changed baton pass speed to 32 from 36, added mention for defensive celebi and a warning about the dangers of increasing speed too much

doing uu pokemon tonight

edit:

aggron
changed subpunch spread to just outpace base 60s

azumarill
made all of azumarill's spread 4 speed -- added mention of other base 50s in optional changes

claydol
calm mind was 4 evs too fast (it hit 213 not 212)

donphan
qc changed rapid spin evs to max hp so i did that (not a speed creep thing but done in the same sweep so it's here)

hariyama
changed choice band to 4 speed evs from 8
changed lead to 44 evs for omastar not 56

mesprit
deslashed hp ground from trick on special set
labeled ou sets "OU x"

miltank
moved 32 speed from utility to defense that's an ou spread -- the 243-245 tier is nonexistent in uu

rhyperior
redid standard set evs they actually make sense now (and they actually do what they claim they do... you can't sub five times with the old evs lol)
speed creep on uu standard, uu cb, ou subperior

spiritomb
lead set had 4 evs too many

weezing
changed sdef set speed evs to sdef + defense
 

supermarth64

Here I stand in the light of day
is a Contributor Alumnus
Outrunning a min speed Pokemon is not considered Speed Creep. 84 Evs would be acceptable, and is already used for the OU CB set.
Didn't want to bring it up in that thread (here: http://www.smogon.com/forums/showthread.php?t=78668), so I brought it here.

Anyway, why wouldn't it be considered Speed creep? The defensive Pokemon that runs 0 Spe can just run 8 Spe and lose a very little amount of defensive capability in order to outspeed that Pokemon that tries to outspeed the defensive Pokemon in the first place, which is Speed creep as I see it.
 
In Ubers, speed creep has gone too far. Jibaku has been known to run 112 Speed Groudons, and when I run 36 Speed Groudon to beat Ho-oh I get informed that "invest more speed, you lose to 40 Speed Ho-oh". The same logic can be applied to anything.

Regarding the analyses, what I suggest is radical, but I feel it is necessary: remove all speed from the basic EVs and allocate them to the defenses of choice. Then write a note in the analysis giving specific defensive targets for the Pokemon to hit, and leaving it up to the reader to decide how much speed he wants to invest.
 

eric the espeon

maybe I just misunderstood
is a Forum Moderator Alumnusis a Researcher Alumnusis a Top CAP Contributor Alumnusis a Tiering Contributor Alumnusis a Top Contributor Alumnus
Didn't want to bring it up in that thread (here: http://www.smogon.com/forums/showthread.php?t=78668), so I brought it here.

Anyway, why wouldn't it be considered Speed creep? The defensive Pokemon that runs 0 Spe can just run 8 Spe and lose a very little amount of defensive capability in order to outspeed that Pokemon that tries to outspeed the defensive Pokemon in the first place, which is Speed creep as I see it.
The creep is the steady increase of speed, the competition in which both analyses can and will gradually raise their speed to try and beat the other. So long as we select a place to stop the creep (in this case, outspeeding 0 Spe spreads) it does not become a problem for the analyses. There is some degree of arbitrariness in deciding to cut it off at beating 0 speed spreads, however it seems by far the best place to stop since it is the lowest speed that lets you beat "some" of the set you are trying to beat. Extra investment beyond that required to beat 0 speed foes is "optional". It is up to the individual player to decide how large a drop in other stats is tolerable in order to outspeed more foes. The metagame will always have speed wars in which players invest a variable number of EVs in speed to beat sets aiming for a similar Speed. This is impossible to avoid and not harmful in any way. What this thread addresses is how the analyses should deal with it, since they cannot follow the day by day change in how many Pokemon run 8 or 16 EVs (especially without super specific stats, other than on PO I guess).

An added advantage to using "you may outspeed 0 speed spreads" and "you may run 4 Speed Evs" as guidelines together is that it puts the slow Pokemon running some speed to beat 0 Speed sets of other Pokemon on an even footing with those Pokemon who don't want to invest more than 4 Speed Evs. It means that, from an analysis perspective, they have the same Speed. In the metagame both will end up investing a little more than the analysis suggests a vast majority of the time when used by competent players, but as previously stated the analyses cannot follow this in any consistent, accurate, and useful way. Giving users the bare minimum investment and letting them make their own decision on how much extra to invest more useful.
 
If we're going to have speed creep rules, we should define the different "types" of Pokemon and what speed they're ideally going to run. Disclaimer: Yes, not all Pokemon will fit some role exactly, this is just in general.

Frail Sweeper: 252 Spe / 252 Atk
They aren't supposed to take hits, so they always run max Spe.
EX: Gengar, Jolteon

Balanced: Atk benchmark / Spe benchmark / rest HP/Def
This one's a little trickier, but this is one where Speed, offense and bulk are all utilized.
EX: BulkyGyara, Rotom-A, ChestoRest Kingdra

Bulky Offense: 252 Atk / minimal Spe / rest in HP
"Tanks", Pokemon that are supposed to take hits and hit hard back. Basically runs max Atk, hits a speed benchmark with like 20-80 EVs invested (or otherwise 0 Spe), and the rest dumped in HP.
EX: Rhyperior, Azumarill, Donphan (thinking UU here)

Walls: 252 HP / 252 Def
Pure defensive Pokemon. Doesn't run any speed unless it's fast enough to outrun a specific slower tier (8 Spe Milotic for example)
EX: Skarmory, Blissey, Weezing

With offensive Pokemon there isn't any problem, since they're all running max Spe or just enough to outrun other max Spe Pokemon. The problem basically comes up with the last two types, where everyone is running the bare minimum Speed. Basically, the rules as I see them are below:

1. Walls are required to run minimum speed, unless they have high enough base Spe to beat max Spe slower Pokemon.

Any wall with less than something like 70 Base Spe should always be running minimum Speed. The exceptions are with things like Milotic and Uxie, who are naturally fast and only need minimal investment for things like Max Spe Adamant Aggron.

2. Bulky offense Pokemon are allowed to run enough to outrun minimum Speed walls.

Best example of this is with Donphan running 84 Spe for minimum speed Clefable and Weezing. Since the Pokemon is offensive in nature, a few Speed EVs are more beneficial to it than to walls.

3. Pokemon can only run 1 point above a speed tier, not 2.

This means that we can't play the game of "Lanturn can run x EVs to outspeed Milotic by 1 that outspeeds Torterra by 1." Drop Lanturn's Speed down to Torterra +1, then say something like "The EVs outspeed Adamant Torterra, but many Milotic also hit that same speed stat... hint hint..."
 

eric the espeon

maybe I just misunderstood
is a Forum Moderator Alumnusis a Researcher Alumnusis a Top CAP Contributor Alumnusis a Tiering Contributor Alumnusis a Top Contributor Alumnus
I'm not sure how much separating Pokemon into categories like "frail sweeper" and "bulky offense" helps. There is rarely a clear distinction, and many examples which blur the line. It makes the most sense to me to just use the best spread possible, and follow the current Anti Speed Creep guidelines (you must aim to beat a a specific constant spread, either a max Speed spread or a min Speed spread). This does already cover your proposed guidelines (1: Walls can run 4 Speed to beat min Speed walls, this is generally the best base spread. 2: Bulky offensive Pokemon can run enough to beat min speed targets if they are important enough. 3: Only spreads which beat the target by one are allowed, no adding an extra one or two.).
 

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

Top