Preventing Speed Creep

[eric the espeon]

Something that's come up a few times is a set adds enough speed to outpace a set which runs enough Speed to outrun the first set and wants to outrun the first. The loss of defenses/attacks from 4 EVs is minimal, so logically it is better to move 4/8 EVs on the first set to Speed, the gain from going first against a common opponent is much greater than the loss. The problem comes when both Pokemon are capable of investing just that little bit more, and bulk is important for both. Since so many people simply follow the analysis what we put in changes what is the best option. In this case, both sets should continually add a small number of EVs to push their own speed above the set they are competing with. There is no good way to tell where to stop this speed creep in most situations, at some point it is better to just drop your speed back to 0 since the loss of defenses from speed investment is too great. Or what about dropping it back to 4, to beat those who dropped it to 0. And then the cycle starts again.

I propose that for common sets only we disallow spreads specifically designed to beat Pokemon which can add a small number of EVs to speed easily, and start a speed creep cycle. This means no 8 Speed Scizor, no "Metagross lead with enough speed to beat Empoleon lead", none of that. A short paragraph explaining the option to slightly raise Speed by any reasonable amount the player chooses would be included in analyses which need it. However, unusual Pokemon may still use these types of spreads since the more common Pokemon would be unlikely to react and raise their Speed to beat a rare threat.

For added justification, consider this: There is no final correct speed for this kind of set. The ideal speed depends on the metagame, which depends on the analyses. We can't tell people which spread is best because by doing so we make ourselves wrong, so the best resolution is to allow players to decide individually how much they are willing to take out of bulk to give them a better chance of outspeeding foes. This obviously does not apply to things like "enough speed to beat the + Speed base 90s", since they can't just add 4 EVs.

[Dubulous]

Seems logical to me; perhaps it wouldn't hurt to mention moving some EVs to speed to outspeed some "speed ties" in Additional Comments or something like that.

[Zystral]

I think if Max Speed isn't being used, then just have "Enough to outspeed certain other threats."
Example; if something has Base 95, it will need 216 to outspeed Lucario, but it can run more to outspeed other things, but that should be mentioned in AC.
I don't think continually changing the EV spread is a good idea, just having a single, minimum thread with a sentence saying "more speed can be taken from HP to make it outpace X."

[supermarth64]

I always thought that the "8 Spe" EVs were for entry hazard damage like Scizor and Machamp. However, it may just be those two Pokemon that come off the top of my head.

[eric the espeon]

The taking EVs out of HP is indeed often to reduce entry hazard damage, however placing them in Speed does nothing to change entry hazard damage. Those EVs would be placed in whichever defense is most appropriate.

[cim]

Speed creep becomes a problem when a damage calc tips a 2HKO or something. This is completely objective and up to the writer to find out.

Speed creep has been a potential issue for years and has never caused a problem, I don't see why we need hard and fast rules to deal with it now, nor do I really want to restrict analysis writers in such a way.

[eric the espeon]

Losing defenses is always a problem, you can't assume max HP. The whole philosophy of "you must provide specific calculations to prove extra stats are helpful" is fundamentally flawed. The issue of speed creep has come up several times, and there has been no decision taken on it. This means there are no guidelines, which for this kind of issue is a fairly large oversight. I would expect almost every single writer would be more than happy to follow whatever guidelines are put in place, and would prefer to have some rather than make it up as they go along with the same debate taking place in multiple threads and taking them off topic. If a writer can argue that extra speed is indeed beneficial to the analysis, and it won't simply cause the set it aims to outspeed to add 8 and render the addition useless, then of course it will be allowed.

[Rising_Dusk]

I agree with the OP that we should only impose Speed EVs when the following can occur:

• You always outspeed Pokemon in a given Speed tier. (224 Speed Timid Suicune always outspeeds 252 Speed Jolly Gyarados, for instance, so it should stay)
• You outspeed minimum Speed variants of things. For instance, 4 Speed Suicune with Roar is handy to outspeed 0 Speed Suicune. No more Speed would be used on either set to "beat" the other, only to beat the baseline.
• You are running max Speed EVs in order to be as fast as possible.

I feel that these are the only three cases where running Speed EVs in analyses are meaningful and should be kept. In every other case, the EVs should be placed elsewhere.

[Mekkah**]

0 EVs is the absolute minimum
4 EVs is outpacing the absolute minimum, and the leftover of a 252/252/4 spread
8 EVs is the leftover of a 248 hp/252/8 spread
(12 EVs outpaces those)

I think you can reason this is where you stop outspeeding "hard" values and the slippery slope starts. 8 speed EVs is the maximum amount of speed a Pokémon carries regardless of speed ties within a bracket.

[callforjudgement]

What about cases where you're trying to outspeed a particular other species of Pokémon trying to outspeed a third? I'm thinking of TyraniBoah using 28 Spe to outspeed minspeed Blissey (sensible, and allowed by this rule), so what happens if a third Pokémon runs just enough speed to outspeed a Boah? I'm having difficulty working out if that's speed creep or not.

[whistle]

well I've made two posts in different update threads about this already so my point of view should be apparent. "Speed Creep" (good name btw lol) is not an empirically damaging issue, but there is definitely a worst-case infinite regression scenario that is theoretically possible. in most cases, theoretical but not empirical problems don't really warrant action. but in this case, I think the warrant for the theoretical issue existing is also a warrant for why bumping Speed is silly, unnecessary, and counterproductive (ete post #7).

in terms of how to resolve this, I agree with Rising_Dusk's proposal. however, I would modify his second point and replace it with Mekkah's "definition" which makes a bit more sense operationally.

cfj - yes. that is no different than, say, using 216 Speed on Gliscor to beat Lucario and having Jirachi come along and think "hmm I wanna beat Lucario [Blissey] but I also wanna beat those Gliscor who want to beat Lucario [Boah]". in turn, Gliscor would bump up its Speed as well since it is nice to be faster than Jirachi until both Pokemon are running 252 Jolly.

[eric the espeon]

I'm not sure we should even have those "leftover" 4/8 evs placed into speed. They can always be put to good use in one of the defenses, or elsewhere, so are not true leftovers. Even moving those 4 remaining Evs to Speed encourages the start of trying t get one up. Especially when almost everything puts 4 more than the baseline, and the new baseline becomes 4 speed. Outspeeding min speed variants of Pokemon is often a good target, but if you have the same base speed as that Pokemon I don't think it is best to suggest even a 4 Speed investment. Other than that, I agree with RD's summary.

[cim]

Originally Posted by Fat eric the espeon

I'm not sure we should even have those "leftover" 4/8 evs placed into speed. They can always be put to good use in one of the defenses, or elsewhere, so are not true leftovers.

Speed is the most important stat in Pokemon. 4 EVs in a Defense will matter a statistically insignificant amount of the time, while 4 EVs in Speed, even if it matters in 5% of battles, is significantly more likely to come into play and significantly more important. I think you're massively overstating what 12 EVs does to a non-Speed stat.

[RaikouLover]

• You always outspeed Pokemon in a given Speed tier. (224 Speed Timid Suicune always outspeeds 252 Speed Jolly Gyarados, for instance, so it should stay)
• You outspeed minimum Speed variants of things. For instance, 4 Speed Suicune with Roar is handy to outspeed 0 Speed Suicune. No more Speed would be used on either set to "beat" the other, only to beat the baseline.
• You are running max Speed EVs in order to be as fast as possible.

I agree with everything here. Also, I have a problem with the speed tie things. For example that same offensive Suicune only needs a spread of 32 / 252 / 224 Speed to accomplish what it is supposed to outspeed. However, every offensive set now is " 252 / 252 / 4" to tie with other Pokemon X. While I believe this should be mentioned in AC for offensive sets, there does come to a point when we are slacking on efficiency for the simple "may as well potentiall tie other Pokemon X with the same speed if they are running the indentical same set" nonsense that started. I'll give an example:

The Kingdra special Rain Dance set lists a 4 / 252 / 252 Spe citing types with opposing Kingdra is gamebreaking. While this is true, the original spread used a good HP investment (92 HP or 23 HP points!) because you only need to outspeed Scarf Jirachi / Flygon. Using 4 / 252 / 252 means you will still lose 50% of time to every other Kingdra (other than Specs, which you will always lose to), which in the grand scheme of things Kingdra is used in something like what? 5% of battles? So this scenario becomes much less efficient for the overall use of the set especially considering that Kingdra's lower base 75 HP and solid 95 defenses mean that any HP investment augments its defenses nicely. Just saying that stuff like that should be re-evaluated because its an endless "max speed to tie with other pokemon running the identical same set," rather than outspeeding what you need to.

It is also a consistency issue. As if you look up to the Dragon Dance + 3 Attacks, the same 252 / 252 / 4 rule is not applied here as it is 252 / 232 / 24, where it should be max to "tie" with other Kingdra for the same reason any other set uses that logic.

Speed is the most important stat in Pokemon. 4 EVs in a Defense will matter a statistically insignificant amount of the time, while 4 EVs in Speed, even if it matters in 5% of battles, is significantly more likely to come into play and significantly more important. I think you're massively overstating what 12 EVs does to a non-Speed stat.

While this is true in Pokemon, from an analysis standpoint it is inefficient because the very nature of speed creep starts here where 178 > 177 > 176 deal.

[whistle]

RL if I'm reading your post correctly I think you're addressing a separate but related issue. you're saying that currently, we have lots of analyses that max Speed just for a vague "speed tie", and that this is a bad idea -- right?

if that's the case, I don't think you're referring to "Speed Creep" (at least in the sense that we are); it's more of a subjective judgment about whether opposing Pokemon X are dangerous enough to warrant a Speed max. example: I agree with your analysis about Special RD Kingdra and why it should have 92 HP EVs. but I think almost everyone would agree that MixMence should run +Speed and 252 Speed EVs, partly (mostly) because of the tie on opposing base 100s. whether or not to maximize Speed from the next tier below it is a call that should be made on an individual Pokemon basis by QC.

the reason I think that is distinct from Speed Creep is mainly because this is a large jump that does not have the potential to increase further - there is a brightline, and that brightline makes the jump much less arbitrary. the first recent example of Speed Creep was in the RestTalk Weezing analysis where it was originally suggested that Weezing run 16 Speed to outrun random bootleg Azumarill who use 88 Speed deliberately to beat Weezing.

while the use of 4 EVs etc in Speed is "technically" Speed Creep, I think we have to include them. pretty much everyone puts excess EVs in Speed, and there is a point at which eliminating Speed Creep means we start writing analyses that are misleading and not representative of what players use (I'm pretty sure that no one puts 4 EVs in Speed because of analysis X; they do it because they want to win ties in case their opponent doesn't). same with 8 EVs for things like Scizor.

[jumpluff*]

Originally Posted by Fat Chris is me

Speed is the most important stat in Pokemon. 4 EVs in a Defense will matter a statistically insignificant amount of the time, while 4 EVs in Speed, even if it matters in 5% of battles, is significantly more likely to come into play and significantly more important. I think you're massively overstating what 12 EVs does to a non-Speed stat.

This. While speed creep should be avoided since it results in arbitrary investments and it should be up to the user whether they want to forfeit some EVs in other stats to outspeed random 16 Spe Pokémon, it's common sense to put the 4 in Speed, because that's where it'll matter most.

ETA: So basically, what I'm saying is I agree with whistle. Sorry, I missed your last paragraph.

[Rising_Dusk]

I think we all agree that Speed is the most important stat in the game. Furthermore, I think we all agree to some extent that leftover EVs should indeed be placed into Speed for offensive sets. What I think we need to discuss right now is defensive sets like CroCune and how leftover EVs should be dealt with in that case.

The current on-site spread for CroCune (my example) is 252 HP / 252 Def / 4 SpD. Those last 4 EVs could be placed into Speed, but is placing those 4 EVs into Speed necessary for such a defensive set? Should these 4 EVs only be placed in Speed for offensive, but slow, sets, or should we instead simply "dump all extra EVs always and forever into Speed"? I personally support a distinction between defensive and offensive sets and Pokemon, where Pokemon like Machamp would invest leftover EVs into Speed, whereas defensive Suicune sets would invest them into a defensive stat.

What's your all's take on that?

[cim]

I think that kind of discussion can happen where it always has; in the threads for that set or Pokemon's revamps.

[Phantom_IV]

Agreed with Chris is me. If it comes down to individual cases, let the writers be the ones to decide.

[eric the espeon]

while the use of 4 EVs etc in Speed is "technically" Speed Creep, I think we have to include them. pretty much everyone puts excess EVs in Speed, and there is a point at which eliminating Speed Creep means we start writing analyses that are misleading and not representative of what players use (I'm pretty sure that no one puts 4 EVs in Speed because of analysis X; they do it because they want to win ties in case their opponent doesn't). same with 8 EVs for things like Scizor.

I dispute the point underlined:
| Scizor | Speed EV | None | 54.2 |
| Tyranitar | Speed EV | None | 20.8 |
| Metagross | Speed EV | None | 36.4 |
| Swampert | Speed EV | None | 90.9 |
| Blissey | Speed EV | None | 98.5 |
| Machamp | Speed EV | None | 64.8 |
| Gliscor | Speed EV | None | 11.1 |
| Skarmory | Speed EV | None | 43.1 |
| Vaporeon | Speed EV | None | 38.6 |
Those are all in the top 20, so a significant majority of people with the option to throw 4 "spare" (they can be put to use elsewhere so spare is not entirely accurate) into Speed often decline it, and place those EVs elsewhere. I do not think placing 4 EVs into Speed necessarily makes the analyses more representative. Otherwise, I agree with whistle's post.

There is a significant difference between, say, Machamp investing enough speed to beat 0 Speed Skarmory and Skarmory investing enough to beat min Speed base 70s. The key difference is that for the analysis, you can easily put in Skarm a small mention of "add a little speed if you want", however the difference between 4 Speed and 128 Speed for Machamp is huge. Machamp in reality may well want slightly more than 128 Speed, in order to beat more than the already fairly common min Speed Skarm, but again this "invest a bit more than 128" can be mentioned. 128 would be the basic spread for a Machamp set for which outspeeding base 70s is very important, and users would build from it. However, the base spread for a Pokemon which is investing 98.4% of it's EVs in non Speed stats (so clearly values defenses) is going to be zero Speed. This applies whether the set is classed as offensive or not (itself a imprecisely defined distinction).

chris, I think you are missing the point in both of your most recent posts. I am not arguing that putting 4 EVs into a non-Speed stat makes the set more effective, but that there is not a consistent non-arbitrary way to prevent speed creep if we allow it (or at the very least, not one outlined so far). Putting the discussion into the individual analyses threads without any guidelines will simply mean wasted effort as the same arguments are rehashed every time this comes up, and quite possibly lead to an inconsistent policy since a few writers will have different views from the majority. And, of course, there will be discussion in the individual threads. All I am setting up here is a framework which will allow those discussions to be resolved efficiently and consistently.

Spreads which outspeed max speed sets and those that max Speed to tie at worst with Pokemon in their own Speed tier seem to be universally acceptable. Spreads which set out to beat other spreads which can simply add a small number of EVs to Speed will not be acceptable. Spreads which beat opposing minimum Speed variants of other sets are the exception and are fine, so long as they are not set to beat a Pokemon which has the same base Speed as you (since this is no different from saying "220 Speed Gliscor to beat 216 Speed Gliscor" if you think about it). The last point is somewhat debatable, but the way I see it this conclusion is the only way to remain both consistent (it does not seem consistent to allow 4 Speed Suicune to beat 0 Speed Suicune, but not 188 Speed Gyarados to beat 184 Speed Gyarados when the value of those EVs outside of Speed is roughly equal) and useful (allowing Speed creep would lead to an impossible amount of nonsense, as I'm sure everyone reading this can see) across the analyses.

[cim]

A significant majority of players also suck.

I'm not missing the point. Speed creep has never been a problem and is often the smart thing to do, so I see no need to make overarching, constraining rules on how to EV sets until it actually does become a problem, as I guarantee the rule right now would be more of a hassle than a help to set writers. I mean, if you're talking leadgross versus leademp obviously you can't stack forever, but I don't see a rule that would make it difficult to even get "4 spe" on an analysis being a productive "solution".

Painting in wide brush strokes will just lead to more problems and tons of "exceptions". What's so wrong with a case by case approach?

[Great Sage**]

I would much prefer that all Pokemon are assigned minimum Speed necessary (i.e. if Starmie needs to outrun Gengar, then it gets Timid and 216 EVs in Speed and no more), and for there to be a universal general understanding that it is at times useful to drop a few EVs into Speed. The 0<4<8<12... game is something for the teambuilders to play with, and does not need to be present in the analysis. This is not some theoretical slippery slope; it's something that has happened and currently exists onsite, and calling for it to stop at any point other than minimum Speed necessary is self-defeating and arbitrary.

Chris is me, the problem isn't whether Speed creep is a smart thing to do; I think most would concur that it is often very useful to put a few extra EVs into Speed. However, representing Speed creep in the analyses is absurd. I'm not sure what you think is so complicated about mandating minimum necessary Speed. If anything, I would think that it would reduce complexity, because there is no more ambiguity about whether Pokemon X's EV spread should be officially recorded with 0 Spe, 4 Spe, 8 Spe, or some other number.

[Phantom_IV]

I dispute the point underlined:
| Scizor | Speed EV | None | 54.2 |
| Tyranitar | Speed EV | None | 20.8 |
| Metagross | Speed EV | None | 36.4 |
| Swampert | Speed EV | None | 90.9 |
| Blissey | Speed EV | None | 98.5 |
| Machamp | Speed EV | None | 64.8 |
| Gliscor | Speed EV | None | 11.1 |
| Skarmory | Speed EV | None | 43.1 |
| Vaporeon | Speed EV | None | 38.6 |
Those are all in the top 20, so a significant majority of people with the option to throw 4 "spare" (they can be put to use elsewhere so spare is not entirely accurate) into Speed often decline it, and place those EVs elsewhere. I do not think placing 4 EVs into Speed necessarily makes the analyses more representative. Otherwise, I agree with whistle's post.

Correct me if I'm wrong, but don't the monthly stats list "None" for some small investments? Less than 12 EVs? Something like that?

Reiterating that the analyses should reflect the state of the metagame, and players DO drop their excess EVs in Speed.

Edit: Checked. None simply means less less than 10 EVs, so stating that nearly 1/2 of Scizor are 0 speed is grossly misleading.

[FlareBlitz]

There are a few simple guidelines we can put into place to cut down on speed creep.

Speed investment is done for one of three reasons:

1) A Pokemon needs to outrun another Pokemon who typically has maximum speed investment (i.e. Gliscor needs to outrun Lucario).

In this case we add ONLY enough speed to outrun the Pokemon. Seems pretty obvious and straightforward, but you'll often see (and I myself have written this, even as recently as in the CAP stat spread thread) "invest x in speed to beat y and other pokemon trying to beat y". That should go away as far as the analysis is concerned.

2) Pokemon needs to outrun another Pokemon who typically has minimum speed investment (i.e. Scizor versus other Scizor).

In this case, 4 SPE EVs are the most valid. Why? Because those 4 EVs in speed let you beat no-investment members of the same speed group 100% of the time. As you can tell from the stats posted above, quite a few Pokemon don't run any speed at all, so those 4 EVs are way more valuable in speed than as an 0.01% increase defense or special defense or something.

3) Pokemon needs to outrun another Pokemon who doesn't max speed but often doesn't have minimum investment either (i.e. Machamp needs to outrun Skarmory).

This is probably going to be the trickiest kind, and probably where speed creep sets in the most. We can't just pretend the latter Pokemon always runs min speed if it simply doesn't do so a majority of the time, but we also can't arbitrarily say "run 60 speed on [this pokemon] to beat 36 speed [that pokemon]".
I think the best solution in cases like this will be to run enough speed to outpace 4 speed versions of the latter Pokemon while noting in the analysis that "this pokemon often runs more than 4 speed, if you want to add more speed you can take it from [this stat]". That way people who want to run their own amount of speed can do so while taking it from the stat where, as determined by us, it will have the least impact.

[Great Sage**]

FlareBlitz, I agree entirely with your first point. I can also agree with your second point in the case of 252/252/4 spreads. However, the third seems unnecessary. Instead of noting that a Pokemon may have more Speed than whatever amount the analysis's spread adjusts for, why not just officially record the minimum Speed necessary and then in the same amount of space note that "Pokemon X often runs more Speed, so you might want to adjust your Speed accordingly".