Policy Review The Over Tuning of CAP Creations


is a Community Contributoris a CAP Contributoris a Forum Moderator Alumnus
Approved by snake_rattler

The past generation of CAP processes have all led to some spectacular Pokémon. Our first CAP of this generation, has been a smashing success as well. But as of recently, a pattern has started to rear its ugly head, and with CAP 28 now arriving at the Concept Poll, now it is pertinent to address this pattern as what it is, a potential issue that could be detrimental to CAP in the long term if left unaddressed. Not a month after creation, Astrolotl has not only cemented itself as an S rank threat, but is currently undergoing a nerfing process. Equilibra, the CAP before, is also currently an S rank Pokémon, and has not only been nerfed last generation, but is currently next line for a second batch of nerfs! I believe this is a broader issue than these two alone, and that as of late, CAP creations have become too over tuned to operate in the metagame.

Less extreme examples include Jumbao, who while not banworthy, is firmly sitting in the S rank in the USM VR, and prior to that Pajantom was universally agreed upon by the community to be an S rank threat, and only fell short of achieving it because VR took too long to update, and by the time it was updated, the metagame spiralled around Pursuit abusers due to how potent Pajantom (and other noteworthy Pokémon) had become. Quite surprisingly, the only CAP created since the start of Generation 7 that has not been either nominated for or attained S ranking are the celebration starter type CAPs, which were held back by naturally occurring pitfalls that are in Starter Pokémon. Now, what problems does constantly creating S rank worthy Pokémon have for our community? Quite a few!

Optics :Pajantom:

I have had discussions in the OU room months ago about CAP and the state of the metagame and also recently with a fellow colleague of mine, both of which involved people who are largely divorced from our metagame and our process, and these conversations all led to si.ilar conclusions - most of the CAPs seem unnecessarily power creeped, in some cases ridiculously so, and these creations serve as barriers of entry to both the metagame and processes, as they think these Pokemon are too powerful to play against and they don't want to participate in the creation of more super powerful S rank level Pokemon. I fear that these sentiments are not isolated, and that if they are carried by a large amount of users, this can be damaging to our brand and to the growth of both of these equally important sections of CAP.

Metagame :Jumbao:

To expand upon my metagame issue, I don't think that introducing a metagame defining Pokemon every 2 or 3 months will lead to a healthy metagame in the long term, as this places a barrier of entry to new players, as I said, who to jump into CAP especially if they land into that awkward lull of the metagame that occurs two weeks or so before the debut of a brand new CAP, spending all that time to learn this new metagame, only to be faced with another Equilibra that turns previously established S ranked threats to B rank nobodies. This also puts a strain on our C&C, even more so than what Gamefreak has decided to throw upon us this generation. What's the point of dedicating a month of your time into forming the best analysis you could create, only for it to be at best in need of a rehaul or at worst be completely defunct. This coupled with the need to do nerf processes for brand new CAPs help detract man hours that we could be using for metagame development, or in Astrolotl's case, in both CAPPL and CAP 28 preparations.

Why CAPs are overtuned :Caribolt: :Smokomodo: :Snaelstrom:

We can definitely look towards our stats and movepools for our recent CAPs to get an idea of what may be causing the problems. It goes without saying that most CAPs BST wise are on another level in comparison to their OU contemporaries, as right now most OU Pokemon do not crack 535 BST, while all of the CAPs outside of the starters have all pushed at minimum 550 (Jumbao hit 576, and I do acknowledge my faults as I created this very spread). Equilibra is somehow the perfect Pokemon statwise, with an absurd amount of bulk coupled with immense attacking power, and a respectable speed stat for a Rapid Spinning tank, leaves it as a demanding presence in any match, being so difficult to account for both defensively and offensively. Even Pajantom in its peak seemed like an absolute menace, being able to comfortable switch in on weak special attacks thanks to excellent special bulk for an offensive break and trap a massive variety of stuff, or otherwise nuke teams with a Never Ending Nightmare with its sky high 133 base Attack Stat. Types for CAPs are all noticeably spectacular as well, but I feel this is more or less an unavoidable issue as we strive to create good competitive level Pokemon. Movepools are another big stinker here, as is the case for Astrolotl, our most recent CAP, has a multitude of powerful move options, and is largely one of the key factors in its eventual nerf. Equilibra too, received a movepool nerf last generation losing phazing and Toxic. A running joke both inside and outside of the CAP community is that "everything gets Rapid Spin", but in reality it comes across as a larger part of the issue, as if hazard removal is brought up, there is little one can do argue against said move's addition, as it is most likely "pro-concept". Its these little things that can pile up and lead towards CAPs having too much unnecessary tools that may or may not push it over the edge. Was Rapid Spin for example, really necessary to give to Equilibra in hindsight? It wasn't particularly pro concept, but we all agreed that hazard removal is good, and because of that, Equilibra should have access to removal. (Note I am not arguing for the removal of Rapid Spin from Equilibra's movepool, as at this point that would be an unnecessary removal and quite possibly do more harm than good.) All of these Pokémon are powerful, and I might say, too powerful for their own good.

Ways we can go about solving this :Equilibra:

These are easier to implement than what you may initially think, as I am glad to say that moves are being made to already address aspects of this issue. But once we recognize what is exactly the problem we can go about doing so with more efficiency. For stat spreads we should be much more open to less hyper optimized stats, and be encouraged to flesh out stats on the lower end of the spectrum. For moves, we can pay key attention in how we distribute potent tools like Knock Off, Rapid Spin and Toxic, in regards to competitive move submissions, focus less on overcrowding the Pokémon with as much variety as possible, and more on fully developing sets around key moves that define our project / process. At the end of the day, I think we need to able to ask ourselves: "How do we make sure CAPs aren't hyper optimized", and be able to have a fruitful discussion about how we go about lessening the chances of that from happening in the future.


In conclusion, the hyper optimization of CAP creations is an issue that should be discussed sooner rather than later, as with CAP 28 starting its process, we have little time to think about how these issues may still be running underneath our process.
Last edited:
More on the BST issue.

13Necturna547Xused to be 570
15Aurumoth600Xdescribed as pseudo-legendary
21Crucibelle540Mega Crucibelle has 640
25Caribolt535limited by framework
25Smokomodo534limited by framework
25Snalestrom535limited by framework

While 535 is the most common limit for normal Pokémon (with odd cases like Arcanine at 555, Florges at 551 and Togekiss at 545), I am extending this limit to 540 as I feel there are enough actual 540s (Snorlax, Gyarados, Milotic, Haxorus...) to relax this analysis a little.

Even then, more than half of the CAPs break the 540 barrier. This has been the case as far back as Syclant, but it has been consistently happening since Volkraken (bar the CAP25s which where limited by their framework and Crucibelle which is on the very edge but has a Mega Evolution).
I know distribution is more important than BST (and an analysis of average actual statistics could be more useful), but this should at least be indicative of something even just at an optics level.
Last edited:


used substitute
is a Community Contributoris a Top CAP Contributoris a Forum Moderator Alumnusis a Battle Simulator Moderator Alumnus
I have long been an advocate of not going overboard with stats, because, as stated already, most good OU Pokemon don't need incredible stats. We have long overstuffed our stats to the point that, unless a concept specifically pegs us as a wall or frail sweeper, it is pretty much assumed every mon we make will be bulky offense, because we will never make them weak, never make them frail, and never make them slow. I absolutely think could do with a shift in perspective on this, and I think that begins in leadership. I'm not sure we need any policy change here so much as we need leaders that are willing to set stat limits and slate stat spreads such that having a weakness (other than the off-attacking stat) is not seen as a crippling flaw.

That being said, it is also important to remember that 600 BST legendaries have long been a central part of the metagame. And, of course, legendary status is just a flavor thing. If a real Pokemon can be balanced with a 600 BST there is no reason a CAP can't. Our issue is far more about stat optimization than stat total. But as a project we are very much about optimization. So while I don't think there is necessarily a need to restrict BSTs, if we are going to keep optimizing like we always have, then we should do so with lower stat totals or we will keep having the same issues.

Now as far as movepools, I have long had issues with our current process, but only a couple of those issues touch on the current topic. That being said, I think we have a problem with the way moves are allowed. Since we made the movepool stage about movesets, I feel we have de-emphasized full movepool analysis. I feel like it has become too common for us to look at movesets in isolation, and then allow the moves on them, without thinking about the implications of the moves from multiple allowed movesets being used together in different, unlooked at sets. And yes I do think we also have been giving out a lot of moves far too freely. Especially moves like Toxic which many of us take for granted due to its long status as a universal TM. Every competitive move should be looked at and scrutinized, regardless of its distribution.

While I personally have other issues with movepools that I think could old be addressed by a total overhaul of the stage, for the issues pertinent to this thread, I think, once again the best solution is not necessarily any policy change, but simply a change in perspective of our leaders. We need a mindset closer to "allow only things that are needed" rather than one that is "allow anything that does not seem like an obvious issue."


Like ships in the night, you're passing me by
is a Site Content Manageris a Forum Moderator Alumnusis a CAP Contributor Alumnusis a Tiering Contributor Alumnusis a Contributor Alumnus
I would be cautious about any broad actions around power level given the creep that will be reintroduced through crown tundra. While we’ve certainly made some overly powerful CAPs before, we’ve often aimed too low as well. The fact that currently some CAPs are too powerful does not overly concern me.
Consider these points in a very general manner.

A point I noticed after my previous analysis: the recent high-BST period I noticed coincides with the XY/SM era which due to Mega Evolution had much higher BSTs in non-Uber tiers. Now that Mega Evolutions are not currently in the game I guess this competitive pressure might be lessened.

Even then, more than half of the CAPs break the 540 barrier.
I have long been an advocate of not going overboard with stats, because, as stated already, most good OU Pokemon don't need incredible stats. We have long overstuffed our stats to the point that, unless a concept specifically pegs us as a wall or frail sweeper, it is pretty much assumed every mon we make will be bulky offense, because we will never make them weak, never make them frail, and never make them slow.
Looking back, these points seem very related. Even looking back at Syclant, its spread is basically a stronger version of Infernape's and Lucario's.

More in general, I think this brings about an important point about what the CAP project is meant to be:
CAP mission statement said:
This is a project dedicated to exploring and understanding the competitive Pokemon metagame by designing, creating, and playtesting new Pokemon concepts.
As already noted actual Pokémon tend to have tighter constraints than CAPs and yet they still succeed: I think giving more consideration to said constraints would help the "exploring and understanding" part more than throwing in more useful moves and more stat points.
I am extremely glad that this has been brought up, as it has been something that has weighed on my mind for much of my time in this community. I have spoken about this as early as specs kerfluffle in its creation meta due to its powerful nature and the pressure it placed on the metagame at the time, even if that may seem farfetched to say now. Now, I am not entirely sure there needs to be any written changes in process. I believe this is purely a mental block that CAP must consciously address. I feel a lot of these overtunings are mental errors that aren't being properly addressed during the process. CAP is in a phase where it is afraid of undertuning due to failings in its past, or what it views as its failing, especially now that it is more metagame focused than ever before. It sees itself as failing if it does not make a strong contribution to the metagame in a way that significantly alters it. In this sense, to me, it has an innate fear that it must overcome. Overtuning is no fault of any one particular part of the process, but in fact the leadership, yet still not any one person. There is a fear of being underprepared, as CAP values its own creation abilities. CAP is extremely good at creation when it doesnt purposefully gimp itself, as it did with plasmanta's concept (at the same time, still, it can overcome poor concepts by the nature of its creation process and this exact overtuning in the face of failure - see jumbao), and it should better embrace its abilities to overcome the challenges it sets for itself. As G-Luke listed, CAP hasnt made anything short of excellent (in terms of competitive viability) in a long time, and I personally think the creation process should embrace that success, and be more forward with how it limits the tools it gives out at all points of the process. I think more conscious and decisive action at every point in the process can help the process help itself out of the rut, without any particular limitation placed on any point of the process.

In brief and on in a more specific sense, I do think that checks and counters stage needs some special attention. In my own personal opinion, it has been lackluster in terms of production. I think looking at astro's pre nerf ability to toss around most of its checks speaks to the current state of that stage in the process.


CAP 1v1 me IRL
is a Community Contributoris an Artist Alumnusis a Forum Moderator Alumnus
So I've been thinking about the stats issue in particular (not just because I'll be running CAP 28's stats) but I've started to think a bit more about the problem of BST. As noted by others above, CAPs generally have pretty high BSTs. While this doesn't always necessarily translate into viability, it is still a bit of a problem of optics to reiterate what others have said. Beyond this, however, due to a discussion on discord factoring in 'dump stats' or rather, stats that a Pokemon has and doesn't really make use of, I've been thinking a little bit more about which proportion of stats are used from some key OU mons to hit the 'representative' or 'actual' BST that you will notice in a game. Naturally this is still not a perfect solution because certain Pokemon such as walls don't really use their speed or attacking stats to a large extent. Nevertheless, I think it does reflect some very significant discrepancies.

A rough analysis of OU Pokemon in this way shows a range of 378 Base Stats used (Amoonguss discounting its attack of 85) up to 500 (Mew with either Attack or Sp. Atk ignored). The average of all of the OU Pokemons 'relative' base stat totals is roughly 450, which is in-line with relatively strong and efficient Pokemon like Alakazam, Rotom-Heat, Mandibuzz, Hippowdon and Excadrill. Interestingly both extremes of Pokemon on the relative BST spectrum roughly ignore 16-18% of their stats. The Pokemon with the lowest percentage of their stats being invalidated are Togekiss at 9%, Alakazam at 10%, Volcarona, Scizor, Rillaboom and Urshifu at basically 11%. The highest relative base stat total of these more efficiently distributed Pokemon is Volcarona at 490, and Urshifu a close second at 487.

CAPs on the other hand, demonstrate much more elevated relative base stats, with the lowest being Syclant and Snaelstrom at 441 (ignoring their Special Attack and Attack respectively), going all the way up to 513 Used (Jumbao ignoring its dump attack). The average is also significantly higher at 475, a solid 25 higher than in OU. In addition, CAPs are generally far more efficient when it comes to their dump stats too, with Pokemon like Malaconda, Equilibra, Volkraken, Mollux and Pajatom only losing out on between 8% and 9% of their stats. Of this group Equilibra and Volkraken are shockingly efficient, maintaining a relative BST of 509 and 510 respectively, far beyond comparable Pokemon in OU. Syclant is again the most extreme example, losing roughly 20% of its BST in its conversion, but that's not too surprising considering it has equal offenses and only runs one at a time, being comparable to Dragapult's drop percentage in OU.

As I've said it's not a perfect metric, especially when it comes to factoring in Pokemon such as Zeraora that do in fact use all of their stats in a single battle as it runs Volt Switch, but nevertheless, I think it illustrates that going forward, participants for the stats stages in particular should be a tad more conservative when it comes to creating totals, or at the very least, making sure that their 'relative' base stat total isn't too high. Not to shamelessly self-plug, but I think my spread for Astrolotl is a good example of a spread that looks overtuned on paper, with a BST of 560, but due to its SpA going unused most matches (or its attack if people choose to run full special) it has a relative BST of 468 or 452 which is 'below average' for CAP, and around the level of Chansey, Magnezone and Rillaboom in OU. Although stats aren't always the full picture, I think comparing a Pokemon like Torkoal which sees OU play with a relative BST of 385 solely because of Drought, and Jumbao which has the highest relative BST in the CAP tier, a proportional loss of only 11% of its BST which is torwards the low end in OU, as well as the same ability, demonstrates the problem with stat inflation in CAP creations quite clearly.

Overall what I hope this reveals, or at the very least quantifies for those who have their suspicions, is that not only are CAP Pokemon's BST inflated, but they are in general also far more stat efficient than comparable Pokemon in OU with far more drastic dump stats leading to CAPs maintaining a higher relative BST. If anyone is interested in the full numbers, I link the pastebin here where I've listed the comparisons between BST and relative BST for each Pokemon in the OU and CAP tiers.


I am the Scientist now
is a Site Content Manageris a Forum Moderatoris a Smogon Discord Contributoris a Top CAP Contributoris a Tiering Contributoris a Contributor to Smogonis a member of the Battle Simulator Staff
I do think the worry about BST here is a bit well, worrisome, as one of my gripes with CAP is definitely the dump stats, as shown in MDS's post, that is, libra having a 50 base attack stat just to help it fit below BSR, with Jumbao having a similarly low base attack stat for similar reasons. We really do need to have a better way to deal with base stat dumping, as any BSR limits placed will only further encourage this practice, which really helps to make CAPs look like hyper-optimized beasts of mons. I don't necessarily have a solution to this other than modifying the BSR formula for strictly physical or strictly special mons. I feel like hard limits on BST could also unfairly punish "less efficient" stat spreads, eg compare 50/150/150 to 100/108/108 which both have equivalent bulk (assuming max hp invest, this becomes more in favor of the latter spread for 0 hp spreads).

As for the overtuning of CAPs, I think it really comes down to a few issues, firstly, people are very afraid of making a truly unviable cap, as well, if a mon is unviable at release, its unlikely to become usable after. I think the ongoing post-play recap proposal should at least help to reduce these fears, and allow us to create "weaker" caps at release without worrying about them being stuck in an eternal purgatory. I also hope my limited buff proposal goes through, so that these fears are even less founded.

Secondly is well, voters, and how they want to see a strong mon produced. This is entirely down to the TLT&TL cause they need to basically try to give feedback about particularly hyper-optimized spreads/typings/movepools in the thread itself, and refuse to slate some if they're exceptionally egregious.

Thirdly, as shown in Equilibra, is the problem of obscure concepts and not properly gauging power. Libra itself was heavily down to people not respecting the central idea behind the concept, and thus voting in incredibly strong options to overcompensate for this perceived weakness. I don't have a solution here other than ensuring that concept assessment fully explores the ideas behind the concept itself, and properly gauges the strength of any obscure mechanics used. I feel that, in the aftermath of the Equilibra concept, we shouldn't face this problem to such an extent again.

Finally, Astrolotl's problem really is "regen op" + the initial overrestriction and then relaxing of the movepool which let some incredibly strong options get into the movepool. As for solving this, basically don't give grindy options to mons that never die, unsure how to really fix this, though I hope my recent moveset PR proposal helps.


We have the technology.
is a Top Artistis a Top CAP Contributoris a Top Smogon Media Contributoris a Site Content Manager Alumnusis a Battle Simulator Admin Alumnusis a Super Moderator Alumnusis a Community Contributor Alumnus
I wanted to start here by thanking G-Luke for the OP here. While I believe this thread was made with good intentions, I'd like you all to look at this from the perspective of a PRC that is looking for policy change. I'm not asking for your stance on whether or not you think CAPs are overbloated, but rather what actionable steps can be made to create change in the process. Because I read something like this, and I don't see anything that can specifically be implemented within the rules of CAP:

Ways we can go about solving this :Equilibra:

These are easier to implement than what you may initially think, as I am glad to say that moves are being made to already address aspects of this issue. But once we recognize what is exactly the problem we can go about doing so with more efficiency. For stat spreads we should be much more open to less hyper optimized stats, and be encouraged to flesh out stats on the lower end of the spectrum. For moves, we can pay key attention in how we distribute potent tools like Knock Off, Rapid Spin and Toxic, in regards to competitive move submissions, focus less on overcrowding the Pokémon with as much variety as possible, and more on fully developing sets around key moves that define our project / process. At the end of the day, I think we need to able to ask ourselves: "How do we make sure CAPs aren't hyper optimized", and be able to have a fruitful discussion about how we go about lessening the chances of that from happening in the future.
Is this a great paragraph? Yes, absolutely. But I can't help but feel... helpless reading this. If you look at the percentage of CAP mons with hazard removal compared to regular mons, the distinction is clear that we do indeed overbloat some of our CAPs. This has been true since Gen4, with the most obvious case being Krilowatt, but even Stratagem could fit into this category of overbloated CAPs. That being said, this is much easier to point out in hindsight than when you're in the thick of creating a Pokemon. And what one CAP creator thinks is too much creamer in their coffee is a fraction of the creamer craved by another creator, so to speak.

This thread gave me some deja vu that I couldn't quite pin down until recently. Then I remembered, this thread reminds me inherently of the Build Triangle PR Thread by DougJustDoug, which concluded with us striving to make Pokemon that aren't Fast, Bulky, and Powerful, but only in two of those broad categories at most. I think we've since adhered to this relatively, but I do think we've failed at some points as well. Jumbao probably is the most notable one here, as it has successfully run Scarf sets in the past, has Wishpassing, and could hit like a truck with LO/Specs sets. I don't think I saw the Build Triangle mentioned once in relation to Jumbao or its process, which goes to show how this sort of PR thread can be widely accepted by CAP veterans, but fall to the wayside over time. There's no one to blame there. It's just how memory and the passage of time works.

That being said, I'd like to propose we implement some sort of CAP main forum guide that includes these sorts of posts. It would house Doug's post on the build triangle, this post on the over-tuning of CAP Pokemon, and more as they are suggested/written. I recently watched the Pirates of the Caribbean movies, and I can't help but bring up the similarities here to the "Pirate Code," where the entire point of this is to create a set of guidelines that we adhere to when possible, but break when necessary. Having a public and visible document written by veterans about a "CAP Code" would be a great resource for onboarding new users, but also give the TL and TLT something to point to directly to tell voters that they should show some restraint.

So my proposal would be that we'd make a submissions thread here in the PRC for entries to the CAP Code. We'd chat about them as a community, tweak them, and then once approved send them over to be posted in the CAP Code for all to see and refer to as they please.

Alternatively, I am fine with us making some sort of hard limit here, as per G-Luke's proposal. We could do something like restricting a certain amount of VGMs from specific pools, or tightening up our BSR calculations to allow for less manipulation. All of those things are fine, but I think most of us will agree that CAP already has a lot of red tape. It behooves us to have some guidelines posted by veterans to keep us on track, but still allows flexibility in the CAP process as needed, which is why I think something like a CAP Code fits our needs more succinctly.
Last edited:


is a Pre-Contributor
I really like the idea of having some kind og "CAP Code" or loosely enforced guidelines that allow us to make sure we're not making CAPS that turn out to be S-tier threats the moment they're dropped into the metagame. Of course, there are two things that we need to immediately consider when doing this kind of things:
  1. These guidelines cannot simply be unspoken. They need to be clear, concise and generally available to anyone who wants to get involved with CAP.

    I don't think this section comes as a shock to anyone. Birkal literally mentioned having a post about these kinds of guidelines in a main forum guide. I wholeheartedly support that proposal and I think CAP processes as a whole will be better for it. Allowing people to understand that we're not trying to make some S-Rank mon every time we do a concept is extremely useful in making sure the CAP process remains interesting and dynamic, and that the product isn't a deterrent for people to play the metagame.

  2. As useful as they might be, these guidelines can't be imposed on CAP28 halfway through the process.

    If we develop these guidelines, I would rather we finish the process we've already started, rather than having to shoehorn everything in in the later stages. Of course, that's assuming that the ball would even start getting rolled before CAP28 wraps up. I think holding off on implementing a formal "code" until after CAP28 helps us maintain the integrity of the process as a whole.

Those are my two-cents on the proposal by Birkal. I think they might be stating the obvious, but someone's gotta do it. As I said before, I support this idea of making sure that we're not making CAPs that are too powerful, and I think this is a smart way of implementing things while also not completely altering the rules of the process.


Master Procraster
is a CAP Contributor
I agree with the CAP Code suggestion from Birkal.

From bits that I have witnessed, I don't think the over-optimization of CAPs that we have been seeing is necessarily a process issue, although you can link the trends we witness today to the shift to building in the CAP metagame instead of OU (I think this is a neccesary evil if it is the source problem, as a switch to the CAP meta was/is an extremely beneficial decision for the health of CAP). I think the over optimization of CAPs is much more a cultural issue, and the CAP Code would help remedy this to some extent, as well as give newer users an easier in for the process.

I think I also like the free suggestion and implementation that has been suggested for this process, though I also think the community should be able to remove a code if its influence pushes CAP in an unfavorable direction, or end up being far too restrictive that the process becomes predictable and linear. I'm not saying that we can't have codes that don't imply or encourage restriction (I think the build triangle is a healthy restriction that limits the amount of "bulky attacker with a strong attack stat and a speed stat almost never under 90" that we see like all the time) but if it comes to the point where people are treating that rule like gospel and its having a noticeably bad effect on the process then it needs to go.

Also I support the notion that CAP Players are scallywags as this suggestion implies. Yarr.

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