Hello, sorry I've been a bit too busy to follow up on this, but I was looking through some OO stuff in ADV analyses and agree with you that outclassed / inconsistent is insufficient as a descriptor, or at the very least, it's not entirely clear what those descriptors mean.
I realize there are actually two questions to address.
1. What moves / set should even make it to the analysis (whether in OO or in set description is a question to be answered later)
2. What moves belong in set description, not in OO?
I realized that I don't have strong opinions on what should be in OO, but instead I feel more strongly about what should not be in OO but in the set description;
My principles of what should be in set description and not OO
- Analyses in old gens rarely get updated and should stand the test of time.
- If you need a piece of knowledge to win a serious game against a good player when you face this set, this knowledge should be available in the set description.
Hence, some examples of what I think should not be in OO, goes into set description:
- Moves/sets that were once used but are off-season due to reversible metagame changes should still remain in the set description, even if they're not explicitly slashed. This gives the player a view of the metagame regardless of when they read the analysis, and makes future updating easy when these moves / sets become popular again because it doesn't involve moving around content.
- Niche moves that fit well on a rare archetype; this should go into set description because the archetype provides context - a good player has a decent probability of bringing said archetype where you have to consider what every mon does on such an archetype (for example, in ADV, Sub/Roar/CM/Surf Suicune on Rain. There is much less reason to think someone would need to understand a niche set that has no context and needs a specifically crafted team, like randomly putting Icy Wind on a Pokemon instead of Ice Beam. I think this ties in to your "usable but lacks a purpose" point, but in a more specific manner, because there are some moves like Thief in ADV that are usable and definitely help one to make progress, but has unclear context because without any archetypal framework, justification has be made on a team-by-team basis.
Some gray area stuff
- What about a set that is unique and reasonably consistent at what it does (thinking defensive Wish Salamence in ADV) so that terms like "outclassed" or "inconsistent" don't apply, but are rarely seen because they open up gaping holes in other aspects (getting Baton Passed on in this example)? In this case, it's kind of a matchup fish. So it's not so much the Pokemon that is inconsistent but the resulting team. Maybe this is a refinement of your point about opportunity cost?
As for whether or not something makes it to OO from nothing, my gut feel is that anything that I've seen remotely work in a game should qualify. Not clear to me how to articulate that though.
I'll post more thoughts when they come to my mind. Feel free to dispute/add on.
In all honesty I don't think I remembered to consider Set Description vs OO, since normally I stick to what's slashed in the set description, but I guess that's an error on my part and something that's inconsistent with others. Anyway, I actually like your criteria for distinguishing between SD and OO. My post was trying to catalogue a variety of different dynamics, which is generally not the best approach since it's hard to be comprehensive and it focuses on the different ways something might jump from nothing to OO or OO to SD without really identifying a boundary. Your definition I think more clearly identifies a boundary while also being more general purpose, and thus not prone to unanticipated dynamics. It's also a good deal simpler, which is nice.
Because I'm not sure how clearly I worded that, perhaps an analogy would be if you decided to walk to the next suburb over. My approach was plotting a range of potential routes- I tried to get most of them, but there are probably others that I missed, and it may not always be certain when you transition into the next suburb. Your approach was to simply identify a boundary where one suburb crosses into the next, that way you know which suburb something's in, regardless of what route you're on.
Regarding the grey area thing, I see 2 issues at hand with the example you describe of WishMence. The first is that it's an entirely distinct set, rather than an alternate option for an existing one, which means it cannot go in SD, and must go in OO, or be its own set. The simplest solution is to put it in OO, but I guess then it might be inconsistent with the standard you describe. Otherwise, I think it could be listed as its own set, provided the corresponding description clearly states the flaws and issues of the set- I think this would be a reasonable compromise. Not sure how you'd incorporate that into a skeleton though
The second issue is that it's heavily matchup oriented, which makes it difficult to gauge its viability. Personally, I think a match-up oriented set is still worth treating as viable, unless it's impossible to build a team that includes that set that doesn't have a fighting chance in an unfavourable matchup. Personally, I'm not a fan of matchup fishing. It's obviously nonsense to expect a team to have no unfavourable matchups, but even if I'm not favoured, I still expect a given team to give me a decent chance to win a bad matchup. If a team has genuinely unwinnable matchups, I don't consider that team viable, even if it's really strong otherwise. If using a specific set would mean that any team it is facing unwinnable matchups, then I think that set belongs in OO.
I guess an example of this would be dual electric teams in gen 1, which I've historically used a fair bit. I think they're usable in RBY despite the existence of Rhydon making them heavily matchup dependent, because I still think it's possible to beat Don teams with a dual electric team, it's just a good deal harder. I tried it ages ago in Stadium, and in that meta I consider the matchup unwinnable, and thus the team to be unviable (team preview cripples it)
Regarding the distinction between OO/nothing, I think that's a good general threshold, but you're right in that it probably is worth defining more. I guess my take on it might be "anything that could be even slightly effective against competent players", which isn't really much of an improvement tbh. One flaw sticks out to me in that luck-based strategies might inherently be favoured by such a definition, since almost anything can be effective with the right amount of luck. It may also be worth explicitly noting that OO is differentiated from SD in that moves in SD are relevant enough that they might need to accounted for against good players, whereas OO is theoretically usable, but you probably don't really have to worry about it
*This is potentially inconsistent with me arguing that Flash Chansey is merely not good, rather than being trash. I'd argue that Flash Chansey gets far more value out of fishing for hax since it finds itself in low risk low reward situations far more often than Gengar, which is why I view them differently