For me, having any OSs at all just runs counter to the stated goal of the game: to be more friendly to semi-competitive players. It doesn't really matter if they're hard, they will be used by the best players. And I use braindead in the wider sense, in that OSs allow players to not have to make decisions. That's BS.
Every OS in existence is a blight upon mankind. The only good OS is a dead OS.
I said it before (though not on here): OSes by itself aren't inherently good or bad. Everybody makes use of OSes at all levels of play, whether they realize it or not. Things such as buffered specials inside of normals when you're playing at a max range poking range are OSes. Safe jumps that have normals or specials buffered into them are OSes. These two things are used by players of all levels - especially the first one. And the first one is something that almost every player encourages people to do when playing at that poking range. It's hard to find someone who *doesn't* use buffered specials at max range poking from my experience or safe jump setups that rely on the opponent putting them in blockstun to eat up an input that would have came out otherwise or buffered back dash setups during blockstrings that is simply just making a status check of true blockstring vs gap in string.
It's more that the risk/reward structure for the OS needs to be properly balanced given the context and OSes that skew this greatly towards reward relative to lowered risk should have at least some attempts to be analyzed and see what can be done or see what the rationalization of it remaining as such should be. OSes (typically) aren't undefeatable and all encompassing. The basic ones are designed around a specific situation and around some of the common responses in that situation.
Looking at probably the most basic one of buffered specials in normals: this is designed purely as hit vs whiff. If the normal doesn't connect, nothing comes out while if the normal connects the special will come out. Practically every single fighting game has this now and the only ones that wouldn't are specifically ones that allow normals to whiff cancel into specials prior to recovery frames have fully completed. But we see lots of times that this OS can be worked around if it's relied upon a lot by walking just in range of the expected normal and immediately blocking to allow for a punish once the special comes out. Is it hard to do? Well, probably considering how quick some normals come out. But other avenues of punishment also exist in the form of whiff punishing. So is this type of OS okay? In my mind, sure - it's not inherently doing games any harm and the implications of attempting to remove this type of OS would probably have *worse* effects on the game than leaving it in.
Safe jumps with buffered inputs: another common one that people at an intermediate (or possibly earlier?) look at as well. This is something that we all should be familiar with at least. There's quite a few things to look out for in this one, but the most simplistic one is probably wake-up block vs invuln back-dash vs invuln reversal. Is this type of OS beatable? Depending on the characters and the game mechanics, sure. If the opponent cannot even punish the invuln back-dash because of [reasons], then this specific OS doesn't even work and a different one would have to be employed such as a character like Slayer in GG who can just BDCFD or has unique reversal options in BDCBite against some oki. Depending on which response the Slayer is electing to take, the opponent has to take different options, even if using an OS simply because some of these OS-able states cannot be fully covered (at least to my knowledge, I'm not sure how one could safejump and cover BDCDoT, BDCBite, BDCDandy, BDCFDJump, and backdash within a single OS). Certain system mechanics inherently make certain OSes have more counters as well (though at an increased execution barrier to the person on the defense) such as parries in 3S or focus in SF4 (though both of these brought a lot of issues as well in terms of OSes). If you want to get rid of *any* safejump OS, it feels like it would take a lot of work around concerning hitstop windows in order to remove these, and if you feel that a simple safe jump with buffered input is okay, then where is the line drawn between a “simple” one that encompasses X options and a “complex” ones that encompasses X+1 options?
Another one that will most likely see increased use is SF5’s Chun-Li being able to OS Lightning Legs vs Spinning Bird Kick in combos depending on if she had enough charge through her 5MP>2MK link or not. How would you even go about preventing this type of OS? If you remove this by having Lightning Legs be a strict 236 input instead of 2369 and special-cancelling the pre-jump frames, then you’re going against your goal of removing OSes for the sake of being “more friendly for semi-competitive players” who may not have 100% perfect execution of the top echelon of competitive players (who also don’t have perfect execution 100% of the time). Similarly if you make the SBK input only accept 2 and 8 inputs as charge start, gain, or execution of charge, then you’re also making things harder for semi-competitive players to maintain charge and execute it as well. So would you just change the input entirely to 214K similar to how lightning legs was moved to 236K? That’s fine and all – but what happens for other characters who can do similar OSes that cannot afford having the motion changed to that due to another move already occupying that motion?
Certain OSes that typically *all* people felt were way too rewarding not to do in terms of how much (little) risk it presented were things like the YRC-Burst OS in early Guilty Gear Xrd. There was literally *no* downside to this except in terms of execution. There was no counter to this outside of hoping the opponent had less than 25% meter or more than 49% meter. But this could be controlled because of tools like YRC, FD, and Blitz. This OS was also made possible due to the priority system for buttons in GG (and practically every other fighting game). Thankfully, this OS was easy to fix and *everybody* saw that the reward was way too disproportionate to the risk (read: infinitely rewarding), and it was changed *without changing any other mechanic in the game* which is something that’s important when it comes to fixing these OSes as fixing this had to affect everybody while also bringing about no unintended consequences which seems to be something that the whole “remove every and all OSes” sentiment doesn’t fully express in their explanation.
Removing a lot of these OSes would require changing in how hitstop works (read: most likely not have any hitstop at all) or changing in how input priority and reading is done (read: more strict). Both of these also make the game less semi-competitive friendly, so by removing the OSes, you’ve just shifted the balance from “semi-competitive players will have to be aware of these OSes and their counters” to “semi-competitive players will require incredibly strict execution at all times and will have to appropriate adjust to no hit-stop on anything” and that *still* doesn’t remove mutually-exclusive state based OSes (like the V-Reversal/Throw OS or Guilty Gear’s Throw/c.S OS or GG’s Throw/Gold Burst OS).
The fact is that SF5 *is* “more semi-competitive player friendly” (I really dislike this term btw, I feel if you refer to yourself as this, you’re competitive, just with not as many financial resources or time available). The 2-frame buffer for normal makes combos *vastly* easier for everybody. There should be no argument about this aspect at all. Things like making Chun’s lightning legs not a mash input makes playing her easier (it’s literally the reason why I went from not playing Chun in SF4 to even considering her in SF5). More options in training mode makes creating oki tools or meaty setups much easier (though the training mode is certainly lacking many things, it’s definitely an improvement over SF4 in terms of available options). The fact I can go online, search for a player, and retrieve that person’s most recent replays instead of just the replays that person uploaded makes learning certain matchups or approaches or strategies easier. Compared to other games like Skullgirls or Xrd or BlazBlue? That’s debatable I suppose, but in terms of *teaching* the player or introducing a brand new player to a fighting game, those three are completely above SF5. In terms of providing tools to the player, I’d say Skullgirls is above SF5, and Xrd and BlazBlue are about at SF5’s level, with possibly SF5 being above due to having wakeup timings and such which are what people are more focusing on now. However, compared to SF4, it’s really difficult for me to say that Capcom failed to deliver on their promise of SF5 being more friendly to people that are not Capcom Cup caliber. Is it more punishing to people when you mess up? Yes, it’s quite punishing to eat a Crush Counter into half your life or something similar, but it’s definitely easier to execute and (probably more important to me) *more fun* compared to SF4.
Should Capcom attempt to remove every single OS it can? No, that would take way too much effort, especially on an already completed game, and we'd probably end up with a game like SF1, just with more characters. Should Capcom attempt to remove disproportionately rewarding OSes? Yes - an OS should have a clear and consistent counter to it or at least one of the options the OS covers doesn't put the opponent in a disproportionately risky state than what would have happened if the OS didn't exist at all.
