Not copy-pasting — just making sure the core issue is visible and clear, especially when it’s still being misunderstood or sidestepped. If the AI’s frame-perfect reaction speed is the root cause of so many in-game problems, then yeah — I’m going to keep raising it until it’s properly addressed. It might feel repetitive, but if that’s what it takes to get a gross fundamental system imbalance fixed, then so be it. Some things are worth repeating.
And let’s not forget — Kabam said in their dev diary that they don’t want to change that “frame-perfect button masher” mechanic. That’s what scares me. Because it means that no matter how smart AI 2.0 is, it’ll still be reacting at 16.7ms. That’s the bit we, as a community, need to question right now — before it’s fully built, implemented and rolled out. Because once it’s done and in the context, it could be another 10 years before it’s changed again… IF it’s ever changed again.
I think at this point, you're better off just seeing it when it arrives. I think you're hung up on a technical point that doesn't have the real world consequences you think it does. You're worried about something that, if it gets implemented, you're afraid you're going to have to live with for years. I get that. But in this particular case, what you're worried about is something the implementers aren't going to address, because it doesn't mean what you think it means, and even if it did, it wouldn't lock the AI into behaviors you're worried it will.
I'll try one last time. You think the difference between MCOC AI and other fighting game AI is that MCOC makes instantaneous decisions with perfect information and other fighting games do not. You think other fighting games somehow implement an AI that is "slow" and "ignorant" - that takes in data and then just kind of sits on it pondering, and then some time later makes a decision. But I'm pretty sure no fighting game AI works that way, because that would be nonsensical. Instead, they all take in data and decide, instantaneously, what to do about what they see just like MCOC. But those AIs have more sophisticated options available to them that allow them to decide, instantly, right now, to take actions in the future. They can decide now to react ten frames from now. They can decide to execute a reaction script that has built in delays in them. They can decide to prioritize acting on what it remembers the player doing in the past rather than what it observes is happening right now. It is going to be just as fast, but what it does with that speed is going to be significantly different than what it does now. And because of that, there's no reason to complain about the speed. It is how it decides what to do that is more important than how fast it decides to do it, and that's what AI 2.0 is intended to address.
Last thing: the dev diary did not specifically state that Kabam doesn't want to eliminate the button mashing part of the AI on principle. Rather, they said this behavior would not radically change initially because changing the AI in a dramatic fashion would be disruptive to players. Players who have learned for years to intuitively play against an AI that is just a dumb button masher would find their intuition no longer working if they did that. So it seems that they will add the new AI features in stages, rolling out increasingly more sophisticated behaviors to see how they affect overall play. Whether they keep that indefinitely will depend largely on the way the playerbase responds to how the AI's behaviors change as they roll the new systems out.
You can see that the button mashing thing is in fact going to go away in the long run, because they explicitly state that one of the goals of the AI 2.0 roll out is to design the new system to roughly replicate the more twitchy random behavior of the current system, specifically to make sure players aren't caught off guard by an AI that does completely different things. But this step would not be necessary if AI 2.0 was still a random button masher. If it was, there would be no need to expend effort to try to get AI 2.0 to replicate the behavior of AI 1.0, and no worry it wouldn't do so exactly. It is a different system, but it is being designed to replicate the behavior of the old system to minimize the break in player expectations when it starts to get rolled out.
Hey all — appreciate the continued discussion, but I think it’s important to drive home just how fundamental this issue is. This isn’t a minor quirk. It’s not about wanting the AI to be easier, or passive, or dumbed down. This is about the core architecture of how the AI functions — and how that architecture creates a fundamentally unfair playing field.
Kabam literally stated the AI is a “frame-perfect button masher” — meaning it can make and validate decisions every 16.7ms at 60fps. That’s 10–12x faster than a pro player, and up to 15x faster than your average player. It doesn’t predict, it doesn’t anticipate, it just sees a gap — even a 1-frame gap — and acts instantly. That’s not intelligent. That’s just cracked.
And this one design choice explains so much of what feels off in high-end content: • Specials getting blocked mid-combo • MLLM into special chains being blocked while MLLL works • Dash-ins being intercepted before your animation even completes • Relics being blocked after landing 1–2 hits • Parry “fails” that probably aren’t actually input drops — just the AI reacting a frame faster than us
The issue isn’t about how “good” the AI is. It’s that it’s allowed to act faster than any human ever could. That’s not something you can outplay. That’s not about spacing or memory. That’s a system-level imbalance.
And to be blunt, as I said before, no other modern fighting game does this. Street Fighter, Mortal Kombat, Injustice Mobile, Tekken, even God of War and Elden Ring — they all build in human-like reaction caps, or buffer windows, or intentional AI delays. Not to make things easy — but to make them fair. Kabam’s AI doesn’t have that. Never has. And with 60fps, it’s now reacting twice as fast as it used to.
DNA, your point about smarter AI is spot on. Smarter is better. But smarter + frame-perfect = even worse. Because now you’ve got an AI that not only reacts faster than us, but makes smarter decisions while doing it. That’s a disaster waiting to happen.
And let’s not forget — Kabam said in their dev diary that they don’t want to change that “frame-perfect button masher” mechanic. That’s what scares me. Because it means that no matter how smart AI 2.0 is, it’ll still be reacting at 16.7ms. That’s the bit we, as a community, need to question right now — before it’s fully built, implemented and rolled out. Because once it’s done and in the context, it could be another 10 years before it’s changed again… IF it’s ever changed again.
So this isn’t about behavior or difficulty. It’s about speed. And the fact that the AI operates on a system that no other competitive game allows. That’s not something we can adapt around forever.
Thanks again for the thoughtful responses. Even where we disagree, I think we all want the same thing — a game that challenges us, but doesn’t cheat us.
So I think you’re off base a bit here. I’m good at fighting game execution. 1 frame combo hitting good so maybe I down play the difficulty of mcoc but it’s not as bad as you appear to think. It really only needs tweaks at the margins.
It’s not possible for us to tell exactly why those things happened that you described. I think it’s less a problem of speed and more a problem of “hitboxes” and hit/block-stun. If those are calibrated correctly, you won’t get those types of errors regardless of the rate of AI input.
It actually doesn’t matter if the AI can roll actions at 30x per second or more as long as it doesn’t do so in the wrong circumstances. If the AI dashes in, it makes no difference if it was frame perfect, you can parry it every time.
It’s the edge cases where the AI’s speed becomes a problem not in general play. I’ve given some examples but it’s really when the AI is right in your face or walking into that range. It’s impossible to react to a frame perfect input (if it rolls right).
So you should be less concerned about why it can act so fast and more concerned about why it’s allowed to get into range and take specific actions.
The AI in MCOC is really good. I’d even say great with some really specific pain points.
Hey, appreciate the thoughtful response — definitely good points raised. I don’t disagree that MCOC’s AI can feel good in many situations, especially compared to other mobile games. But I think you’re underestimating just how fundamental the frame-perfect issue is.
It’s not about edge cases or reaction vs. prediction — it’s that the AI can always act with perfect information, instantly, every 16.7ms. In fighting games like Street Fighter, Tekken, Mortal Kombat, Injustice Mobile, even at the highest AI levels, the AI has baked-in reaction delays to simulate human limitations. They do that on purpose — because they understand that if you allow the AI to react frame-perfectly with no cap, you eliminate the concept of “earned” openings entirely.
You mention that if hitstun or blockstun were calibrated better, it wouldn’t matter how often the AI rolls actions. I get the logic — but in practice, if the AI can always test every available option every frame, and instantly switch gears with no human hesitation, you’ll still always run into BS scenarios: • specials blocked mid-combo, • relics getting blocked mid-string, • dash intercepts at ranges we physically can’t react to •. Ai light intercepting you out of nowhere • parry “fails” that aren’t failed inputs, just losing by a single frame.
These aren’t just bugs or quirks. They’re symptoms of a system that’s fundamentally skewed.
I 100% agree it feels great when it works — but when it fails, it’s not because of behavior tuning or hitbox issues. It’s because the AI is playing on a completely different reaction timeline than we are, and with the addition of 60fps from 30fps, this issue has only been exacerbated further.
Bottom line: “Smart AI” is good. “Frame-perfect AI” is not. If Kabam truly wants AI 2.0 to feel rewarding and fair, they’ll need to limit reaction speed like every other serious fighting game does — not just tweak behaviors or ranges.
Thanks again for the conversation. These discussions are exactly why it’s important to bring this stuff up before AI 2.0 launches, not after.
Hey @DNA3000 appreciate you sticking with the conversation and explaining it clearly. I actually think we agree on more than it sounds like. But there’s still a core difference that matters hugely.
You’re right that smarter AI is the goal. Smarter = good. Smarter AI that can plan, recognize patterns, remember past behavior — all great things.
But: Even a smarter AI that still reacts instantly on a frame-by-frame basis will feel cracked and unfair, because it will capitalize on gaps humans physically can’t avoid. That’s the danger.
Other fighting games like Tekken, Mortal Kombat, Injustice Mobile do react instantly in decision-making, but they deliberately add reaction delays, buffered moves, or cooldown states after that decision. The games don’t let the AI instantly block, intercept, or counter off a 1-frame window. That’s why human players still have windows to bait, punish, and fight back. It’s planned fairness, not just smarter thinking.
In MCOC, unless they add some kind of reaction timing limit or decision buffer, smarter AI will still abuse the raw frame-perfect system — it will just abuse it more intelligently. That’s the real risk. The AI doesn’t have to be malicious — just mathematically faster and sharper than us every single time. And that would make top-end gameplay feel oppressive, not rewarding.
Also — I get what you’re saying about them trying to minimize disruption at first. But Kabam’s wording was careful: They said they want to replicate the random/twitchy behavior “at first” so players aren’t caught off guard — but they didn’t promise to change the fundamental speed mechanic behind it.
That’s why raising the issue now matters so much. Once AI 2.0 is fully live everywhere, it’ll be much harder to argue for deep systemic changes like speed-limiting. And it could define how the next 5–10 years of the contest feels to play.
And just to be clear — if this sounds repetitive from me, I get it, and I’m sorry for hammering it so hard. But it’s because this is the fundamental issue underneath so many things players complain about — parry inconsistencies, specials being blocked mid-combo, instant intercepts, light attacks through relics, and so on.
It’s not about Kabam acting maliciously — I genuinely think, as with other things recently, they just kicked the ball and didn’t look where it would land. 60fps was meant to make the game smoother — but because the AI works on a frame-by-frame basis, now it’s acting twice as fast as before. And if we’re going to keep 60fps (which we obviously should), we have to address the frame-perfect reaction speed alongside it. Otherwise the problem only grows worse with smarter AI brains.
So I’m pressing this point now — before AI 2.0 is fully implemented — because once it’s baked into the system, it will be almost impossible to untangle. And it could be years, maybe decades, before we ever get the chance to fix it again — if ever.
That’s why I care about this so much. That’s why it needs community pressure now — not after the fact when it’s already too late.
At this point, I think it would be really helpful if someone like @kabamdork or @KabamPinwheel could weigh in with clarification — just so we know whether the frame-perfect reaction mechanic is intended to remain permanently, or if it’s something still open for review as AI 2.0 evolves. I also made a Reddit post on this which is now at about 23,000 views, with a lot of players sharing examples and discussing this. I’ve linked this forum thread there too so people can easily go between them. Would be great if the mods could weigh in here to give the community a bit of clarification — I’ll link the Reddit post below for context. Would really appreciate any insight they can share.
Thanks again for the thoughtful responses everyone. Even where we disagree on some of the points, it’s clear we all want the same thing — a game that challenges us, but is fair and doesn’t cheat us.
When you say fair, do you also think we should let the AI learn about nodes and power stings and effects? Because we can know that. It's only fair we the AI does too.
You'll essentially make sure every power sting Champion becomes worthless.
When you say fair, do you also think we should let the AI learn about nodes and power stings and effects? Because we can know that. It's only fair we the AI does too.
You'll essentially make sure every power sting Champion becomes worthless.
That’s not what I’m saying at all — and honestly, by this point in the forum thread (especially after you asked if I was just “copy-pasting the same comment or if it was deja vu”), you should know that. Nobody’s asking for the AI to “know more.” We’re asking for it to react fairly — with human-like reaction time — instead of “frame-perfect,” instant decisions.
It’s about speed, not behavior or buff awareness. Two completely different issues.
When you say fair, do you also think we should let the AI learn about nodes and power stings and effects? Because we can know that. It's only fair we the AI does too.
You'll essentially make sure every power sting Champion becomes worthless.
That’s not what I’m saying at all — and honestly, by this point in the forum thread (especially after you asked if I was just “copy-pasting the same comment or if it was deja vu”), you should know that. Nobody’s asking for the AI to “know more.” We’re asking for it to react fairly — with human-like reaction time — instead of “frame-perfect,” instant decisions.
It’s about speed, not behavior or buff awareness. Two completely different issues.
You are asking it to be fair because apparently the AI being fast is unfair to the player. That's the entire baseline of your argument. FAIRNESS.
I'm saying that if we're using fairness as a metric, then it's only fair that the AI has the advantages that we have too.
@Average_Desi —You’re misunderstanding again and this really isn’t that complicated.
The goal isn’t to strip the AI of tools or give it player-specific buffs or knowledge. It’s simply to balance the ‘fairness’ of reaction time so that fights are decided by skill, not by an AI that can act literally faster than humanly possible. No other fighting game allows the AI to act faster than humans because that’s literally bad game design, it’s that simple.
You’re trying to argue about “advantages” when the core issue is that timing windows themselves are broken. It’s about leveling the speed of interaction — not its “awareness” or “advantages.” If the AI reacts faster than we physically can, that’s not fairness — that’s rigging the timing of the entire fight.
@PickL1e89 — Nailed it in one. Speed, not behavior. The frame-rate doubling to 60fps is what really exposed it. The AI was already faster than humans at 30fps — now it’s just fully cracked, and the issue shows way more blatantly.
Again, if you want to balance out fairness so that "skill" is rewarded, why are you only singling out the only place which AI has a supposed advantage and not the factors where we as a player have advantage?
Why is it okay that we have an unfair advantage over the AI, but the AI is not allowed to have any?
Again, if you want to balance out fairness so that "skill" is rewarded, why are you only singling out the only place which AI has a supposed advantage and not the factors where we as a player have advantage?
Why is it okay that we have an unfair advantage over the AI, but the AI is not allowed to have any?
@Average_Desi — Again, you’re mixing two separate concepts.
Nobody’s asking for the AI to be nerfed, or to loose power/tools. We’re asking for the reaction speed system — the foundation of how fights function — to be fair and human-like.
Players having access to buffs, masteries, or node knowledge isn’t the same thing. Those are designed asymmetries. They’re intentional parts of how the contest is balanced — just like defenders getting nodes that give them massive boosts in attack, armor, power gain, regeneration, unblockable effects, you name it.
The AI already has huge advantages through nodes. That’s normal, and nobody is complaining about that.
But the base timing rules — the fundamental ability to react to an event — should be fair for both sides.
If the AI can react frame-perfectly, and we physically can’t, that’s not skill testing — it’s a fundamental system-level imbalance and bad game design.
Every serious fighting game limits AI reaction timing to keep the contest skill-based. We’re simply asking for MCOC to follow the same basic standard.
Skill should win fights — not literal robotic speed the player physically cannot match.
Again, if you want to balance out fairness so that "skill" is rewarded, why are you only singling out the only place which AI has a supposed advantage and not the factors where we as a player have advantage?
Why is it okay that we have an unfair advantage over the AI, but the AI is not allowed to have any?
@Average_Desi — Again, you’re mixing two separate concepts.
Nobody’s asking for the AI to be nerfed, or to loose power/tools. We’re asking for the reaction speed system — the foundation of how fights function — to be fair and human-like.
Players having access to buffs, masteries, or node knowledge isn’t the same thing. Those are designed asymmetries. They’re intentional parts of how the contest is balanced — just like defenders getting nodes that give them massive boosts in attack, armor, power gain, regeneration, unblockable effects, you name it.
The AI already has huge advantages through nodes. That’s normal, and nobody is complaining about that.
But the base timing rules — the fundamental ability to react to an event — should be fair for both sides.
If the AI can react frame-perfectly, and we physically can’t, that’s not skill testing — it’s a fundamental system-level imbalance and bad game design.
Every serious fighting game limits AI reaction timing to keep the contest skill-based. We’re simply asking for MCOC to follow the same basic standard.
Skill should win fights — not literal robotic speed the player physically cannot match.
Again, if you want to balance out fairness so that "skill" is rewarded, why are you only singling out the only place which AI has a supposed advantage and not the factors where we as a player have advantage?
Why is it okay that we have an unfair advantage over the AI, but the AI is not allowed to have any?
@Average_Desi — Again, you’re mixing two separate concepts.
Nobody’s asking for the AI to be nerfed, or to loose power/tools. We’re asking for the reaction speed system — the foundation of how fights function — to be fair and human-like.
Players having access to buffs, masteries, or node knowledge isn’t the same thing. Those are designed asymmetries. They’re intentional parts of how the contest is balanced — just like defenders getting nodes that give them massive boosts in attack, armor, power gain, regeneration, unblockable effects, you name it.
The AI already has huge advantages through nodes. That’s normal, and nobody is complaining about that.
But the base timing rules — the fundamental ability to react to an event — should be fair for both sides.
If the AI can react frame-perfectly, and we physically can’t, that’s not skill testing — it’s a fundamental system-level imbalance and bad game design.
Every serious fighting game limits AI reaction timing to keep the contest skill-based. We’re simply asking for MCOC to follow the same basic standard.
Skill should win fights — not literal robotic speed the player physically cannot match.
Fair enough. Understood. I don't agree with the conclusion or the solution, but I get where you came from.
Again, if you want to balance out fairness so that "skill" is rewarded, why are you only singling out the only place which AI has a supposed advantage and not the factors where we as a player have advantage?
Why is it okay that we have an unfair advantage over the AI, but the AI is not allowed to have any?
@Average_Desi — Again, you’re mixing two separate concepts.
Nobody’s asking for the AI to be nerfed, or to loose power/tools. We’re asking for the reaction speed system — the foundation of how fights function — to be fair and human-like.
Players having access to buffs, masteries, or node knowledge isn’t the same thing. Those are designed asymmetries. They’re intentional parts of how the contest is balanced — just like defenders getting nodes that give them massive boosts in attack, armor, power gain, regeneration, unblockable effects, you name it.
The AI already has huge advantages through nodes. That’s normal, and nobody is complaining about that.
But the base timing rules — the fundamental ability to react to an event — should be fair for both sides.
If the AI can react frame-perfectly, and we physically can’t, that’s not skill testing — it’s a fundamental system-level imbalance and bad game design.
Every serious fighting game limits AI reaction timing to keep the contest skill-based. We’re simply asking for MCOC to follow the same basic standard.
Skill should win fights — not literal robotic speed the player physically cannot match.
Again though it feels like you're misunderstanding what that time actually represents. It's like for example saying it's unfair to us (the players) that we get to see the arc of thrown ball for multiple seconds and get to adjust to the arc make plenty of decisions compared to the ai which in this example opens its eyes with a ball 30 cm away into its face, and the fact that more often than not its doing something else that prevents it from reacting in time and even if it was open to react, it's literally chance on whether or not it makes the "right" decision
Who would you rather be in that example?
You may not have the robotic reaction time but you have the luxury of actually remembering what has happened in the last couple seconds and that's plenty to bridge this gap you feel is unfair
@PickL1e89 — Exactly. What you’re describing — that “lag spike” effect — is actually the AI reacting faster than humanly possible. It’s not lag; it’s the AI responding frame-perfectly at 60fps (16.7ms). Thanks for calling that out — it’s a big part of the problem we’re trying to get addressed before AI 2.0 finalizes.
@victor158 — Respectfully, you’re misunderstanding the core problem.
First — the AI already has built-in advantages, and that’s perfectly fine. Just like we have masteries, relics, synergy bonuses, and node knowledge, the AI gets attack bonuses, regen nodes, unstoppable buffs, power gain, and so on.
Those are designed asymmetries — intentional parts of balance.
But mechanical asymmetry — reaction timing that humans physically cannot match — breaks the foundation of fair fights.
Players have natural input limits: • Animation recovery (after throwing specials or attacks) • Input lag (both device and engine-based) • Physical human reaction time (~150–250ms+)
Meanwhile, the AI has zero such limits. It can instantly react — block, evade, intercept — within 16.7ms. No human can possibly match that, no matter how skilled.
To use a better analogy:
It’s like getting into a boxing match where your opponent can instantly dodge the second you throw a punch, but you have to wait a full second to even see their movement before you can react.
You can plan, strategize, be smart — but mechanically, you’re doomed to get hit more often than you land hits, simply because your body can’t react instantly like a machine.
Even if Kabam codes in the AI making occasional mistakes (“eg in the boxing match example it losing its footing”), it doesn’t matter if the AI can instantly recover faster than players can punish it. • We’ve seen this when trying to punish specials: You dex a special perfectly, dash in — but the AI immediately blocks your hit, because its reaction speed is frame-perfect. • We’ve seen it when trying to combo into specials: Some champs (with longer mediums) can’t safely chain medium–special attacks anymore because the AI blocks between the hits. • We’ve seen it when dash-attacking: Even if the AI looks aggressive, it can instantly backdraft intercept because it reacts on the frame you dash forward.
Correct play is being actively punished — and that breaks the basic premise of skill-based fighting.
In MCOC, historically, if you played correctly — dexed a special, timed a punish, spaced an intercept — you were rewarded with an opening.
Right now at 60fps, you’re often punished for playing correctly — because the AI is simply reacting faster than humanly possible. That same opening we used to have from fully dexing a special is now non existent or unsafe. Correct gameplay should NOT be punished.
Every serious fighting game in the world (Tekken, Mortal Kombat, Street Fighter, Skullgirls, Injustice/Injustice 2 mobile, God of War etc etc) enforces reaction delay on AI — to protect skill-based gameplay and avoid robotic unfairness.
MCOC needs to do the same.
Again — We’re not asking for dumber AI. We’re not asking for easier fights.
We’re asking for fair fights — where skill, prediction, timing, and correct play are rewarded, not punished by robotic speed.
Hey @DNA3000 appreciate you sticking with the conversation and explaining it clearly. I actually think we agree on more than it sounds like. But there’s still a core difference that matters hugely.
You’re right that smarter AI is the goal. Smarter = good. Smarter AI that can plan, recognize patterns, remember past behavior — all great things.
But: Even a smarter AI that still reacts instantly on a frame-by-frame basis will feel cracked and unfair, because it will capitalize on gaps humans physically can’t avoid. That’s the danger.
You still seem to think other fighting games don't do this, even though they do. I'm guessing you didn't read the article I linked that described how this worked in an actual fighting game known for being a fighting game. So let's put it this way, AI 2.0 is going to do what those other AIs do, more or less. And when you ask "but is it still a frame accurate AI with instant reactions" I'm going to suggest to the devs that they simply stop saying that, and instead state that the net result will be a system that does not appear to do that.
Every serious fighting game in the world (Tekken, Mortal Kombat, Street Fighter, Skullgirls, Injustice/Injustice 2 mobile, God of War etc etc) enforces reaction delay on AI — to protect skill-based gameplay and avoid robotic unfairness.
1. I think you're going to need to start citing references here.
2. Interesting that you mention Injustice. I played Injustice and Injustice 2. If any game has AI similar to MCOC's AI 1.0, it is Injustice. If anything, it is much less forgiving than MCOC's is, particularly in its defensive response speed. At highest difficulty the Injustice AI could respond instantly to player attacks, even moreso than MCOC's does/did. So in that least one of those cases, I would assert you're incorrect.
Something that I think has gotten lost in the shuffle is the fact that the problem with MCOC's AI is not, and has never been that it is "too hard." Players regularly win against even the hardest fights. The problem is the AI is not intrinsically predictable, which means when a glitch or error is detected and reported, if behavior is seen that is not desirable, the devs have no easy way to address it. If the AI does something unintended, it cannot easily be told to not do that. AI 2.0 is not being written to make the AI smarter or dumber, faster or slower. It is being written to make the AI more predictable, and have more tuning knobs to make that predictability more controllable.
The whole point to AI 2.0 is if the devs see the AI doing something to the players they don't want it to do, they can just push a button and tell it to stop doing it. It doesn't matter if AI 2.0 is faster or slower. What matters is if it does what it is told to do. Right now, most of its behavior is either random or emergent - it randomly picks what to do, and the options for what it can do are determined by intrinsic situation and state which are complex and variable. It cannot be told "not do this over and over" or "don't immediately intercept the player when they do XYZ" because they can't turn those actions off completely, and they can't account for all the myriad ways the circumstances that could allow this to happen might arise.
Something I want to bring up real quick after scrolling through this thread and the Reddit post again: MLLLH combos are not supposed to be a thing. That’s why characters like Doom, Wasp, and Silk all apply a passive stun if they attempt to throw a heavy attack mid-combo; the heavy attack wouldn’t land otherwise. If you think this is inhuman reaction speed, I absolutely assure you that if an AI landed a combo on me and threw a heavy attack in the middle of it, I would 100% be able to dodge out of it. The only reason why this combo worked for you in early acts is because the AI in early acts is absolutely horrendous and plays like it doesn’t know what any of the buttons do.
Injustice AI is terrible. There you can't even completely avoid specials. In MCoC even matchups where you have a significant disadvantage by rating can be won with skill.
In Injustice that is impossible. The AI there can keep hitting you without giving you time to recover
Something I want to bring up real quick after scrolling through this thread and the Reddit post again: MLLLH combos are not supposed to be a thing. That’s why characters like Doom, Wasp, and Silk all apply a passive stun if they attempt to throw a heavy attack mid-combo; the heavy attack wouldn’t land otherwise. If you think this is inhuman reaction speed, I absolutely assure you that if an AI landed a combo on me and threw a heavy attack in the middle of it, I would 100% be able to dodge out of it. The only reason why this combo worked for you in early acts is because the AI in early acts is absolutely horrendous and plays like it doesn’t know what any of the buttons do.
To add,you can still pull of the MLLLH combo in arena since the AI there is just weird.
You mentioned that AI 2.0 aims for predictability rather than adjusting reaction times. While predictability is valuable, it’s crucial to recognize that reaction time directly impacts fairness. If the AI can react within 16.7ms (frame-perfect at 60fps), and human players average 150–250ms, this disparity can lead to situations where: • Players are punished for correct play, such as being blocked immediately after trying to punish a successful special dex. • The AI can counter mid-animation, making certain strategies ineffective (eg relics somehow being blocked on hit 2/3.)
This isn’t about making the AI easier but ensuring that skillful play is rewarded, not negated by inhuman reaction speeds.
Regarding MLLLH combos, the concern isn’t about specific combo sequences but about the AI’s ability to react instantly. When the AI can: • Block or counter between hits in a combo. • Intercept dash-ins immediately after a special.
It creates an environment where players can’t rely on established strategies, not due to their skill but because of the AI’s unnatural speed.
While Injustice’s AI may have its challenges, the broader point is that most competitive fighting games implement AI reaction delays to maintain fairness. For instance: • Tekken and Street Fighter introduce deliberate input delays to simulate human-like reactions. • Skullgirls allows players to adjust input delay buffering, acknowledging the importance of reaction time in gameplay balance.
These design choices ensure that players are tested on strategy and skill, not on competing with machine-level reflexes.
____
In Summary:
The core issue isn’t about the AI being too hard but about it being unfairly fast. To foster a competitive and enjoyable environment: • Implementing a reaction delay for AI actions can bridge the gap between human and AI capabilities. • Ensuring that correct play is consistently rewarded will enhance player satisfaction and skill development.
By addressing these concerns, we can move towards a more balanced and engaging gameplay experience.
• Players are punished for correct play, such as being blocked immediately after trying to punish a successful special dex. • The AI can counter mid-animation, making certain strategies ineffective (eg relics somehow being blocked on hit 2/3.)
The first one is not a speed thing. It's possible for us to do too. It depends on the animations of the specials. You can throw cgrs special Into an opponents block and you will not be punished even if they block it.
The relic whiff is a bug unrelated to AI speed. If we could be hit by relics, we can do it too.
You mentioned that AI 2.0 aims for predictability rather than adjusting reaction times. While predictability is valuable, it’s crucial to recognize that reaction time directly impacts fairness. If the AI can react within 16.7ms (frame-perfect at 60fps), and human players average 150–250ms, this disparity can lead to situations where: • Players are punished for correct play, such as being blocked immediately after trying to punish a successful special dex. • The AI can counter mid-animation, making certain strategies ineffective (eg relics somehow being blocked on hit 2/3.)
This isn’t about making the AI easier but ensuring that skillful play is rewarded, not negated by inhuman reaction speeds.
Regarding MLLLH combos, the concern isn’t about specific combo sequences but about the AI’s ability to react instantly. When the AI can: • Block or counter between hits in a combo. • Intercept dash-ins immediately after a special.
It creates an environment where players can’t rely on established strategies, not due to their skill but because of the AI’s unnatural speed.
While Injustice’s AI may have its challenges, the broader point is that most competitive fighting games implement AI reaction delays to maintain fairness. For instance: • Tekken and Street Fighter introduce deliberate input delays to simulate human-like reactions. • Skullgirls allows players to adjust input delay buffering, acknowledging the importance of reaction time in gameplay balance.
These design choices ensure that players are tested on strategy and skill, not on competing with machine-level reflexes.
____
In Summary:
The core issue isn’t about the AI being too hard but about it being unfairly fast. To foster a competitive and enjoyable environment: • Implementing a reaction delay for AI actions can bridge the gap between human and AI capabilities. • Ensuring that correct play is consistently rewarded will enhance player satisfaction and skill development.
By addressing these concerns, we can move towards a more balanced and engaging gameplay experience.
⸻
If you’re going to post relentlessly about this without actually processing anyone else’s comments, just start all your comments with
“FOR ME, the most important issue in recorded history is the speed of the AI…”
Regarding MLLLH combos, the concern isn’t about specific combo sequences but about the AI’s ability to react instantly. When the AI can: • Block or counter between hits in a combo. • Intercept dash-ins immediately after a special.
It creates an environment where players can’t rely on established strategies, not due to their skill but because of the AI’s unnatural speed.
The AI isn’t breaking any of the game’s fundamental rules to do any of those things. I could do all of those things, too, if it’s possible. If you’re giving the AI an opening to block your hits mid-combo then you’re not executing your combo fast enough. The AI is able to intercept your dash-ins after specials because of the recovery time after certain special attacks. Some special attacks have a very short recovery time which means you can’t punish them. It has nothing to do with the AI’s speed, the things you mentioned are all possible for a regular player to do.
Comments
I'll try one last time. You think the difference between MCOC AI and other fighting game AI is that MCOC makes instantaneous decisions with perfect information and other fighting games do not. You think other fighting games somehow implement an AI that is "slow" and "ignorant" - that takes in data and then just kind of sits on it pondering, and then some time later makes a decision. But I'm pretty sure no fighting game AI works that way, because that would be nonsensical. Instead, they all take in data and decide, instantaneously, what to do about what they see just like MCOC. But those AIs have more sophisticated options available to them that allow them to decide, instantly, right now, to take actions in the future. They can decide now to react ten frames from now. They can decide to execute a reaction script that has built in delays in them. They can decide to prioritize acting on what it remembers the player doing in the past rather than what it observes is happening right now. It is going to be just as fast, but what it does with that speed is going to be significantly different than what it does now. And because of that, there's no reason to complain about the speed. It is how it decides what to do that is more important than how fast it decides to do it, and that's what AI 2.0 is intended to address.
Last thing: the dev diary did not specifically state that Kabam doesn't want to eliminate the button mashing part of the AI on principle. Rather, they said this behavior would not radically change initially because changing the AI in a dramatic fashion would be disruptive to players. Players who have learned for years to intuitively play against an AI that is just a dumb button masher would find their intuition no longer working if they did that. So it seems that they will add the new AI features in stages, rolling out increasingly more sophisticated behaviors to see how they affect overall play. Whether they keep that indefinitely will depend largely on the way the playerbase responds to how the AI's behaviors change as they roll the new systems out.
You can see that the button mashing thing is in fact going to go away in the long run, because they explicitly state that one of the goals of the AI 2.0 roll out is to design the new system to roughly replicate the more twitchy random behavior of the current system, specifically to make sure players aren't caught off guard by an AI that does completely different things. But this step would not be necessary if AI 2.0 was still a random button masher. If it was, there would be no need to expend effort to try to get AI 2.0 to replicate the behavior of AI 1.0, and no worry it wouldn't do so exactly. It is a different system, but it is being designed to replicate the behavior of the old system to minimize the break in player expectations when it starts to get rolled out.
It’s not possible for us to tell exactly why those things happened that you described. I think it’s less a problem of speed and more a problem of “hitboxes” and hit/block-stun. If those are calibrated correctly, you won’t get those types of errors regardless of the rate of AI input.
It actually doesn’t matter if the AI can roll actions at 30x per second or more as long as it doesn’t do so in the wrong circumstances. If the AI dashes in, it makes no difference if it was frame perfect, you can parry it every time.
It’s the edge cases where the AI’s speed becomes a problem not in general play. I’ve given some examples but it’s really when the AI is right in your face or walking into that range. It’s impossible to react to a frame perfect input (if it rolls right).
So you should be less concerned about why it can act so fast and more concerned about why it’s allowed to get into range and take specific actions.
The AI in MCOC is really good. I’d even say great with some really specific pain points.
It’s not about edge cases or reaction vs. prediction — it’s that the AI can always act with perfect information, instantly, every 16.7ms.
In fighting games like Street Fighter, Tekken, Mortal Kombat, Injustice Mobile, even at the highest AI levels, the AI has baked-in reaction delays to simulate human limitations. They do that on purpose — because they understand that if you allow the AI to react frame-perfectly with no cap, you eliminate the concept of “earned” openings entirely.
You mention that if hitstun or blockstun were calibrated better, it wouldn’t matter how often the AI rolls actions.
I get the logic — but in practice, if the AI can always test every available option every frame, and instantly switch gears with no human hesitation, you’ll still always run into BS scenarios:
• specials blocked mid-combo,
• relics getting blocked mid-string,
• dash intercepts at ranges we physically can’t react to
•. Ai light intercepting you out of nowhere
• parry “fails” that aren’t failed inputs, just losing by a single frame.
These aren’t just bugs or quirks. They’re symptoms of a system that’s fundamentally skewed.
I 100% agree it feels great when it works — but when it fails, it’s not because of behavior tuning or hitbox issues. It’s because the AI is playing on a completely different reaction timeline than we are, and with the addition of 60fps from 30fps, this issue has only been exacerbated further.
Bottom line:
“Smart AI” is good. “Frame-perfect AI” is not.
If Kabam truly wants AI 2.0 to feel rewarding and fair, they’ll need to limit reaction speed like every other serious fighting game does — not just tweak behaviors or ranges.
Thanks again for the conversation. These discussions are exactly why it’s important to bring this stuff up before AI 2.0 launches, not after.
I actually think we agree on more than it sounds like. But there’s still a core difference that matters hugely.
You’re right that smarter AI is the goal. Smarter = good. Smarter AI that can plan, recognize patterns, remember past behavior — all great things.
But:
Even a smarter AI that still reacts instantly on a frame-by-frame basis will feel cracked and unfair, because it will capitalize on gaps humans physically can’t avoid. That’s the danger.
Other fighting games like Tekken, Mortal Kombat, Injustice Mobile do react instantly in decision-making, but they deliberately add reaction delays, buffered moves, or cooldown states after that decision.
The games don’t let the AI instantly block, intercept, or counter off a 1-frame window.
That’s why human players still have windows to bait, punish, and fight back.
It’s planned fairness, not just smarter thinking.
In MCOC, unless they add some kind of reaction timing limit or decision buffer, smarter AI will still abuse the raw frame-perfect system — it will just abuse it more intelligently.
That’s the real risk.
The AI doesn’t have to be malicious — just mathematically faster and sharper than us every single time. And that would make top-end gameplay feel oppressive, not rewarding.
Also — I get what you’re saying about them trying to minimize disruption at first. But Kabam’s wording was careful:
They said they want to replicate the random/twitchy behavior “at first” so players aren’t caught off guard — but they didn’t promise to change the fundamental speed mechanic behind it.
That’s why raising the issue now matters so much.
Once AI 2.0 is fully live everywhere, it’ll be much harder to argue for deep systemic changes like speed-limiting.
And it could define how the next 5–10 years of the contest feels to play.
And just to be clear — if this sounds repetitive from me, I get it, and I’m sorry for hammering it so hard.
But it’s because this is the fundamental issue underneath so many things players complain about — parry inconsistencies, specials being blocked mid-combo, instant intercepts, light attacks through relics, and so on.
It’s not about Kabam acting maliciously — I genuinely think, as with other things recently, they just kicked the ball and didn’t look where it would land.
60fps was meant to make the game smoother — but because the AI works on a frame-by-frame basis, now it’s acting twice as fast as before.
And if we’re going to keep 60fps (which we obviously should), we have to address the frame-perfect reaction speed alongside it. Otherwise the problem only grows worse with smarter AI brains.
So I’m pressing this point now — before AI 2.0 is fully implemented — because once it’s baked into the system, it will be almost impossible to untangle.
And it could be years, maybe decades, before we ever get the chance to fix it again — if ever.
That’s why I care about this so much. That’s why it needs community pressure now — not after the fact when it’s already too late.
At this point, I think it would be really helpful if someone like @kabamdork or @KabamPinwheel could weigh in with clarification — just so we know whether the frame-perfect reaction mechanic is intended to remain permanently, or if it’s something still open for review as AI 2.0 evolves. I also made a Reddit post on this which is now at about 23,000 views, with a lot of players sharing examples and discussing this. I’ve linked this forum thread there too so people can easily go between them. Would be great if the mods could weigh in here to give the community a bit of clarification — I’ll link the Reddit post below for context.
Would really appreciate any insight they can share.
https://www.reddit.com/r/ContestOfChampions/s/k2tqaqJ741
Thanks again for the thoughtful responses everyone. Even where we disagree on some of the points, it’s clear we all want the same thing — a game that challenges us, but is fair and doesn’t cheat us.
You'll essentially make sure every power sting Champion becomes worthless.
Nobody’s asking for the AI to “know more.”
We’re asking for it to react fairly — with human-like reaction time — instead of “frame-perfect,” instant decisions.
It’s about speed, not behavior or buff awareness. Two completely different issues.
I'm saying that if we're using fairness as a metric, then it's only fair that the AI has the advantages that we have too.
The goal isn’t to strip the AI of tools or give it player-specific buffs or knowledge.
It’s simply to balance the ‘fairness’ of reaction time so that fights are decided by skill, not by an AI that can act literally faster than humanly possible.
No other fighting game allows the AI to act faster than humans because that’s literally bad game design, it’s that simple.
You’re trying to argue about “advantages” when the core issue is that timing windows themselves are broken. It’s about leveling the speed of interaction — not its “awareness” or “advantages.” If the AI reacts faster than we physically can, that’s not fairness — that’s rigging the timing of the entire fight.
@PickL1e89 — Nailed it in one.
Speed, not behavior. The frame-rate doubling to 60fps is what really exposed it. The AI was already faster than humans at 30fps — now it’s just fully cracked, and the issue shows way more blatantly.
Why is it okay that we have an unfair advantage over the AI, but the AI is not allowed to have any?
Nobody’s asking for the AI to be nerfed, or to loose power/tools.
We’re asking for the reaction speed system — the foundation of how fights function — to be fair and human-like.
Players having access to buffs, masteries, or node knowledge isn’t the same thing.
Those are designed asymmetries. They’re intentional parts of how the contest is balanced — just like defenders getting nodes that give them massive boosts in attack, armor, power gain, regeneration, unblockable effects, you name it.
The AI already has huge advantages through nodes.
That’s normal, and nobody is complaining about that.
But the base timing rules — the fundamental ability to react to an event — should be fair for both sides.
If the AI can react frame-perfectly, and we physically can’t, that’s not skill testing — it’s a fundamental system-level imbalance and bad game design.
Every serious fighting game limits AI reaction timing to keep the contest skill-based.
We’re simply asking for MCOC to follow the same basic standard.
Skill should win fights — not literal robotic speed the player physically cannot match.
Decisive @ST4RL0RD08 Victory
Who would you rather be in that example?
You may not have the robotic reaction time but you have the luxury of actually remembering what has happened in the last couple seconds and that's plenty to bridge this gap you feel is unfair
@victor158 — Respectfully, you’re misunderstanding the core problem.
First — the AI already has built-in advantages, and that’s perfectly fine.
Just like we have masteries, relics, synergy bonuses, and node knowledge, the AI gets attack bonuses, regen nodes, unstoppable buffs, power gain, and so on.
Those are designed asymmetries — intentional parts of balance.
But mechanical asymmetry — reaction timing that humans physically cannot match — breaks the foundation of fair fights.
Players have natural input limits:
• Animation recovery (after throwing specials or attacks)
• Input lag (both device and engine-based)
• Physical human reaction time (~150–250ms+)
Meanwhile, the AI has zero such limits. It can instantly react — block, evade, intercept — within 16.7ms. No human can possibly match that, no matter how skilled.
To use a better analogy:
It’s like getting into a boxing match where your opponent can instantly dodge the second you throw a punch, but you have to wait a full second to even see their movement before you can react.
You can plan, strategize, be smart — but mechanically, you’re doomed to get hit more often than you land hits, simply because your body can’t react instantly like a machine.
Even if Kabam codes in the AI making occasional mistakes (“eg in the boxing match example it losing its footing”), it doesn’t matter if the AI can instantly recover faster than players can punish it.
• We’ve seen this when trying to punish specials:
You dex a special perfectly, dash in — but the AI immediately blocks your hit, because its reaction speed is frame-perfect.
• We’ve seen it when trying to combo into specials:
Some champs (with longer mediums) can’t safely chain medium–special attacks anymore because the AI blocks between the hits.
• We’ve seen it when dash-attacking:
Even if the AI looks aggressive, it can instantly backdraft intercept because it reacts on the frame you dash forward.
Correct play is being actively punished — and that breaks the basic premise of skill-based fighting.
In MCOC, historically, if you played correctly — dexed a special, timed a punish, spaced an intercept — you were rewarded with an opening.
Right now at 60fps, you’re often punished for playing correctly — because the AI is simply reacting faster than humanly possible. That same opening we used to have from fully dexing a special is now non existent or unsafe. Correct gameplay should NOT be punished.
Every serious fighting game in the world (Tekken, Mortal Kombat, Street Fighter, Skullgirls, Injustice/Injustice 2 mobile, God of War etc etc) enforces reaction delay on AI — to protect skill-based gameplay and avoid robotic unfairness.
MCOC needs to do the same.
Again —
We’re not asking for dumber AI.
We’re not asking for easier fights.
We’re asking for fair fights — where skill, prediction, timing, and correct play are rewarded, not punished by robotic speed.
2. Interesting that you mention Injustice. I played Injustice and Injustice 2. If any game has AI similar to MCOC's AI 1.0, it is Injustice. If anything, it is much less forgiving than MCOC's is, particularly in its defensive response speed. At highest difficulty the Injustice AI could respond instantly to player attacks, even moreso than MCOC's does/did. So in that least one of those cases, I would assert you're incorrect.
Something that I think has gotten lost in the shuffle is the fact that the problem with MCOC's AI is not, and has never been that it is "too hard." Players regularly win against even the hardest fights. The problem is the AI is not intrinsically predictable, which means when a glitch or error is detected and reported, if behavior is seen that is not desirable, the devs have no easy way to address it. If the AI does something unintended, it cannot easily be told to not do that. AI 2.0 is not being written to make the AI smarter or dumber, faster or slower. It is being written to make the AI more predictable, and have more tuning knobs to make that predictability more controllable.
The whole point to AI 2.0 is if the devs see the AI doing something to the players they don't want it to do, they can just push a button and tell it to stop doing it. It doesn't matter if AI 2.0 is faster or slower. What matters is if it does what it is told to do. Right now, most of its behavior is either random or emergent - it randomly picks what to do, and the options for what it can do are determined by intrinsic situation and state which are complex and variable. It cannot be told "not do this over and over" or "don't immediately intercept the player when they do XYZ" because they can't turn those actions off completely, and they can't account for all the myriad ways the circumstances that could allow this to happen might arise.
MLLLH combos are not supposed to be a thing. That’s why characters like Doom, Wasp, and Silk all apply a passive stun if they attempt to throw a heavy attack mid-combo; the heavy attack wouldn’t land otherwise. If you think this is inhuman reaction speed, I absolutely assure you that if an AI landed a combo on me and threw a heavy attack in the middle of it, I would 100% be able to dodge out of it. The only reason why this combo worked for you in early acts is because the AI in early acts is absolutely horrendous and plays like it doesn’t know what any of the buttons do.
In Injustice that is impossible. The AI there can keep hitting you without giving you time to recover
You mentioned that AI 2.0 aims for predictability rather than adjusting reaction times. While predictability is valuable, it’s crucial to recognize that reaction time directly impacts fairness. If the AI can react within 16.7ms (frame-perfect at 60fps), and human players average 150–250ms, this disparity can lead to situations where:
• Players are punished for correct play, such as being blocked immediately after trying to punish a successful special dex.
• The AI can counter mid-animation, making certain strategies ineffective (eg relics somehow being blocked on hit 2/3.)
This isn’t about making the AI easier but ensuring that skillful play is rewarded, not negated by inhuman reaction speeds.
⸻
@Toyota_2015:
Regarding MLLLH combos, the concern isn’t about specific combo sequences but about the AI’s ability to react instantly. When the AI can:
• Block or counter between hits in a combo.
• Intercept dash-ins immediately after a special.
It creates an environment where players can’t rely on established strategies, not due to their skill but because of the AI’s unnatural speed.
⸻
@Average_Desi:
While Injustice’s AI may have its challenges, the broader point is that most competitive fighting games implement AI reaction delays to maintain fairness. For instance:
• Tekken and Street Fighter introduce deliberate input delays to simulate human-like reactions.
• Skullgirls allows players to adjust input delay buffering, acknowledging the importance of reaction time in gameplay balance.
These design choices ensure that players are tested on strategy and skill, not on competing with machine-level reflexes.
____
In Summary:
The core issue isn’t about the AI being too hard but about it being unfairly fast. To foster a competitive and enjoyable environment:
• Implementing a reaction delay for AI actions can bridge the gap between human and AI capabilities.
• Ensuring that correct play is consistently rewarded will enhance player satisfaction and skill development.
By addressing these concerns, we can move towards a more balanced and engaging gameplay experience.
⸻
•Block or counter between hits in a combo.
• Intercept dash-ins immediately after a special.
Cause these are thing we as players can do reliably. So it's not a speed thing.
Also, could you clarify what you mean by the second point?
If you mean that the AI can throw a special and do a standup intercept, we can do that.
• The AI can counter mid-animation, making certain strategies ineffective (eg relics somehow being blocked on hit 2/3.)
The first one is not a speed thing. It's possible for us to do too. It depends on the animations of the specials. You can throw cgrs special Into an opponents block and you will not be punished even if they block it.
The relic whiff is a bug unrelated to AI speed. If we could be hit by relics, we can do it too.
“FOR ME, the most important issue in recorded history is the speed of the AI…”