Theories on how Hitstun Degradation Works

Hello all,

I’ve been playing MvC3 since it came out and I’m a little curious as to why no one has really taken the time to dissect exactly how hit stun degradation actually works. Knowing the exact mechanics could help in developing resets with variable timing, longer combos etc etc.

The common assumption would be that it’s somehow related to the number of hits in a combo or that certain moves contribute to further hitstun (throws being the obvious ones) while some ignore it completely (followup air combos to TAC) and others still, reset it completely (DHC glitches and tagging in other characters after throws).

I personally think it’s a combination of things. At first glance the most important factor might seem to be the number of hits, but I think it has more to do with a “hitstun timer.” That is, the longer a combo goes on, the more likely an opponent will flip out.

To illustrate this, I made a couple of videos:

In the first, I’m doing a typical Doctor Doom combo and stretching to the point where any hesitation will have the opponent flip out at the end. Notice that I’m airdashing down and whiffing his gun to land faster and “beat the clock” in order to land a few more hits than would normally be possible:


In the second example, I’m doing the same combo as before, but this time I’m letting Doctor Doom just fall down on his own after the sj.S. Notice that I’m unable to go for a fourth launch this time:


For the third example, I use the Doom missile glitch to illustrate that as long and the initial volley of hits is done in quick succession (40+ hits in this case), you can still get a full combo. This seems to indicate that the number of hits don’t actually contribute to hitstun decay per se, but rather, what’s important is how long it actually took to inflict those hits.


Other examples like being able to land a full combo after Iron Man’s level 3 point to the same conclusion:


Any thoughts on this?

My own experience with X-23 makes me think that there is a kind of Blazblue/Guilty Gear type thing going on where different hits prorate hitstun by different amounts. If I start a combo with c.L c.L c.L c.M (five hits) I can still do two full loops using s.H xx feint hop, but if I start that same combo with [s.H xx feint cancel] x 5 into s.H xx hop I can only do one loop before the other character pops out of the combo.

Maybe there is a combination of a kind of untechable window that starts at X value and slowly decreases over time (like you are suggesting), that is added to the normal prorated hitstun of a move? That would help explain why Magneto combos work the way they do… he uses a lot of fierce hits that apparently prorate very little, mostly out of superjumps (more hitstun on moves when in superjump) and he can do them all very quickly, so he can naturally take advantage of all the games combo systems perfectly…

I had a thread where I was documenting a bunch of hit stun times (mostly for Zero). Apparently it’s gone now after the forum change, but it was at, and still seems to be cached by Google, so I could repost if need be.

There are some factors besides just time. If you do a bunch of hits to a grounded opponent, it will cause less hit stun decay than if they were airborne. After a launcher, hit stun is increased by a certain amount. All moves also seem to have a minimum hit stun. TACs usually remove all hit stun decay until you hit the ground, but there are some exceptions, such as Trish’s Peekaboo. DHC glitches also don’t always remove 100% of the hit stun decay.

One thing I noticed later, which goes along with your findings, is that I can combo Zero’s cr. :m:, :h: (3 hits), :f::h: (3 hits), :l::m:, etc. only if I cancel the :h: during the hit pause of the third hit. If I wait at all, the :l: won’t have enough hit stun to allow the :m: to combo after it, even though the combo was exactly the same. This also probably means I’d need to go back through almost all the hit stun timings I recorded and make sure I was doing them as tightly as possible for them to be accurate. I also noticed that if I do a combo midscreen and use :m: Hienkyaku to get into position to do an :m: Sentsuizan for an assist relaunch, it causes more hit stun decay than if I had just done an :m: Sentsuizan in the corner.

Taking the X-23 loop example, I quickly found out that juggling after a Tron assist hit allows me to do two hops, but if I do it while juggling out of Doom’s beam, I can only do one hop, as the opponent will tech out before I can hit after the second. I do think this means that different attacks, as well as their properties (such as forced knockdown or something) affect the hitstun differently and it’s not really dependant on time.

At least for X-23. The time method seems to be true for Doom if we’re to look at those demonstrations. Perhaps different characters have different processes of hitstun degradation? If that is true, it might just be the cause for the DHC glitch.

I will agree with the hypothesis that hit-deterioration is based off of a timer, and that the amount of time spent kocked down also contributes to this timer. And I will also agree with Caudrath that ground hits don’t contribute too much in hit-stun deterioration in the air. Though I will go even further in thinking that hits on a standing opponent and hits on an airborne opponent, in fact, have separate timers entirely.

I would also like to add that launchers seem to stop the timer until a certain duration or until certain actions are taken.

With Magneto, for example, I’m able to do:

:s: > j. :h: > addf. > j. :h: > :s: > j:h: > adf. > j. :m: > j.:h: > :qcb: :s: > j.:l: > j.:h: > adf.
> j.:h: > adf. > j.:h: > adf. > j.:h: > addf. > j.:h: > j.:s: > :h: > :qcb: :l:

However, if I forego the relaunch, I would have to actually reduce the amount of hits I could do compare to the combo just posted. In fact, I could even space the hits in the original relaunch combo, while not doing the relaunch would actually make it harder and less adaptable for less payoff.

And with Ironman, I’ve been told that you can space some hits in an air combo after a relaunch. However, if you include an airdash or double jump as part of the relaunch combo, you wouldn’t be able to space the hits even if you add fewer hits and even if you were in the air in less time. This could mean that certain jump actions, like airdash or double jump, deteriorate air combos even further on a launch.

you should note that some characters have separate ground and air hitstun deterioration values, while others are shared.

Hitstun is based on a timer, with some moves ignoring the timer (hard knock downs) and others reducing the time of the hitstun.

If you do a really long ground string with Zero while keeping the opponent grounded the entire time, then launch and immediately check the hit stun on the next attack, it will have more hit stun taken off than if you had done a shorter ground string, so the timers aren’t completely separate.

Sougenmu {lmhx3,f.hx3,lmhx3}, f.hx3,launch,l (29 hits): 15 hit stun
lmf.hx3,lmhx3,f.hx3,launch,l (15 hits): 15 hit stun
compared to
lmhx3,f.hx3,launch,l (10 hits): 16 hit stun
launch, l (2 hits): 17 hit stun

Are you talking about hitstun values or hitstun deterioration values (if such a thing really exists)? I think it’s important to try and make the distinction.

All moves seem to have a set base hitstun duration (and minimum hitstun duration) which is then affected by a number of modifiers (timer, if it was started off a throw, air hit).

What I’m trying to figure out is what all these modifiers are and how they are applied.

For the record I’m also almost positive DHC’s reset hitstun degradation completely. The reason it might not always seem that way is because the timer resets the moment the opponent goes into the spinning reel and most characters have to wait a bit before being able to combo some more. That is, the hitstun timer keeps going before you even land the “first hit” of the combo so to speak.

Same goes for tagging in after something like one of She-Hulk’s command grabs.

It’s the hit stun value. The hit pause frame + the number of frames before the flip recovery animation occurs.

And, I take it, not how that one hit will affect the hitstun inflicted by any subsequent hits.

Regarding “minimum hitstun duration”, can you give an example of two moves (that do not ignore hitstun decay) that have different value for this? I’m just not sure if this is actually a value that differs among these types of moves.

I see it like this, there are two types of moves (in regards to hitstun decay):

  • moves that ignore it completely (always do the same amount of hitstun, or the same knockdown)
  • moves that have a base histun amount, which gradually decreases over time, eventually to basically nothing.

I don’t really see any overlap between these two types (you’re either one or the other!)

This is not to say that all moves deteriorate at the same rate, or are equally deteriorated by how much time has passed, but with the exception of moves that ignore hitstun-decay completely, I think they will all get down to the same point eventually (basically nothing, and they immediately air recover).

So, minimum hitstun duration really only exists for moves that are already at their *maximum *hitstun, as these are the moves that ignore hitstun decay completely.

Seems like it would be incredibly tedious to test this. You’d probably need to do a few TAC’s and have an assist that would allow you to land said moves after the OTG.

You might be right that all moves with no special properties end up having the same minimum hitstun value once all modifiers have been maxed out. Care to test it?

If the minimum were universal, then the infinites that were removed from the game would have affected all characters. For example, Captain America’s Shield Slash clearly has a lower minimum hitstun than it did before, but that change didn’t affect any other moves or characters.

Well, there probably is a true minimum, just long enough for the hit to register and stuff… but I haven’t seen this differ amongst moves that are affected by hitstun deterioration in the first place

As for testing, I’ve tested it with MODOK? :slight_smile: Can you think of any other moves affected by hitstun deterioration that might have a minimum hitstun that differs from the rest? I can’t… I don’t mind testing though if you do!

I don’t think this is true actually (I also play Cap! :))

Shield slashes are not affected by hitstun deterioration. Each hit always does the same amount of hitstun, and that’s it.

To remove his infinite, all they did was change his grounded H-slash so that it is now affected by hitstun deterioration. The base hitstun is still the same, but it now deteriorates over time, to the universal minimum of basically nothing. (The other slashes remain the same, and ignore hitstun deterioration). So there is really no reason this would affect any other character.

Hmm, how about Thor’s st.H? Maybe it just has a crapton of base hitstun, but it’s not a hard knockdown and they always seem to fly at least some distance after you land it.

Sent Drones assist also always seem to connect no matter how late in a combo you do them and they don’t exactly hit in quick succession nor do they create a hard knockdown.

I would check Akuma’s Air Tatsumaki Zankukyaku, Captain America’s Shield Slash, and Amaterasu’s f+HHHHH, since we know they were changed.

hmmm… cant be bothered to read the whole thread, so sorry if all this has already been hypothesized and whatnot or already stated, but imho hitstun degradation has a BUNCH of factors that all start to kick in depending on what type of combo is used:

repeated moves: this seems to be one of the biggest ones that will increase the hell out of hitstun deterioration. multiple l’s just seem to have lots of hitstun deterioration because they are the easiest moves to repeat in a combo.

air and ground: these have differences in deterioration. transferring from ground to air seems to “give back” alot of deterioration.

moves that give back: i think alot of moves give back (example they subtract from hitstun depending on when they’ve been done in a combo, and how many times they’ve been done in a combo. some examples would be launchers and h attacks, and certain special moves.

some moves arent affected by hitstun UNTIL THE END of them (ammys F+H series pre-patch being a very good example) i think this was so that certain multi hit moves wouldnt stop comboing after the first hits connected… imagine using ammy or akuma and having your F+h series or tatsu stop comboing in the middle and getting punished on block cause of it…

and of course some moves just straight up cause alot of deterioration.

i like the op’s hypothesis that the longer a combo goes on the more deterioration it gains as well. this seems very plausible. it would mean that a ticker starts counting very slowly at the beginning of everycombo and would reduce hitstun for all moves (or only certain ones) every frame… like maybe for evry 10-30 frames in hitstun the combod character would take 1 frame less of hitstun for every hit… depending on how this was implemented it would be a very strong deterrent to infinites and long ass loops.

anywho, all i know is that whatever is the actual hitstun formula… its redundantly complicated, and unless we actually get in the guts of the game it will probably never be fully figured out… capcom probably has a bunch of bogeys programmed in to have random hitstun (within certain limits of course) to keep us on our toes. in fact capcom could program any number of moves to have random hitstun properties besides just the actual hitstun.


Capcom’s ninja nerfed characters in patches before via this, Viper now has to rely on a different combo than she used to thanks to some change in her B&B that I can’t find where the adjustment came from but hitting the near the end B, B, C, 9, B, C, S simply doesn’t work any more and needs to be replaced with B, B, C, 9, C, S. =/

I want to think that it might actually be quite simple.

  • each move is assigned a base hitstun value that can vary greatly (launchers vs jabs, etc)
  • as time progresses during a combo, these values will be reduced (until they reach the a universal minimum?)
  • however, some moves are completely immune to the timer’s effects, and always do base value no matter what

…and that’s it. No more, no less.

Some preliminary frame counting leads me to believe that the universal minimum histun on normals might be 4 frames?
1 frame of initial hit + X number of hitstop frames (move dependant) + 4 frames of actual histun