MvC2 System
100% Technique Explanation
First off, this isn't going to be very useful for anyone that isn't interested about how program pads work or how MvC2 works in terms of frame skipping and the cycles. If you don't have interest in the matter, you shouldn't read it. Also, there is another page/article explaining the basics of a programmable controller. I suggest you check that out before venturing into this one.
There is a way to ensure that the programmable controller will execute a combo 100% of the time, it also allows you to know when certain attacks/combos/whatever won't work. Up until now, programmable pads would sometimes 'mess up' the combo or situation they were told to execute. Programmable controllers are flawless, it's the video game that isn't truly 60fps. In order to get accurate results every time, the programmable controller needs to be calibrated every time before the main combo/situation. The way the solution works is, you insert a sequence before the actual combo (the 100% sequence and the actual combo are in the same overall sequence). This is a great tool for complicated and over elaborate combos that come up pretty often. This technique becomes invaluable when combos or situations get out of hand and have a less than 10% success rate or when trying out a short but complicated string.
Before I explain how the technique works, it's best to understand how MvC2 works in terms of frame skipping. MvC2 on Normal speed runs at 72 frames per second (last sheet), when you change it to Turbo you're forcing it to skip frames. So it runs like, "1, 2, 3, 4, 6, 7, 8, 9, 10, 12...". That shows the game will always display four consecutive frames before skipping one; the skipped frame is inaccessible (that is to say that if a just-frame occurs on this frame, it will not be registered). The frame skipping cycle begins as soon as you turn on the DC or after the loading screen. This doesn't affect inputs in a normal match but can be responsible for a few "errors". However, those situations are pretty rare. Usually it's the timing of 'reversals' that are affected the most, not combos.
An example of a combo or block string that would use this would be if someone wanted to link LP into LK as soon as possible, they'd need LP to be a displayed frame, the null frame would have to be in the time you have to wait before pressing LK, and the LK would have to be a displayed frame. Here's Storm doing her S.LP followed by a S.LK. Normally, the duration of the S.LP is 10 frames but it has a 1 in 4 chance of being 9 frames due to frame skipping. If you divide 9 by 4 you'll get a number that ends in .25 or 1/4th. That was the case in the Storm clip, she did a S.LP for 9 frames and pressed S.LK on the 10th frame instead of the 11th.
There are a few ways to calibrate the ppad, one method is having the character do a triangle jump as fast as possible. This poses a few problems because you'll need to do the triangle jump twice in a row (I'll explain in a bit) and not everyone in the game has a triangle jump. The reason you'd have to triangle jump twice due to the numbers in the sequence coming out to .5 or 1/2 the time. That is to say that you'd need to see the triangle jump happen twice in a row to know that you've landed on the correct frame.
Here's Magneto triangle jumping twice. This shows which frames were displayed in the game:
Magneto's Triangle Jump Sequence -
d = displayed, s = skipped
s--d---d--d---d---s---d---d----d-----d----s-----d----d-----d----d---s---d--d---d---d---s---d-----d-----d----d-------s----d-----d---d
n---2---8---8---8---8---8---8---2pp---n---hp---hp---hp---hp---2---8---8---8---8---8---8---8---2pp---n---hp---hp---hp---hp---n
If the displayed/skipped frames were off by just one, that sequence won't work. However, after performing that, I'd need to check if the sequence actually works by annexing another just-frame move. In this case, I'll use Magneto's Tempest super.
In order to make his Tempest special, I did a fake tiger-knee motion and held the UF part for 3 frames. The way this works is if the UF part is 1 or 2 frames, Magneto will not leave the ground, if it's 4 or higher, he'll leave the ground 100% of the time. However, if it's 3 frames there is a 75% chance that he will leave the ground and the special/super move won't come out. However, if he doesn't, that's an indication that you're on the right frame. This shows Magneto's previous 100% sequence followed up by a 'special' Tempest super.
The Tempest Portion -
d---s---d---d---d---d--s---d---d---d
n---n---n---2---1---4---7---7---7---kk
Since only 2 frames are displayed for the super jump, the character stays on the ground and does the special/super.
Below is a sequence of images showing each frame of the tempest start up:
1
2
3
4
5
6
Notice that from the 5th to 6th frame, the tempest occurred even though my input was to hold up-forward for 3 frames. This is because of the frame skipping taking place during 1 of the up-forward frames.
The second method used to adjust the program pad to the game's frame cycle was mentioned above, the fake tiger-knee method. Each method gives you the same end result except that the methods for doing it can be moderately unpleasant to downright annoying sometimes. Since having moves come out 100% of the time is a necessity in some cases, using the technique is a must. This shows Magneto doing a 'tiger-knee' Hypergrav followed by a 'tiger-knee' Tempest. If the tempest was attempted a frame earlier, it would have caused the frame cycle to be thrown off and thus not work. The problem with this method was that 75% of the time it would be a super-jump, this meant having to wait and move the character(s) back into position and try again. It wasn't crucial but it was definitely unpleasant. Furthermore, some of the characters have specials that will hit the opponent, this would really mess up situations where specific health was important to both characters. A work around for Storm was to do her lightning attack facing backwards, however, it would still cause her to SJ 75% of the time.
The third method is the less time consuming by far. This method was discovered by joo; he used the data he found for length of character's normals to deduce which normals will output the desired result. In my opinion it takes some insight to understand that you'll want a certain number that when divided by four will result in a number ending with .25 or one-fourth. If you use a number that ends in one-half you can still calibrate the program pad, it'll just take two repeats to narrow it down. .25 is the smallest number that exists in the game when dealing with probability, this is what gives away which moves to use for the calibration. Relying on a move that's 14 frames normally, but 1/4th of the time 13 (13/4=3.25) will save time in repeating the calibration process. If I were using a move that's 15 frames most of the time but sometimes 14 (14/4=3.5), I would require two consecutive results of the normal coming out on the 15th frame instead of the 16th. The numbers I'm interested in would be: 2, 6, 10, 14, 18, 22, 26, 30, 34, 38, 42, 46, 50, 54. Those numbers minus 1, divded by four would result in "_.25".
The only caviat with this method is some characters don't have normals that are any of those numbers so doing something that works 50% of the time twice is necessary. For example, Ryu: he does S.LK, S.LP twice. This isn't nearly as unpleasant as waiting for a character to land from his super-jump and walk him/her back to the corner. An interesting note to is that two of the characters have taunts that are 54 frames and 1/4th of the chance, 53. So even taunts could he helpful in calibrating the ppad.
The other aspect of using the 100% technique is you can actually control the frame skipping cycle that the game produces. The game's frame skipping cycle is set until you pause the game/change levels/restart/etc. If you pause before you execute a complicated sequence not using a 100% sequence before it, nothing will change. However, if you're using a 100% sequence before your actual combo and you pause the game, chances are you're sequence won't work 100% of the time.
I learned this the hard way when I was trying to execute something and I started to realize that it wouldn't work 100% of the time. I thought I had changed the wait time before the main combo (1-4 are the variables), but when I realized none would work, I was dismayed. I paused the game to change speeds because up until then I thought changing the game's speeds was the only way to restart the frame cycle. When it didn't work I paused the game again but I remembered I had to try the variables before I was certain the sequence wasn't going to work. So I un-paused the game and tried the sequence again and it worked! I eventually figured out that using pauses was all you needed to change the game's frame skip cycle. Using pauses will greatly help your chances of finding things in MvC2. If the frame cycle is set and you're using 100% technique, there is a small chance that you've landed on a frame that will prevent your combo from working 100% of the time. This is where pausing comes in. Inserting a pause for a few frames and unpausing it on a certain frame will allow the cycle to restart. This allows different results in sequences that use the 100% technique.
If I wanted to insert a pause to a technqiue I'd have to make the pause 4 frames, the neutral is 3 frames and the unpause 1 frame. This results in the cycle changing and therefore allowing different results. Really crazy and elaborate combos can come about from
this the only drawback is pausing causes the voice that was speaking to mute...which isn't really a problem. This shows Sentinel doing his cr.hp unblockable on Zangief with and without the pause before the 100% sequence.
That's everything regarding the 100% technique and how it works as far as I know. The technqiue is beyond useful when programming really elaborate stuff. However it doesn't eliminate all 'randomness' from the game, it can't do anything about mashing. Prime example of this rule are thorws. Sometimes throws will do absurd amounts of damage if mashed right; other times (even if mashed the same way every time) not. I'm not sure as to why but my guess is the answer is embedded somewhere in the programming; maybe it really is a truly random moment.
The other kind of mashing is supers and specials. Those can't be controlled either, that's kind of dissapointing since there are times I want to control AHVB or or Silver Samurai's Lightning super.
The other kind of randomness is due to the game slowing down during some statges causing ST-esque errors that do not make any sense at all. This makes emulation an interesting subject, if you were to have a super-powerful PC that can run magical emulators that allow the emulator to work just like a DC, you wouldn't get the random stuff that happens in normal matches on normal DCs or arcade hardware since your PC is too powerful to have slowdown due to characters calling assists/dhcing at the same time -- at the same time. Also, it probably wouldn't allow some useful (for combos) bugs to happen as well. Also, emulators are usually accompanied by emulation tools such as macros and save states, those would eliminate the randomness of mashing. Mashing and random supers wouldn't be random due to the save states. Anyway, I don't have any hope for emulators so I'll stop talking about this.
So if you're interested in seeing every character's 100% seequence, go here.
Special thanks goes to:
Joo
Maj
ECZangief
Powered by Zachd.com
