Shields: Everything you need to know about them

[Battlecow]

Yes, Pete's correct. Teleport-Usmash shieldbreak/Usmash against wall shieldbreak is a well-known awesomething, you'd know that if you watched more VIDS.

Think I brought that up last time we were discussin' your theory.

 

[The Star King]

Yes, Pete's correct. Teleport-Usmash shieldbreak/Usmash against wall shieldbreak is a well-known awesomething, you'd know that if you watched more VIDS.

I've seen it, fool. That doesn't make it a sure thing. It could have been slightly damaged before the Up-Smash broke the shield; remember, it would only take 5% damage on the shield for the Up-Smash to break.

However, I tested it and I think pete's right. You win this round.

Think I brought that up last time we were discussin' your theory.

Yeah, you did, when you used Samus's Up-Smash to try and debunk my 55% theory. Who turned out to be right? YAWN

Spoiler

You're probably right this time, though.

 

[Battlecow]

ME! I TURNED OUT TO BE RIGHT!

I'm so clevar.

 

[The Star King]

I conceded. What more do you want?

U mad?

 

[Sangoku]

I haven't checked samus usmash yet but I wasn't so convinced about your theory =S.

I'm surprised not more people commented on the last part =(. Was that already common knowledge?

 

[Battlecow]

I didn't know the frame data, but I knew it recovered to ~half and that mashing/high damage would get you out faster.

But yeah, really nice stuff to have REALLY nice. It's great of you to work on this; I'm fascinated. Anyways, someone should check out whatever the melee equivalent of this thread is and see how their shields work; the comparison would be interesting (might shine light on why there's so much shieldstun/shield breaking in 64).

Also, it might be worth your while to investigate how much extra attack lag/hitstun/whateveritscalled there is for the attacker after they hit shield vs after they hit enemy.

 

[Sangoku]

Thanks for the input! I personally didn't know mashing was helping lol.

Concerning Melee, their shield obviously drains much quicker, but doesn't seem to be affected that much by attacks (both HPs and shieldstun). I might take a look at their boards for further details.

Concerning the attacker's hitlag I don't think there is a difference between attacking a shield or an opponent. I will check it nevertheless.

 

[Blue Yoshi]

TAS is so hard lol.

Anyways, it's late right now, but I'll start doing useful stuff soon.

 

[Supreme Dirt]

Coincidentally, 30 shield HP is the same amount as is programmed in Brawl, though Brawl has 50 initial HP.

So no need for an extra code, yay/

 

[The Star King]

Yesterday I broke Legendary's Falcon shield with Yoshi on the very edge of DL, and instead of landing on the stage, he barely missed the edge and fell down helplessly to his death.

I thought it was funneh.

 

[dandan]

my best shield breaker was against kabal, his shield was hit a lil beforehand and on the middle of hyrule, i reflected his full charged shot and he shielded and i managed to tele usmash to break his shield.
actually tried it afterwards with mahie and it does not work alone, but you can shine then tele sh fair usamsh and it does work.

 

[Battlecow]

I'm sure that there is a difference between hitting attacker and shield. Check it out.

 

[firo]

This is all very interesting stuff.

Mind sharing the memory addresses for displaying the shield hp, shielbreak recovery, and other stuff (if you are using them/they exist)? It would be interesting to check it out.

I think there are a lot of other interesting questions associated with shields, like how the radius of the shield is affected by how much hp you have.

 

[Sangoku]

There is something I should have mentioned in the OP, sorry. All the values I get are empirical results. I'm not a programmer / hacker or whatever so the only thing I can do is try to describe what I see.

Fortunately everytime I got some results I could elaborate simple theory which seemed to work.

This obviously means that I could be all wrong. For example, the shield's natural drain could not represent the smallest reduction from shield (that is the unit of the shield's health). However I would have noticed quickly if some attacks' power weren't a multiplication of the unit I defined. This didn't turned out to be true (all the powers I've checked so far are integers), therefore I concluded my theory was correct.

My empirical method is obviously not the optimal one (longer, can lead to error, ...), but since we don't have anything better (yet?) that's the one I'll use. And as I said, I don't know anything about programming and to be honest I don't even know what are the memory addresses you mentioned. I remember I once saw a post by antd where he was talking about z-cancelling, that there was some number that was reset after pressing z or something... Well what I mean is (as mentioned in the OP), I'm not as good as antd, I have less experience in TASing and I don't know much about the real game's mechanics. If you know more than me (which is highly probable since I basically know nothing) and you don't mind explaining the basics in simple terms I'd be really grateful.

The shield's radius depending on the HP was something I would have liked to look at too, but with my method it would only be approximation (I could only count the pixels, but that number varies depending on how zoomed in it is). I could however define ratios. For example how many percent the shield reduces (in size) each time one HP leaves. Or compare full shield versus smallest shield. I don't know if there will be some general pattern though. For example jigglypuff's shields reduces in size by much more than falcon's going from full shield to almost broken one.

 

[Supreme Dirt]

Another thing which helps is that all this data makes sense when I look at the actual values in Brawl. And putting the values into Brawl with codes basically gives me 64 shield mechanics, minus shield damage because I have to do that on a move-by-move basis.

 

[Battlecow]

I'm kinda pumped for Brawl 64 now.

 

[Supreme Dirt]

It'll be a long while in coming.

Everyone who joins the team ends up flaking.

Though I do have someone who is supposedly working on a disassembly of 64 right now.

 

[SSBPete]

You have to start somewhere Sangoku and I think this is a great place to start.
Really love the effort your putting into this, keep it up

 

[Kabutox]

This is a great start Squall! If you have any questions you can always ask me.
Also, you should check out fox's jab SS.

(and if you could maybe count the frames before you can do another jab, just for me? <3)

Could you maybe explain me how you measured the amount of moves that are needed to break the shield? Like fox's lazer, how did you leave out the 'time-shield-drain-factor'?

Yesterday I broke Legendary's Falcon shield with Yoshi on the very edge of DL, and instead of landing on the stage, he barely missed the edge and fell down helplessly to his death.

I thought it was funneh.

I actually have me doing that on video vs Olikus ^^

 

[Sangoku]

First thanks Pete and Fays for your support lol. Second I apologise for not providing the weekly update I hoped I could make =/.

Fays: I'll check fox's jab as soon as possible =). To count the NNB you have two ways: either you determine the attack's power and divide 55 by it (and round off to the greater integer or apply the roof function or whatever lol). Or you can use the "shields don't drain over time" cheat so you can spam lazors =P. You can find this cheat in the big cheat file gaudy uploaded.

 

[firo]

So there is a much easier way to do this then what you are doing now...

I was able to find the memory address in cheat engine of player 1's shield. In versus mode, its 4b952f84 You can even mess with it and make your shield huge and take up the entire screen.

If you use cheat engine and enter this address manually, and also use the gameshark code so it doesn't reduce over time, you can find out the attack values and such. So far your information seems correct. A bit later maybe I can explain a bit more in depth on how to do this so you can use it to gather more info.

 

[Sangoku]

Concerning all the properties and "behavior" of the shield, I think we know everything already (Blue Yoshi made a last discovery, which he will hopefully explain soon in the first post) so no need for the "cleaner" method.

However, concerning all the frame datas, if your technique is faster than ours (which it probably is) then that would be great =). That is really nice from you to have gone into this and I would really appreciate if you could explain your method (knowing I know nothing lol).

 

[Blue Yoshi]

The only thing that I am having troubles with is the HP calculation. The way I am doing it right now (which is the way Sangoku told me, so I assume he is doing it too) is looking for when the shield shrinks. However, I'm worried I might miss the first time it does that and may be one or maybe two frames off in the final result. From what you are saying Firo, if I understand correctly, you are saying there is an easier way to determine the HP loss by using the memory location. If that is the case, then I think that would be MUCH simpler.

Though I think with further tests, it is possible that there may be consistent results as to when this first shrinking of the shield happens (e.g. X frames after shield stun). But still, having a second method to double check is always good.

As for this discovery, I'll try to re-word it properly (describing it to Sangoku is easy because I can use frame data, and he can test it on TAS to see what I am talking about. But describing it here... I may need to record it somehow and post it that way). That said, time to ask something which is better fit in the Q&A thread...



Edit: Well, I'll post both of my messages I sent about this. If people want to read alot, you can read them. Since it's 2am right now, I don't really feel like summarizing it right now (kinda wanna go to bed), but tomorrow, I'll summarize it all.



Message 1:

Spoiler

I discovered something which may need to be looked into, and may change everything (or at least become a major factor).

I found this by using Jigglypuff vs Jigglypuff (using the first jab move). All frame values shown (frame 13, etc.) are frames after the white shield (i.e. the number shown at the bottom left of the screen when starting a recording on the frame where you see the white shield).



What I have been doing so far as far as shield stun goes is what you said. I press shield, once the white frame shows up, I let go of shield, start the recorder, and proceed frame by frame. Normally I would stop when the shield disappears (takes 22 frames). All was good, or so I thought.

I noticed that Puff starts moving around frame 14, in what seems like a shield dropping animation, but still with the shield throughout. I didn't think of it too much at first and moved on to other moves, but then, several moves "completed" later, I decided to test it out just in case. I first used 3 Jigglypuffs, had one jab like I usually do, but this time I counted the frames, and had the third Jigglypuff jab the target puff during this "shield drop" animation. Turned out that I had jabbed her late in her animation, but since I hit the shield, she returned to her shield stun stance. What was very interesting is that the second time she entered her "shield drop" animation (after the second jab), instead of dropping her shield on frame 22 like she usually does with one jab, she actually dropped it way sooner (frame 15-17, wasn't counting), but continued her animation just without a shield.

So I began testing to see if you could input commands during this time. I soon found out that if any command were to work, you had to hold shield, or else she would enter her standard shield drop animation. I tried jumping out of shield, rolling, and grabbing. It turned out that for Puff's jab, while I originally thought that there was shield stun until frame 22, it turned out that you could actually input a command on frame 13 (provided you can perform it in a shield, e.g. roll, grab, or jump), and on frame 14, the action would come out. On the same lines, we can include shield dropping in this category too, as it starts in frame 14, but has a different effect.

So now, with all that said, what I initially thought was a 4 frame advantage for the attacker (attacker finishes on frame 18, thought shield ended on frame 22) actually turns out to be a 4 frame advantage for the shielder (shielder can make actions starting on frame 14). However, still being in the shield, the options available are limited, and depending on the character shielding, it may not be an advantage after all).



So therefore, just because the shield hasn't disappeared yet, doesn't mean that the shield stun is not over yet. As in my example, Jigglypuff was out of shield stun on frame 14, even though the shield ended on 22. However, just because this takes 8 frames to happen, it may not always take 8 frames. As I noticed in my two attacking Jigglypuff example, when the shield leaves during the shield drop animation is not constant, as the first time, it disappeared on frame 22 of the jab, but the second time, it disappeared well earlier.

Another thing to note is that a character does have options available to them from a shield. In my example above, if on a platform, Jigglypuff can input "down" on frame 13, then "B" (while still holding down) on frame 14, resulting in a shield drop rest hitting on frame 15.



As annoying as it may sound, I think we will have to go through each move and find out when exactly an action is doable, and use that as the shield stun frame count. Attacker stun will be the same, but due to the new shield stun, the frame advantage will be different. It will be annoying, we'll have to do each move frame by frame and find when an input is doable, but that's the only way to get consistently accurate results.

Hopefully this isn't too much crazyness... but it is definitely something we need to include in our results.

Well, I'll talk to you later. As annoying as this is, I hope this helps lol.

Message 2:

Spoiler

After more testing, it's not nearly as bad as I thought it was, and actually adds an interesting new element to shields.

The "first frame action" is the earlier number (obviously, since you can perform a jump etc). However, what I noticed is that the later number that I have been getting is actually due to another factor which I figured out after lots of testing.

When you shield for one frame and let go, your shield stays out 8 frames. If you shield and hold shield (for more than 8 frames), as soon as you let go of shield, there will be no shield the next frame (shield drop animation lasts the same number of frames, but without the shield).

After testing it out with attacks, I figured out that the reason why there is a difference on when the shield gets removed when you let go of the shield button turns out to be because of one thing:

You must last 8 frames without shield stun before letting go of the shield will remove the shield. Otherwise, the remaining frames will continue during the shield dropping animation.

This number, however, adds up. If I held my shield for 2 frames, took a hit from an attack (experience shield stun), kept holding my shield for 2 frames, got hit by another attack (experience shield stun), I now have 4 more frames remaining before I can drop my shield via letting go of the shield button.

One thing to note: if you JUMP, ROLL, GRAB, or PLATFORM DROP (on platforms only), these 8 frames are ignored, and the action happens immediately (assuming you are out of shield stun frames, where you can input an action). The only way these actions can be performed are if the shield button is pressed down, otherwise, the standard "getting out of shield" animation will take place.



What does this mean for shield breaking?

Letting go of shield is NOT the best way to avoid getting your shield broken, as you will still have up to 8 additional frames where an opponent can hit you. Ironically, as said in the paragraph above, the soonest way to let go of your shield is by holding down shield, and performing one of the actions above (jump, roll, grab, or platform drop).

If you are against someone who is trying to escape by simply letting go of their shield, then you have an up to 8 frame window where you can mess up. However, lets assume that you do manage to hit on the first frame of the shield (therefore they have to endure the entire 8 frames), and your opponent lets go of their shield and does nothing else, and you perform an attack that lasts 3 frames longer than your opponent's shield stun frames, your opponent will now have lasted 3 of the 8 frames required to drop their shield via doing nothing. If you keep doing this, then on the fourth time you do this attack, your opponent would have lasted 9 un-stunned frames, therefore being able to drop their shield. This can explain why some shield breaking combos seem to work perfectly early on, but fail near the end of the combo (the opponent's 8 frames have passed).



Hope this explains it

If you have any questions, just ask me.



I am now curious as to what "standard" shield break combos are now no longer true shield break combos because of this.

I'll summarize it tomorrow sometime. But for now, that's what it is. I was going to make a video to show getting out of stuff faster among other things... but we'll see.



Edit: Added a huge chunk of text to the OP

 

[firo]

To help get the data you'll need Cheat Engine, or a program like it. You can google it and download it, its the first result.

Your first step is to start SSB and get into a versus mode match. Before you do anything open cheat engine, click the blinking computer icon in the top left, and choose the PJ64K smash window. Now you have two options:

You can click the "add address manually" button and input 4b952f84 which is the address for the hp of player one's shield. Alternatively, you can search for it, which will probably be a better way to learn how it works.

The process is pretty easy to understand. We are looking for where the value of your hp is stored, so assuming that you are at full hp, you want to put in 55 in the value field, and then hit first scan. A huge amount of values come up.

Next. put up your shield or get hit or something, and then go back to the cheat engine window. Now we know that your hp is lower than it started. So under scan type, choose decreased value. Then drop the shield and let it increase, and next scan should be increased value. You can do this until you'll only get one or a couple values on the left, and this is your shield hp. Double click it and it goes into the bottom. You can now set this value to what you want or freeze it. From here its only a few steps away from making a gameshark code.

Note that this method won't work for all values. It works for a lot of stuff in ssb, but if you check the value type (for this it should be 4 bytes) there are a lot of ways values can be stored. You can use this in conjunction with frame advance and gameshark codes. You can probably see that the process can become pretty complex for some stuff, but the shield hp shouldn't be hard to work with. You also don't have to do this with Project64k, any emulator will work.

For more info on this you can check out a video on youtube. There are a good amount of tutorials on this. Here is one: http://www.youtube.com/watch?v=fNaxReZm0h8

 

[TANK64]

Such a good read Blue Yoshi.
This thread is awesome.
Also so glad we have active TASers =)

 

[Blue Yoshi]

I always wanted to TAS, I just never had the right tools for it, and was too lazy to find them. Now that I am able to, I'll do a bunch of TAS-related stuff

 

[SilentSlayers]

I always wanted to TAS, I just never had the right tools for it, and was too lazy to find them. Now that I am able to, I'll do a bunch of TAS-related stuff

Contact me if you need any help with TASing.

 

[Sangoku]

Thank you very much firo. I haven't read it all yet, but I'm going to try this method as soon as I can. Hopefully this will save us a lot time.

 

[Blue Yoshi]

Assuming that Firo's method works, I already stopped counting the HP and am working only on stun frames and other relevant important info (basically everything except HP)

 

[Sangoku]

Okay so Firo's method turns out to be great =D. Really faster than my old stupid method. I've been able to almost finish Luigi in 20 minutes and I was still making mistakes so it should be even faster when I get used to it. Hopefully with the help of Blue Yoshi we will quickly finish to count shieldstun.

 

[firo]

Glad I could be of some help.

If anyone has any technical questions like this I'd be happy to help out. It's possible that you may be able to count the shield stun for the moves with this method as well, but that would require some more searching and frame advance (ie, progressing frame by frame after shield hit and searching for decreased value). That number might be 2 bits or something as well.

 

[Blue Yoshi]

My method for shield stun is actually pretty fast. But I really need another method to calculate HP.

 

[Sangoku]

I've finished Luigi, Mario and Donkey Kong already. Took me less than an hour in all.

Just to say how great that method is lol. I won't update the first post until I'm finished with all chars though. Any suggestions to make a nice table thing for the OP?

 

[Blue Yoshi]

I'd say by character. I've already got a table going for each character (I'll show you when I finish a little more), and we can go from there.

 

[Sangoku]

This is all really disturbing... Except for aerials that I haven't checked, all Falcon moves give the same % on opponent and on shield (number of % corresponding to HPs obviously) except dash attack and falcon kick. Can't find the logic =/. If only falcon punch was different too maybe The Star King's theory could have turned out to be true... And why dash attack? I will try to compare with every attacks until I find something coherent.

edit: yeah really no logic since falcon punch doesn't get a boost but full charged DK punch is 48 instead of 36 % (4/3 multiplier).

 

[The Star King]

Wait, wut? How does "all Falcon moves give the same % on opponent"? I think I'm misunderstanding you, because what that sounds like to me definitely isn't true.

If Samus's Up-Smash is indeed a shield-breaker than my theory is already ruined lol

 

[Battlecow]

They give same % on opponent and shield, meaning that the damage they do to opponents is the same as the damage they do to shields

 

[Sangoku]

Thanks battlecow and sorry Star King lol. Yeah that's what I meant. An attack that deals 16 % on opponent has a power of 16HP on shield.

And I thought we could have just modified a bit your theory for it to work, but I guess that's not the case.

 

[The Star King]

Yeah, sorry, I thought you meant that they all do the same %, like every move does 12% or something. My bad.

So dash attack does a different % on shield, compared to on the opponent? Are you sure you're comparing the same parts of the move, i.e. not the initial part of the move to the later part?

 

[StealthyGunnar]

This thread is awesome!

Thank you!