Project Database

[JOE!]

Ok, but how else could this be tested without factoring some "wiggle room"?

 

[rPSIvysaur]

By making it a freelance project.

If you want some form of support from the Smash Lab, you could get a few replays from pros and figure out the average variance from the actual FAF you get by setting up the same scenario. (it would talk a bit of work/time, but I would find that kind of stuff reliable)

 

[JOE!]

So how is the Smash Lab gonna gather accurate/practical weight or rather "survival" data?

 

[rPSIvysaur]

So how is the Smash Lab gonna gather accurate/practical weight or rather "survival" data?

'practical' - therein lies the problem. If we can find the approximate reaction to the ending of hitstun pros can do, then we can figure out the optimal DI. This would take a really long time too.

 

[JOE!]

true, but even if we find the "optimal" breaking time, we then have to figure out if it's humanly possible to react to that frame, and then figure out with that factored in if the frame difference between pro and avg is even significant to save you from X to Y %, or even if X-Y is a diff of like 1 or even 0.5% (this exists)

 

[rPSIvysaur]

I believe we should be using 'pro DI' otherwise we would get extremely inconsistent DI.

 

[JOE!]

allrighty then, so to do this I think we'd have to go around to each board and essentially ask them what are thebest ways for their characters to DI when launched from (Side/Up/Down/Angle/etc)?

then test those methods using frame data for when the aerial/spot dodge/etc occurs (how most DI is done) and go from there?

 

[rPSIvysaur]

I guess you would ask them, but we could also test it ourselves.

Then, I think it would be nice if we could get some replays of people like M2K, Ally, ADHD and other top players at interlaced 30 FPS or deinterlaced 60 FPS, so we can find how many frames after the hit they reacted. We then could reenact the scenario and find the FAF during the hitstun. We then would have a ball-park range of what is acceptable for human capacity.

 

[JOE!]

I think the best way to go about this would be to just have a community character by character basis...

first by finding kill % for each of their moves on each character (daaaaaaaaaaaaaaaaaammmnnnnnnnnn)

then, compiling that data per character to find out how long they live vs each character

then, find their "average survival" from that...

after we of course find out the DI stuff for both Pro and Avg to find out the "human spectrum" for optimal DI

 

[rPSIvysaur]

community character by character basis...

I'm a bit confused as to what this means. Please enlighten me.

 

[JOE!]

as in, we go to each character board and ask them what their best methods of DI are for theri character vs certain launch directions, instead of having to find out ourselves, as they know their chars better than we should...

 

[JOE!]

well, that is easy...

Bowser's best move for DI'ing is allways Fair.

(assuming we go over him 1st)

 

[Luxor]

Oh hey rPSCIENTIST, can you link me to anyplace detailing how to swap background textures, the grid texture, and give me a few sound words of sagely wisdom? kthxbai

 

[Indigo Jeans]

Actually his original name when he registered was rocketPSIence.

 

[rPSIvysaur]

Oh hey rPSCIENTIST, can you link me to anyplace detailing how to swap background textures, the grid texture, and give me a few sound words of sagely wisdom? kthxbai

Oh hey Luoidjspodifjsd :3c

http://www.smashboards.com/forumdisplay.php?f=232

 

[Luxor]

@Indigo: Aware of this. I liked the clever wordplay.

@rPSI: Looked there, can't find it. I even tried *gasp* the search bar. Although I am top-tier at overlooking things in plain sight IRL.

 

[rPSIvysaur]

It's called asking in there. lol

 

[Luxor]

Once again... overlooked.

 

[KayLo!]

I have the grid custom stage texture if you still need it.

 

[Luxor]

That'd be great.

 

[KayLo!]

http://www.megaupload.com/?d=M5JWEAC5

Goes over the futuristic custom stage parts.

 

[saviorslegacy]

I would like to research tethers. Review all that is known + this: http://www.youtube.com/watch?v=a8KkRGzPcRM&feature=related
I was looking into a little bit and found something out about Olimar. Everyone knows that if he jumps and times a tether properly he can tether the ledge but land on the stage.
Well, I figured out that he can cancel the end of this and go into his second jump with no lag. So it looks like he bounces off of the stage and into the air.


I just wanted to review and see if there are any easter eggs that have been over looked in tethers.

 

[rPSIvysaur]

Umm... first just ask around in here if anyone else knows mechanics and try to collaborate with them.

 

[saviorslegacy]

Umm... first just ask around in here if anyone else knows mechanics and try to collaborate with them.

True, I really didn't know how to go about getting research started on something.

Then the case is, "What do you all know about the vid and about that Olimar tech?"
It seems like Samus is tethering, barely touching the stage and then falling off. I tried it with other characters, but I couldn't get anyone else to do it. TL is very close though.
Opinions?
Also, what are some other things that you can do with tethers other than droping it from a distance and using the tether above the ledge to get its full length?

 

[rPSIvysaur]

Lucas can dusterslide. Video of frame by frame: http://www.youtube.com/watch?v=BCbnoXkavqE

For this project, I recommend gathering as many frame maps as possible to figure out what's going on.

 

[saviorslegacy]

Lucas can dusterslide. Video of frame by frame: http://www.youtube.com/watch?v=BCbnoXkavqE

For this project, I recommend gathering as many frame maps as possible to figure out what's going on.

Interesting

I have a pretty good idea what you are talking about but just to be 100% certain, what exactly is a frame map?

 

[rPSIvysaur]

Input maps are pretty much the best thing ever for frame data research. All my "raw" frame data test results are recorded in this format and it allows me to instantly know the right frame for inputs to get a desired response in frame advance. Say I want to look at MK's perfectly autocanceled fair followups. Now I don't have to guess at when to input the fair. I just input the jump at the 00 timer point for a SHNF, check my list and see that I should be inputing fair 16 frames later, at the 56 millisecond point on the timer.

The advantages in recording in this manner is that its a very easy format to derive values from and work with. It'll save us all a ton of time and give us a set of tools to work with with given characters in frame advance.

Format:
frame* : timer** : input/effect (framework)

*in relation to the noted input frame (Jump input maps always use the jump input as frame 0, for example)

Examples: MK Single Jump optimal input maps (I still have to find the optimal FF inputs )

MK (vBrawl) SH/FH input maps
in process of redoing because of a glitch in the old tester .pac

Bowser

00 : 00 : input jump
25 : 58 : optimal FF input (SH)
45 : 25 : SH lands
65 : 191 : FH lands

Jigglypuff

00 : 00 : input jump
07 : 88 : airborne
28 : 53 : optimal FF input (SH)
34 : 43 : optimal FF input (FH)
51 : 15 : SH lands
65 : 191 : FH lands

Lucas

00 : 00 : input jump
05 : 91 : airborne
21 : 65 : optimal FF input (SH)
33 : 48 : optimal FF input (FH)
38 : 36 : SH lands
61 : 198 : FH lands

AC: autocanceled on the earliest possible frame
LC: landed the frame immediately after the first hit frame of the noted hit

I left out options that were simply sub-optimal, like LC fair 3rd hit, which is simply inferior frame-wise to AC fair. FF options always assume earliest FF. Thus some options which are amazing with later FF were left out cause... I got lazy lol and finding all the SH/FHFF permutations is very time consuming.

Frame maps.

 

[Luxor]

Got BrawlWall/Brawlbox set up and working with all the necessary stage .pac's, so I don't think I even need to get the stage textures working anymore. Getting the numbers for all these stages should be super quick, most of the time will be because of making pretty pictures to show the numbers.

 

[KayLo!]

I really, really want to do a project on shield poking, but I don't think I'll have the time to run one without significant help.

Is anybody else interested? I know someone offered before.

 

[Yikarur]

I really, really want to do a project on shield poking, but I don't think I'll have the time to run one without significant help.

Is anybody else interested? I know someone offered before.

what do you want to research in terms of shield poking?

 

[saviorslegacy]

I really, really want to do a project on shield poking, but I don't think I'll have the time to run one without significant help.

Is anybody else interested? I know someone offered before.

Actually, things are to busy right now for me to take on a project such as tethers.
So I'm in on this. shield poking is something that has always interested me.

 

[KayLo!]

what do you want to research in terms of shield poking?

Mostly which characters/attacks/strings have the best poke potential, which characters are most vulnerable (ex: big characters who tend to have body parts sticking out of their shields vs. small characters with ****ty shields to begin with vs. tall characters whose legs are left unprotected), a more in-depth look at shield tilting, and visual aids for the masses.

I have no detailed ideas on how I want to structure the project yet, so any suggestions are much appreciated!

Originally I was going to do it on my own for Pikachu and Zelda, but I figured I might as well open it up for anybody who wants to contribute for other characters. Personally, however, I plan on focusing on those two.

Actually, things are to busy right now for me to take on a project such as tethers.
So I'm in on this. shield poking is something that has always interested me.

Sweet.

I'll write up a thread sometime soon. Might take a few days tho.

K Prime, you're helping too.

 

[saviorslegacy]

I'll start on Sheik and Ike.
The rest should just be divvied up between everyone.

 

[rPSIvysaur]

I'll start on Sheik and Ike.
The rest should just be divvied up between everyone.

What?

 

[Two-Ell]

Lol Savior are you referring to everyone pledging to Kaylo's project or everyone else here? Kaylo I'm interested in this(Peach and Link), but I'm spending this week studying things around here to make sure I'll be competent enough to contribute to yours and future projects. You can count me in after I finish though.

 

[saviorslegacy]

Everyone was referring actually to myself, Kaylo, Yikarur, apparently 2link now and to anyone else who would be volunteering after we finished the characters that we were interested in (and thus giving it a start).
I know that not literally everyone can participate in each and every project. There are just way to many and we all have schedules to follow.

But yes, I did not mean EVERYONE on the board. That would just be dumb. lol




Anyways, anyone have an idea on how to test shield pokes? All I can think of is to go into training, set the speed at 1/4 speed and use your toes to control the controller with the characters whose shield you are testing while you attack with the controller in your hands.

 

[Luxor]

Frame advance >>> 1/4 speed lol. Plus with frames you have access to values and formulas and whatnot for things like shield life, shield damage, etc. This could be a highly mathematical project if you choose to make it so, or you could stick with qualitative description. I would start with something nice, like seeing at what shield life (every shield lasts 361 frames) each character's untilted shield gets poked by MK's Dtilt. That gives you a starting reference for pokability anyway, so you don't have to sample the whole cast every time you need to check something- just the extreme points and a central one. Then I guess you'd get the critical shield life # for a bunch of other common shieldpokes to poke a shield on a few representative characters, and then you could use the shield damage formulas and whatnot to establish at what shield life % you MK can, for example, Nair-->Dtilt poke someone's shield, assuming no shield tilt. I'd ignore shield tilts completely for this project, actually.

tl;dr Use frame advance, see when in each character's shields lifetime (make sure you use Veril's timer conversion or you'll be stuck counting up to 361 frames) MK's Dtilt will shieldpoke. Get those 42 numbers and you've got usable, hard data to further manipulate.

tl;dr for tl;dr GET SOME NUMBERS

 

[KayLo!]

Ninja'd by Lux on the frame advance part. Never use 1/4 speed..... there's no use for it when you have debug mode.

Anyways, I like that approach, but I still plan on including tilting in my personal findings. This doesn't have to be a deadlined project, after all..... it can be open and last for as long as people feel like working on their respective characters once we have some basic data. Thanks for the input, though, Lux..... your suggestions will be a perfect starting point.

I also have a partial list of key moves I want to test in-depth once we get some basic numbers: MK tornado/uair, Peach dair, Marth fair/uair, Pika dsmash/uair, and Zelda's smashes are just a few.

Oh, and for people who're experienced with hacks: If you know any shield-related codes that might be useful for this, please let me know! I've looked around for some but haven't found anything that'd be particularly helpful. I'm notoriously bad with code stuff, though.

Ugh, I really need to make an organized thread for all this. x.x Tomorrow....... !

 

[Two-Ell]

Thanks Luxor that explanation is pretty understandable. I'll be watching out for your thread, Kay. I'm pretty interested in Peach's aerials as well.

 

[Luxor]

Kaylo, here's a useful shield-related code for you.

Spoiler

Colored shields [Phantom wings] (20 lines) trollface.jpg
4A000000 80F59A60
14000088 FF3AB700
14000090 FF000000
140000A0 FF000000
140000A8 00000000
4A000000 80F59ED4
14000088 00FFFF00
14000090 0000FF00
140000A0 0080FF00
140000A8 00000000
4A000000 80F5A348
14000088 FAF9E100
14000090 FFFF0000
140000A0 FFFF8000
140000A8 00000000
4A000000 80F5A7BC
14000088 88FCAE00
14000090 00FF0000
140000A0 00C00000
140000A8 00000000

I like your list as well, though I'd include MK/Marth Dtilt. In any case, establishing an order of "pokability" is key. That order might be different depending on whether you attack from the top or bottom, though- I can't think of an easy "MK Dtilt" style procedure for attacking from the top, since Nado is always hard to work with. I'll try to dig up Veril's formulas for shield damage, etc., too.