• Welcome to Smashboards, the world's largest Super Smash Brothers community! Over 250,000 Smash Bros. fans from around the world have come to discuss these great games in over 19 million posts!

    You are currently viewing our boards as a visitor. Click here to sign up right now and start on your path in the Smash community!

Project Database

#HBC | Dark Horse

Mach-Hommy x Murakami
Joined
Jun 12, 2010
Messages
3,739
I hope you don't.

Here's what I'm talking about: You know when somebody flops on the ground due to being hot by a move with locking properties? And you know how you can hit them again?

What about finding how many frames you have to hit them again?
 

rPSIvysaur

[ɑɹsaɪ]
Joined
Jun 7, 2009
Messages
16,415
You mean a Jab Lock. And yes I'm going to start the mechanics of missed techs soon.
 

#HBC | Dark Horse

Mach-Hommy x Murakami
Joined
Jun 12, 2010
Messages
3,739
Okay.

Wait, why is it just called jab lock? It's also possible with other moves, such as falco's lasers and TL's arrows.
 
Joined
May 3, 2009
Messages
7,190
jab lock more like weak hit lock mirite

Seems to be that a move has to have a very low BKB and growth rate to lock. Which would explain why jab1 tips lock while jab1 closecontacts don't.

...I think. I'm going to look into what is neccessary for a move to lock.

Also, I'm going to take a peek at various suspect aerials in PSA/OSA/w/e to compile EACris landing frame data. According to Veril you need to land before RAbit16 is set, so I'm going to work on that along with conservative testing to find out which aerials can feasibly be EAC'd consistently/easily.

Prolly going to make a new thread for that though since it's reasonably different from what's in my junk mechanics thread.
 

#HBC | Dark Horse

Mach-Hommy x Murakami
Joined
Jun 12, 2010
Messages
3,739
Just a suggestion, but what about a stage mechanic project?

(You try to discover patterns in the more interactive legal stages)
 

rPSIvysaur

[ɑɹsaɪ]
Joined
Jun 7, 2009
Messages
16,415
Any stage project would probably be about the floats found withing the .pacs
 

Luxor

Smash Champion
Joined
Jul 13, 2009
Messages
2,155
Location
Frame data threads o.0
A comprehensive stage boundary study would be nice. It's easy enough to replace the normal background with the custom "grid," right?
 

rPSIvysaur

[ɑɹsaɪ]
Joined
Jun 7, 2009
Messages
16,415
A comprehensive stage boundary study would be nice. It's easy enough to replace the normal background with the custom "grid," right?
That's pretty easy to do TBH. I can show you how to do it and you can set it up. (using BBox of course)
 

rPSIvysaur

[ɑɹsaɪ]
Joined
Jun 7, 2009
Messages
16,415
I don't even look at those threads... they're terrible. Perhaps the only one I actually find worth noting is the Brinstar one, but that's only for the lava levels.

If you're honestly going to do the 8-minute timer thing, I'd force you to do it >250 times. Seriously, that's the only way to get a reasonably semi-accurate figure with "see what happens".
 
Joined
May 3, 2009
Messages
7,190
rPSI, that's a little too much... just stick with the scientific method and run some six-nine trials.

And I confirm, I'm not particularly fond of those threads either, but mainly because 90% of the time their objective is to justify the stage as legal, instead of for the sake of presenting and gaining knowledge.
 

#HBC | Dark Horse

Mach-Hommy x Murakami
Joined
Jun 12, 2010
Messages
3,739
rPSI, that's a bit overkill. Probably somewhere around 5-7.

@IG

What if the data was purely for research?
 

DMG

Smash Legend
Joined
Feb 12, 2006
Messages
18,958
Location
Waco
Slippi.gg
DMG#931
Getting more info on projectiles would be nice. Like how far they travel per frame, how arcs in the path affect it, etc.
 

Luxor

Smash Champion
Joined
Jul 13, 2009
Messages
2,155
Location
Frame data threads o.0
Getting more info on projectiles would be nice. Like how far they travel per frame, how arcs in the path affect it, etc.
Doing this with, for example, Diddy's banana would be great. I know a lot of people would be grateful to no end if there was a photoshopped image of where Diddy's banana is frame-by-frame, along with a little bit of data saying, "at max range diddy has 16 frames to banana throw OoS to hit MK after shielding Dsmash" or whatever. Maybe glide toss numbers too; somebody should at least go make that picture, though. I think I heard you suggest it in the first place, actually >.>


small sample size is small
Luxor used Counter!
 
Joined
May 3, 2009
Messages
7,190
I've also been wondering about that for-you guessed it-Gyro. It'd be nice to get that, but there'd me just so many permutations for it that it'd be mentally scarring. Distance after being charged x amount of frames? Now after x+1 amount of frames? So tedious.

Also, I've been thinking about Snake's shield grenade grab, or whatever it's called. Does anybody know/looked into the exact mechanics of it? So far I've gleaned that 1) it can only be initiated on a droppable surface, and 2) it is some form of janky airdodge grab. What bother me about that last part is this, though: You can SGG from a plat to fall off and still catch the nade, with an airdodge animation, but you can't do the same thing off the lip of a stage.

Also, the inputs. I'm going to try and create a frame map for it, but put simply, it's hold B to pull a nade out, tilt/tap down, then immediately FLICK shield and release B. More or less. However, if you omit shield from the equation, nothing happens. If you omit down, you shield drop the nade. So I'm thinking that there's a small window in which "down tap" lingers, and when you flick shield, you buffer both a shield fall through and an airdodge, which, because it is initatiated so quickly, is cancelled by the platform. This is also evidenced when you do a "perfect" SGG, where the shield doesn't appear and only the pickup graphic appears nears Snake's head. Snake also seems to be fresh out of his soft drop animation.

Thoughts?
 

Luxor

Smash Champion
Joined
Jul 13, 2009
Messages
2,155
Location
Frame data threads o.0
Random speculation, but I think that "shield platform soft-tilt fallthrough" tech is involved. It would explain why you can't do it off the ledge, but shouldn't really make you airdodge. Anyway.
 
Joined
May 3, 2009
Messages
7,190
Well, I did include that near the end of my post...

But the input order is what confuses me. Normally soft-tilt shield drop is shield THEN tilt. In this case it's tilt THEN shield.
 

JOE!

Smash Hero
Joined
Oct 5, 2008
Messages
8,075
Location
Dedham, MA
anyone ever just try going through each character's (kill) moves on everyone else, then finding out realisticlaly how long they live? (with or without DI)
 

rPSIvysaur

[ɑɹsaɪ]
Joined
Jun 7, 2009
Messages
16,415
It's subjective. Too realistically find how long they live, we would have to find a tolerated level of acceptable DI. This varies from player to player. And we can't expect players to do Frame Perfect QCDI and what not. =/
 

JOE!

Smash Hero
Joined
Oct 5, 2008
Messages
8,075
Location
Dedham, MA
true, but perhaps we could simplify it a bit to "start"?

For example, we have say Mario and Bowser.

Fsmash kills Bowser at Y % naturally. Kills at Z% with "standard" DI (jump/hold toward stage/recover move?). And maybe even "technically kills" if at X % the character simply cannot recover the distance?
 

rPSIvysaur

[ɑɹsaɪ]
Joined
Jun 7, 2009
Messages
16,415
Is the aerial/double jump frame perfect? What frame should they occur?

BTW, I'm not shooting your idea down. I just want a rigid criterion in which this will be based on.
 

JOE!

Smash Hero
Joined
Oct 5, 2008
Messages
8,075
Location
Dedham, MA
problem with that is that it's all IASA based on the move and character....

so i dont think there'd be a possible way to measure the exact frames each move ends the hitstun/etc on...

however, simply buffering in an aerial/airdodge/DI/recovery move/potato should get in the ballpark of that exact frame, no?
 

rPSIvysaur

[ɑɹsaɪ]
Joined
Jun 7, 2009
Messages
16,415
The exact frame data in which you're in hitstun can be very hard to figure out. If you can get some very rigid criteria to base this project upon, I will let it pass a Smash Lab project, you will have to give it A LOT of direction as it has a lot of variables to it.
 

Luxor

Smash Champion
Joined
Jul 13, 2009
Messages
2,155
Location
Frame data threads o.0
The question is, do you WANT that exact frame? Are you measuring M2K's DI or the average JOE!'s (pardon the pun)
Because the average human can't do things frame perfect, so what can he do short of that but more than nothing? You can't buffer during hitstun.
 

rPSIvysaur

[ɑɹsaɪ]
Joined
Jun 7, 2009
Messages
16,415
The question is, do you WANT that exact frame? Are you measuring M2K's DI or the average JOE!'s (pardon the pun)
Because the average human can't do things frame perfect, so what can he do short of that but more than nothing? You can't buffer during hitstun.
This too.
<ten chars.

Edit: Moving this convo. into the project thread
 

JOE!

Smash Hero
Joined
Oct 5, 2008
Messages
8,075
Location
Dedham, MA
what I meant by buffer is essentially "as soon as possible", as frame-perfect is never gonna happen.

What I propose is just a method of a few tests per move...per character:

1) Mario Fsmash kills Bowser at X% (Natural, no DI, blast-zone KO)

2) Mario Fsmash kills Bowser at Y% ("Testable DI", meaning you DI at a good angle to recover back to the stage, use say Fair/Dair depending on trajectory)

3) Mario's Fsmash "Technically" kills bowser at Z% (Technically as in he DI'ed past the blast zone bubble, but simply could not get back to the stage due to the distance. Wont be applicable for some characters)
As for actually doing the testing, I think each move's X/Y/Z % should be tested say, 3-5 times to get an idea if there's any variation to it/room for error...
 

rPSIvysaur

[ɑɹsaɪ]
Joined
Jun 7, 2009
Messages
16,415
I don't really think the Technically should be used TBH. That could be done freelance.

And when you mean as soon as possible, you mean?
 

JOE!

Smash Hero
Joined
Oct 5, 2008
Messages
8,075
Location
Dedham, MA
well, weight and recovery both tie into overall survival, doesnt it?

but yeah, that can be taken out (for now :p)

and by as soon as possible I mean as early as the move allows, seeing as we have no idea the exact frames each character will end hitstun of each move they get hit by....

It is not very exact, but the other solutions is A) A nightmare to crunch all the seperate numbers and B) not really possible for humans to do.
 

rPSIvysaur

[ɑɹsaɪ]
Joined
Jun 7, 2009
Messages
16,415
well, how do we know everyone will be using the earliest possible frame? Is it not possible that they may be off a few frames?
 

JOE!

Smash Hero
Joined
Oct 5, 2008
Messages
8,075
Location
Dedham, MA
Thats the issue, how are we gonna test every single move's frame data against every single character's for something that is pretty much impossible to practically do (frame perfect) in the first place?
 

JOE!

Smash Hero
Joined
Oct 5, 2008
Messages
8,075
Location
Dedham, MA
Ive been saying this the whole time :p

But I've also offered a method of testing (multiple tests per move to get an "average") to make up for the lack of perfect prescision, as prescision will not happen with this sort of thing in a match...
 

rPSIvysaur

[ɑɹsaɪ]
Joined
Jun 7, 2009
Messages
16,415
I don't even look at those threads... they're terrible. Perhaps the only one I actually find worth noting is the Brinstar one, but that's only for the lava levels.

If you're honestly going to do the 8-minute timer thing, I'd force you to do it >250 times. Seriously, that's the only way to get a reasonably semi-accurate figure with "see what happens".
rPSI, that's a bit overkill. Probably somewhere around 5-7.

@IG

What if the data was purely for research?

You must have no dignity. Releasing that kind of data would be worse than killing yourself.

You are bad and should feel bad for suggesting that.
How I feel about averages. :p
 
Top Bottom