Editor, Experienced Forum User, Published Author, Experienced player
(818)
Joined: 5/2/2015
Posts: 671
Location: France
Ready Steady Yeti wrote:
I want to one day make a TAS of me unlocking everything in the game possible.
On a different subject, it is impossible to 100% this game as of now. Why? You need passwords to unlock the last parts, and only one has been released. To get 100%, you'd have to actively dig around in the game memory, find the passwords (if they are not randomly generated based on clock time or region locked lol) and enter them. This alone is a task that I'm not sure anyone can do (and even less sure anyone is interested in), unless they really like digging in memory.
Edit: Did I mention to 100% the game, you need to finish Story Mode 35 TIMES!!!!!, do all 100 missions, do Survival Mode (which is battle the 34 other chracters without losing), and get the password and product codes all to get all the 350ish Beyblade parts and 100%?
.. yeah, see you in seven hours.
Editor, Experienced Forum User, Published Author, Experienced player
(818)
Joined: 5/2/2015
Posts: 671
Location: France
Ready Steady Yeti wrote:
Yes, please please do submit this. Idk about you but personally I love this TAS and will definitely give it a Yes vote. I want to one day make a TAS of me unlocking everything in the game possible. Maybe you could help me with that also. Please please submit it. You have no idea how much I want it officially submitted! :D
Editor, Experienced Forum User, Published Author, Experienced player
(818)
Joined: 5/2/2015
Posts: 671
Location: France
Link to videoFull game "Arcade, Gingka" (click this link)
Start from clear sram (if you don't know how, just make a savestate right at the start of the movie.) Encode coming when my internet stops dropping.
Don't know if I should submit it here, as it's pretty boring.
Editor, Experienced Forum User, Published Author, Experienced player
(818)
Joined: 5/2/2015
Posts: 671
Location: France
Link to videoMovie file
First 2 fights.
Notes:
- This is the Gingka path, so we have to take Gingka as lead.
- We take Random and wait a few frames because it is faster to Random and do RNG manipulation than just select the characters. This saves 3 frames. We must have the chracters in right order; Gingka at the lead to get his story, Masamune because he has more knockback.
- LET IT RIP
- For the first fight, I wait a bit before doing a cyclone dash to make sure he doens't block. I then do another one, clipping him out of bounds.
- LET IT RIP
- The second fight starts off with a Savage Noise, then we do what is called a dash cancel. By doing our special (Savage Noise) at the same time we finish a flame dash, we can immediatly cast Savage Noise. This makes Ryutaro get stuck inside a wall and get both knocked back by the Savage Noise AND the flame dash, clipping him out of the field and much further than you would expect.
- Because you only need to win two fights, we will not use Tsubasa for the entierety of the run (he's ugly anyways boo)
Editor, Experienced Forum User, Published Author, Experienced player
(818)
Joined: 5/2/2015
Posts: 671
Location: France
Ready Steady Yeti wrote:
So I was just wondering, should I even waste my time playing these games to test them out for TASing purposes
Alright, I'm going to tell you something for out there. Unless this game is really bad (which it doens't seem to be), the game will be interesting to watch if you put your heart in it.
For breaking the game, findings tricks, let your will guide you. You can go at whatever pace you want; you have no time limits. The entire community will be waiting here for your progress.
If you want to talk about TASing, and just general help (even more specific help; you may find people interested for the current game you're TASing) don't be scared to check out the TASVideos IRC Channel (at the top of the forum page.)
Editor, Experienced Forum User, Published Author, Experienced player
(818)
Joined: 5/2/2015
Posts: 671
Location: France
Ready Steady Yeti wrote:
my emulator was so slow I didn't get very far at all.
Assuming you're using DeSmuME, do you have Advanced Bus-Level Timing/Dynamic Recompiler on? (Config > Emulation Settings)? Did you try setting frame skips to 1, 2 or 3+ if you are running it very slowly? These are the main factors that influence framerate.
Editor, Experienced Forum User, Published Author, Experienced player
(818)
Joined: 5/2/2015
Posts: 671
Location: France
z1mb0bw4y wrote:
You're well aware that the only difference between your definition of "fooling the OCR" and what they're planning to do is the location of the extra pixels they're drawing, right? You can literally do the same thing by placing extra lines or pixels within the lines you've already drawn, from my understanding (though admittedly, that process is slightly more complicated and requires more effort). This is just a more convenient way to do it consistently...
The only reason OOB differs from edgebreaking manipulation is that there's two extra edges to account for. That's it. The fact that we are nullyfying the drawing by out of bounds makes doing drawings without out of bounds actually slightly easier, as you can just position the two edges on the top left and bottom right of the drawing, then manipulate with godspots to get your right number. The reason it is easier is because since we aren't doing out ot bounds, the keypoints we are making by drawing register. As such, as long as you have 3+ keypoints (complicated drawings would use around.. 40 to 200?), it is possible to manipulate almonst anything you want, and it becomes easier the bigger the drawing is.
The challenge here isn't manipulating big drawings and making them fool the OCR by some sort of wizardry: it's fooling very small drawings that give little in the way of manipulation. We have very big control over the drawings by just keeping track of the two edges of that drawing.
Editor, Experienced Forum User, Published Author, Experienced player
(818)
Joined: 5/2/2015
Posts: 671
Location: France
dwangoAC wrote:
I assume the consequence is the artwork will be less impressive to look at but will be more impressive technically.
Not true: almonst any drawing, unless you're on a 63/42/81 and it has less than three keypoints or keypoints close togheter in a bundle spaced out from the next bundle of keypoints (which will never happen, unless in the . . . screens) can be made into anything from my experience.
Editor, Experienced Forum User, Published Author, Experienced player
(818)
Joined: 5/2/2015
Posts: 671
Location: France
micro500 wrote:
However, I do find the ability to draw nothing (visually) on screen and still answer the question entertaining, so I am on board with answering one question like that. It would show off the neat glitch we found without deceiving the audience too much. We can tell the audience how that worked without ruining their expectations for the rest of the run. This does mean we need to find a way to make our drawings work for all answers, but that is doable, and it would be an even more impressive thing to tell anyone who asks about the technical details.
Doing a blank answer doens't actually need to use out of bounds, since you will not be drawing anything: if anything, it should help the audience get the gist of what's happening. The game is very complex in its inner workings and it would be hard to explain everything in a 8 minute segment.
Editor, Experienced Forum User, Published Author, Experienced player
(818)
Joined: 5/2/2015
Posts: 671
Location: France
To describe the tricks in more details, it is recommended to have a basic understanding on how recognisition of numbers (aka recon) works in Brain Age. If you haven't, check out the first and second Brain Age post by me in the DS forums.
The first trick is called edge breaking. When you draw something in a single input (no matter how long it is), edges are placed at the start and end of your line. Edges are also used to manipulate models. Manipulating anything with 6-8 edges is almonst impossible. This would mean that we would have to do most drawings in a single stylus input without stoppi g input.
However, by going on the edges of the drawable field for 1 frame then going back in wherever we want, we can manage to stop the line we are currently drawing and start it somewhere else without generating an edge. This means all the drawings in the run will have 2 edges (thank god) or 4 if for a double digit answer. This is seen as okay with everyone since it is recon manipulation.
The other thing with edge breaking is that when you are out of bounds with edge breaking, you can go back in bounds for 1 frame to make one input invisible. Note that while out of bounds, you still generate keypoints (they are just pushed back to the border of the drawable field.)
The second glitch is called godspots. By placing inputs on a very specific position (x:246) we can make the game not show up the pixel since it is too high and is blocked by the top of the screen, but will still count as a keypoint/edge. Using this, we can make completely invisible models. Note that drawings still generate their own keypoints/edges, so we have to carefully make completely in visible edges or keypoints (and sometimes redraw over alderaly deawn portions of the drawing field if we need to be back in bounds for more than one frame), so manipulating with godspots is extremely tough.
The glitch everyone argues on is out of bounds drawing. By making our first frame draw out of bounds, then going back in bounds, our entire drawing will not be recon and will not have keypoints or edges. This breaks the entire game and makes the drawing and invisible models separate, what most people are against.
Editor, Experienced Forum User, Published Author, Experienced player
(818)
Joined: 5/2/2015
Posts: 671
Location: France
Edge breaking has never been so broken. Inspired by micro500's results, I tried somethings, and it looks like you can keep up edge breaking to set up multiple keypoints and an edge. Yes, you read that right. The other big breakthrough is that if you edge breaked on a non-drawable part of the playing field, you can then continue egdge breaking and effectively go out of bounds, but if you edge break on a out of bound part, you have the next frame to either continue the out of bounds, or edge break back in for 1 frame, drawing an invisible keypoint (or edge if it is your last input by edge breaking) anywhere on the drawable field, then go out of bounds again (you can then edge break in the playing field again. This means, if we generate multiple keypoints by edge breaking, only the last will be an edge (following normal rules of edge generation: start when you first input, end when you stop input for one frame)
This leads to the fact that you can generate models while ENTIERELY EDGE BREAKING. In short. because we go out of bounds with edge breaking, and are able to sgo in bounds again then continue the out of bounds, we can draw invisible models!! We can essentialy do every answer in the game with one pixel (or two if it is a mutli-digit answer) with only invisible models, not having to draw over drawings to manipulate their outcome.
This leads to the most hype 2015 meta speedrunning strats in history.
Link to videoMovie file if you somehow get it to sync
I'll do a more thorough explaination on this tommorow, but you can see the gist of what's happening by watching the movie yourself.
At this point, what do we need to have complete control of the game? Right now, we can manipulate a double edge (quadruple edge if double digits) drawing very easily. A bot could be made right now to manipulate any double and quadruple edge drawing. Our only problem now is what to do with drawings with more edges?
We have two solutions:
- either limit every drawing to be drawn in one line (exciting)
- find a way to manipulate numbers easily with more than 2 (4 on multi digit) edges
But at the point we're at, something can be quickly done in the way of botting, and as such should be left to micro500 (and other people who are willing to help, just chill on the tasvideos irc!) and try to find other tools that could possibly improve manipulation.
Editor, Experienced Forum User, Published Author, Experienced player
(818)
Joined: 5/2/2015
Posts: 671
Location: France
xy2_ wrote:
This should take about the same time as Fluctuant Travelator to make; expect it relatively soon.
This was a terrible thing to say. Swathe is infamous for being a copy of B4 in Super Monkey Ball, except all the edges are bigger, which leads to completely different and extremely tough strats. I haven't figured out a good way to tie the human record atm thanks to stupid cycles.. so don't expect it out soon.
This is the only level in the way of World 1 (along with Spiral Staircase, which is really hard). Hoping to finish this in two or three days.
Editor, Experienced Forum User, Published Author, Experienced player
(818)
Joined: 5/2/2015
Posts: 671
Location: France
Link to video
(If you wonder what my browser is, it's Maxthon: what all the cool kids use after Opera became bad)
So something I found but didn't document here is that models are also recon by what we call 'edges'. The best way to explain this, is to look at the 1 model. See where it starts and where it ends? That's the two edges of the 1 model. We need these edges in proper position for the 1 model, and at least two keypoints that can be assigned to this model. Edges are placed basically where your line starts and ends, and it's the reason stupid things get recon.
Now, remember how I said model detection was also based on lines? Using line-breaking (a long known trick,) we can manipulate where edges are placed. After that, we have a 1 frame window to place the edge wherever we want on the touchscreen.
What does this do? It manipulates the number recon without even drawing a single pixel. Yes, you read that right. This is what I call 'breaking edge', since it both breaks a line and manipulate edge placement.
Why does this happen? I have no clear idea, but a speculation:
Since we draw offscreen and never actually draw any more pixels beside that one pixel, the game confines the model of whatever number we're doing to that pixel only (remember, model dimensions are modulable down to a single pixel!). Following that, that edge gets placed, but not drawn: the game still continues, thinking we made a model succesfully (two edges, enough keypoints - as long as it's drawn, the game recons it as a keypoint, and it takes keypoints off that single pixel - and in the right arrangement) and the breaked edge stays at its location, only mattering for its angle at which it is placed. The game stills recons it, and assings it to a number based on its location and which model it makes, but still considers it to be in the pixel we just made (since edges can't logically be on non-drawn locations.) Think of it like a rubberband :)
In this picture, the green dots are both edges, while the orange dots are keypoints. The bright green dot reprensents where the game thinks the second edge is, while the blue dot represents where it actually is. Note that from the second keypoint to where the supposed edge is, the same angle is kept.
This is the most logical explaination I can provide if we follow my speculations of models, keypoints and edges; though the game could work completely differently.
Editor, Experienced Forum User, Published Author, Experienced player
(818)
Joined: 5/2/2015
Posts: 671
Location: France
https://youtu.be/KTb4FLrEgfU?t=5m19s
Swathe is up next, and oh boy. Collision galore. (Go to 5:19 in the video for the WR, by Silver)
This should take about the same time as Fluctuant Travelator to make; expect it relatively soon.
Editor, Experienced Forum User, Published Author, Experienced player
(818)
Joined: 5/2/2015
Posts: 671
Location: France
Link to videoMovie file
Rerecords: 1351
Shoutouts to cycles for making this level a pain.
Note that I take a wierd angle to hit the goal at the end, bouncing off the left goal pillar; it isn't faster than going straight through the goal (I tied), so I just did this for extra style points.
