The Yoga Book Hyper questions/translation requests

Please note:



  3. The Yoga Book Hyper overlaps a lot of the info on my ST page. Make sure your question hasn’t been answered there first.

  4. Keep your translation requests reasonable. “Translate Ryu’s strategy section!!” is not a reasonable request.

Amazing work, NKI. Great information just in that translation!

Just to re-add what I had added on the other thread:

When studying the hitbox data, the coloring scheme for the boxes is as follows:

green > blue > red

Green will cancel green, and will beat blue and red

Red will “whiff” red, beat blue and lose to green

Blue will “whiff” blue, and lose to red and green

Another note: if you look at the normal move data, you will notice 3 boxes to the upper right of the picture of the move animation. If these boxes are highlighted, then it indicates that the move has the box(es)’ properties. The last box, easily enough to decipher, which has “SC” in it, represents super cancelability. From the best I can tell, the middle box represents special cancelability, and the first box represents the move is comboable into itself. NKI, can you confirm this?

More later…

I’m wondering, does the book mention anything about how they came up with the hitbox diagrams? I know the CPS1 SF2’s have a visible hitbox mode that can be activated with hardware dipswitches.

The CPS2 ones don’t have have one that I know of, but some of the scans of the hitbox diagrams in YBH that I’ve seen seem to actually have these in-game hitbox markers visible (they mark the corners of the boxes). Check out the Zangief and T. Hawk pics on the errata page for a good example.

If there is a way to view the hitboxes in-game for Super Turbo, I’d love to know how.

This book is the bible. :nunchuck:

All you need to know is the Condor Beak.

Yes, if I remember correctly. (It should say ??? or something similar in the legend.)

I don’t remember where I heard it (maybe from T-Akiba?), but I heard that Capcom actually gave them a dev kit to work with, so they could do all kinds of awesome stuff, like turn on hit boxes.

can you help me translate some stuff and show some diagram/hitboxes on blanka’s, backhop, forward hop and anti air ball?

Okay, some more video/animation basics, for better understanding of how the game works…and how it works within it’s environment. For the frame data in YBH to be of any use, you must have first have a good understanding of the following.

When we think of frames, there are actually two types of frames. There are video frames and there are animation frames. For the Street Fighter II series (and most arcade games in general), video frames occur at 60 frames per second, or one frame every ~.167 seconds. This is constant, and will never vary. Within the video frames, “animation” occurs, or the illusion of motion that is created by rapidly changing images. So, there are also animation frames, which exist within the constant video frames. When you look at the frame data for YBH, you see (for example) a move having 8 startup frame, 5 hitting frames, and 8 recovery frames. This is the number of total video frames involved in each aspect of the move. And by this, we can further the understanding by taking the total number of frames, which in this example is 21, and say the move lasts a total of 35/100 of a second, or roughly 1/3 of a second. Within those video frames, and the respective aspects of the move, you may have 5 animation frames being displayed in progression over the course of that 35/100 of a second. The first animation frame may last 4 frames, the second lasts 4 frames, the third lasts 5 frames, the fourth last four frames, and the fifth lasts 4 frames. It is these animation frames that create the “appearance” of the move, and these too have very specific properties, which will occur at each video frame of their duration and behave specifically for each progressive frame of the animation. Let’s take walking for an example of this. There are specific animation frames involved in a character walking forward or backwards. For each video frame, the character’s animation “moves” a certain number of pixels from one frame to the next. This creates the movement, and the number of pixels that the character “moves” from one frame to the next is absolutely and specifically defined. This, too, applies to every animation in the game, be it special moves, jumping, crouching, super moves, even background animation. So, to understand how these things can be substantially quantified and held to be unyieldingly true, we first must understand the basic premise that is the Street Fighter II game engine.

Oh God… I wish I could get one of those.:looney:

Is there any plan for anyone translating the black and white strategy section at the end of the book? I’m interested in what it has to say about Zangief. Other than the japanese text this book is amazing. (same goes for the DVDs)


OMG thank you so much for this! <3 I just checked everything out and omg. thank you dude!

I think the most useful thing regarding the strategy guide section at the back would be a translation of what the 8 questions are. The questions are the same for each character, and you can figure out a lot of the answers just by looking at the movelists, even if you don’t read japanese.

Rough pass (I cant read many kanji, so this is mostly guesswork)

  1. Opening moves
  2. super combo gauge, ie moves to fill it
  3. useful strings
  4. how to attack a dizzied opp
  5. anti-fireball
  6. reversals
  7. anti-air
  8. crossups

NKI, couple questions. I’ve started doing some of my own analysis, including documenting the animation frames involved in each move. Does the book denote which speed is being analyzed (I would assume turbo 3, but would like clarification)? Also, is there any information on the “slow down” effect, caused by meaty attacks or certain special moves? In my analysis, doing a shoryuken while this slow down occurs (during an opponent’s hitstun, resulting from a hadouken) produces extra frames. This makes perfect sense, as the “slow down” most likely doubles frames for a certain duration, or something along those lines. I’m sure it is quantifiable,
but am curious if YBH explores this at all.

I would like to quantify the duration of each animation frame, but for now, here is a teaser of what I’m working on. Not quite as graphically nice as YBH, but this is just a start.

‘Hitting attack’ refers to an attack that hits a crouching, non-blocking opponent at close range (obviously, the most opportune attack).
-Startup frames refers to the frames that occur before the attack can hit
-Hitting frames refers to the frames that can hit the opponent (red hit box present, if cross-referrencing YBH)
-Recover frames are those at the end of the attack which have no red hit box present, and the attack is completely vulnerable
-I broke these down within the total frame duration of the attack; i.e. for WW Ryu, the 1st-6th frames are startup, 7th-50th are hitting, and 51st-73rd frame are recovery, with a total duration of 73 video frames (or, 1.2167 seconds).

‘Non-hitting’ attack refers to an attack that does not hit (whiffed attack, practice attack, etc.).

‘Non-hitting with slow-down’ is the attack performed while slow-down is being produced (from a fireball hitting the opponent, in these instances).

WW Ryu jab shoryuken:

ST Ryu jab shoryuken:

One other thing I’ve noticed: the total number of video frames for a given move is not 100% constant. There appears to be some sort of “frame insertion” or variable animation (specifically defined, I would assume, but nonetheless present) occurring. I have witnessed this with ST Ryu’s jab shoryuken on both turbo 2 and turbo 3, as well as (and most importantly) Ken’s fierce shoryuken. Ryu’s “hitting” turbo 3 jab shoryuken is either 43 (as documented by me) or 44 (as documented in YBH) total video frames. I would be willing to dismiss this as some short coming in my caputre/analysis procedure, however, (and this is why it has proven to be most important in this theory) Ken’s fierce shoryuken, in rare instances, produces 1 startup animation frame. This attack is commonly held as, and documented in YBH as, being a first-frame hitting attack. Roughly 90% of the time it is, but roughly 10% of the time, there is a startup non-hitting frame. While my hardware could certainly be dropping a frame randomly, producing variable results, in no way would it insert a startup frame on a move that is documented in YBH (and again, commonly accepted) as having no startup frames. Seeing this consistently in other moves offers some validity to the existence of a “variable animation” purposely programmed into the game. Any mention of anything like this in YBH, NKI?

A friend of mine wanted to ask you NKI about p.31 for Blanka, it lists his crouching short and forward as both having the same startup, active and recovery time. Is this correct?

Also wondering if you have the link to the translated data for CPS1 characters reversing throws? It goes along the lines of

"if you use CPS1 Ryu and try to tap throw your opponent (CPS1) then the results are:

defender does nothing - Ryu throws
defender tries to reverse throw- Ryu throws
against a command grab like SPD - SPD always wins

it’s not 100% because there is a 50% chance Ryu will get a normal attack instead of a throw but the defending CPS1 character can never reverse the throw.

against a CPS2 character:

defender does nothing - Ryu throws
defender tries to reverse throw- defender wins
against a command grab like SPD - SPD always wins"

it’s not word perfect as that’s from memory and i read it a while back but it’s along those lines. i want to double-check the info, but can’t remember where it is.


I promise I’ll get back with you guys once I get home, at the end of the month.

r3k0, it appears they do have the same general properites, with forward having greater range.

So, does anybody know if YBH’s frame data is from ST arcade, or from ST versions in HSF2? I did some analysis of Blanka’s crouching short and crouching forward in HSF2, on Turbo 0 (Asia board, so turbo choices are 0-3), and came up with the following. Results are the same for both strengths, with the only difference being that forward has greater range.

Startup - 4
Hitting - 4
Recovery - 6
Stop frames - 14

Going to turbo 3 gives consistent results of 3-3-4-9 for short, but some anomalous results for forward. Results were 3-3-5-9 and two instances of 3-4-4-9. Very peculiar. Anyhoo, any confirmation of the version and/or platform from which YBH’s results were proffered?

I believe it was arcade ST, Jap version, Turbo 0, because that’s the way T.Akiba tested all his stuff, and he had a hand in The Yoga Book Hyper.

If you narrow that down a little bit, I’ll help you out, but I don’t have time to translate Zangief’s whole strategy section.

I believe you’re referring to hit-stop, which is explained on page 173. (Hit stop adds 14 frames.)

There is random frame dropping in ST, but the emulator you’re using and the fact that it’s not on Turbo 0 will probably make it more dramatic.

That used to be up on my ST page, but I took it down because after testing it myself, I came to a different conclusion.

I don’t really want to upload any images of the hit boxes (I’d rather people buy the book), but what exactly do you want translated?

help me check how many frames of startup and recovery are for the backhop/forward hop and anti air ball and how many invincible frames are there

Checking that stuff is easy. Basically, if you look at the bars below a move they’re color coded like this:

Gray bars: Non-hitting. Usually shown for startup and recovery, but also for moves that have gaps between hits(ex. ken’s hurricane kick).

Red bars: This indicates a hit. Each separated block of red is a new hit, even if they’re right next to each other(ex. sim’s headbutt)

Blue bars: The blue bars below the gray/red bars indicate invincibility.

So, those moves for Blanka go like this:

Back hop: 29 frames from start to end, and the first 13 are invincible.

Forward hop: 29 frames from start to end, with no invincibility.

Jab Upball: No startup. Hits for 30 frames, 3 frames of non-hitting, a second hit for 8 frames, then 17 frames of recovery.

Strong Upball: No startup. Hits for 29 frames, 3 frames of non-hitting, a second hit for 8 frames, then 17 frames of recovery.

Fierce Upball: No startup. Hits for 28 frames, 3 frames of non-hitting, a second hit for 8 frames, then 21 frames of recovery.

You don’t really need to know japanese to read the moves. Color-coding makes it easy :tup: