The doc got featured on Eventhubs:
Front page already in Eventhubs? That was a bit too early. I’m yet to go to check Ken and find out why his 4f jab that is +3 on hit can link into each other.
I updated most of lei’s data, but the thing is that I cannot check the start-ups of some moves (such as follow-ups in Razor Rush), so I left them blank.
EDIT: Moved on to Alisa
- cr.LP is +4 on block (cr.LP -> cr.LP is a true blockstring)
- far/close st.LP is +3 on block
- far st.MP is -2 on block
- far/close st.LK is -1 on block
- cr.LP (Destructive) is -2 on block
If that’s the only stuff from Alisa, I’m glad. Will test that now again.
Lars is finished btw, and ready to be proof tested by you. 
Maybe you can rapid fire his lp?
Well Lei can wait till we have finished the other characters. I fear him along with Christie the most.
Edit:
Confirmed and changed s.lp and c.lp. Also her Chainsaw c.lp is -2 (-3 on Paul). only tested chainsaw s.lp further, which was correct.
Oh haha. You changed destructive cr.LP before I added it in my post 
It would be a work in progress. Alisa is taking a bit longer than I thought in testing stuff, especially the chainsaws. Luckily all the normals are confirmed save the ones I pointed out in my post above.
You CAN rapid fire Ken’s cr.LP, but he cannot cancel into special in that case. I was specifically able to do cr.LP -> cr.LP xx SRK cancel.
Confirmed lk and mp.
I recorded ken’s LP (far. cl. c.) on hit and block, and the data seems correct. So that might be another puzzle to solve. Maybe you can chain/cancel them in the last recovery frame and they become cancelable that way (like we had with knife c.hp)?
That certainly is a possibility. I’m confirming Ken’s data first thing after DLC characters are done.
EDIT: Done testing Alisa. Everything else seems to be fine.
Lars
- close LP is 6 on hit, 2 on block
- Avalanche Stomp 28f start-up
(Avalanche Stomp actually becomes active before you can hit a standing dummy, so if you want to check the fastest start-up you have to use it against jumping dummy. I believe this would change frame advantage against tall characters such as Hugo) - Zeus has 14f start-up
I think this is it for Lars. Testing got much easier as your dummy changed to someone else from Paul 
Found the culprit causing the Paul blockstun issue. It was my 2nd guess, they messed up paul’s guard animation timing. While they moved “Damage” scripts into a common file, “Guard” scripts are still stored inside each individual characters .bac / .bcm files.
Here is what the script header for all other characters (at least the half dozen I checked) looks like (this example is taken from Ken)
<Script Name=“GUARD_ML” Offset=“0x83f4L” HitboxStart="-0x1" HitboxEnd=“0xf” IASA=“0x10” End=“0x11” Unk4=“0x60” XOffset=“0.0” ScriptFlags=“0x0” EndsOn=“0x0” Loop_=“0x0”>
This is what the script header for Paul’s Guard_ML
<Script Name=“GUARD_ML” Offset=“0x85d0L” HitboxStart="-0x1" HitboxEnd=“0x11” IASA="-0x1" End=“0x11” Unk4=“0x60” XOffset=“0.0” ScriptFlags=“0x0” EndsOn=“0x0” Loop_=“0x0”>
If you notice they are basically identical other than the offset (not important AFAIK) and the “IASA” value which judging by what I know from SF4 modding IASA stands for “Interrupt As Soon As” (in the SF4 mod tool it is called “Interrupt at Frame”) and the “hitbox end” value is different. I’m not sure what the Hitbox End does as there is no actual hitbox for the move and we’ve yet to figure out what that value does in SF4. However the interrupt as soon as value being -1 instead of 10 likely is the reason.
Likewise the only applies to Pauls Guard_ML, Guard_MM, not any of his other guard scripts. That is why for instance, Cody’s close s.LP and close s.MP have 1F less blockstun but his close s.HP has the same blockstun on paul as everyone else.
That is awesome info! Thanks for clearing that up. I assumed it had something to do with moves that hit a specific area.
I’ll forward this to the post over on unity.
That means, that was in Vanilla also the case.
Edit:
Would you be able to check out why Cody’s Knife c.hp links into his c.hk (and other 7F moves)? It only yields 6F hit adv. The post above also has a different setup (c.lp, c.lp xx SRK) done by Ken, which should not be possible since c.lp has 4F start up but only has 3F hit adv. If you skim through the posts, we have a theory, but have not confirmed it yet.
Ken’s c.LP is chain cancelable. If you look at Ryu in SF4 you’ll see that his c.LK is only +2F on hit and has a 4F start up but can combo into itself.
c.HP with knife has more stun when the move hits than when it is blocked. The hitstun for on hit is listed as 0x14 while for block it is 0x12 (not sure if this is a 1 to 1 ratio of frames but its IS different.)
I meant hit adv.^^ it yields 6F hit adv but you can combo into 7F c.hk.
And yeah, it’s rapid cancelable, but then you couldn’t cancel the 2nd c.lp by srk.
Thanks to Eternal’s charge time data, I was successfully able to add Cody’s upper body invincible frames in google doc. I posted Lars changes above, so if you could test those it would be great.
Also, I was wondering if you could test Cody’s Knife cr.HP using recording & frame counting:
- Knife cr.HP -> jump
- Knife cr.HP -> any one of Knife cr.MP/cr.MK/cr.HK (it must combo)
Recording these two, then comparing the frames of Knife cr.HP would let us verify the theory.
I wouldn’t be able to confirm anything before Monday, so b4 I have to go I’ll delve into Sakura’s frame data… a bit.
<blockquote class=“Quote”>
<div class=“QuoteAuthor”><a href="/profile/16564/street11">street11</a> said:</div>
<div class=“QuoteText”>That certainly is a possibility. I’m confirming Ken’s data first thing after DLC characters are done.<br>
<br>
EDIT: Done testing Alisa. Everything else seems to be fine.<br>
<br>
Lars<br>
- close LP is 6 on hit, 2 on block<br>
- Avalanche Stomp 28f start-up<br>
(Avalanche Stomp actually becomes active before you can hit a standing dummy, so if you want to check the fastest start-up you have to use it against jumping dummy. I believe this would change frame advantage against tall characters such as Hugo)<br> - Zeus has 14f start-up<br>
<br>
I think this is it for Lars. Testing got much easier as your dummy changed to someone else from Paul <img src="/plugins/Emotify/design/images/4.gif" alt=":D" title=":D" class=“bbcode_smiley” height="" width=""></div>
</blockquote>
Tested cl.lp again and I can only verify 5/1 again (vs Cody and Jin). Don’t tell me there is another glitch… <br><br>I tested Avalanche Stomp against air borne for the start up. (for block\hit adv agaisnt standing obviously) and done it again, his leg goes right through the opponent on frame 28 and 29 and doesn’t hit until frame 30. Hits standing Cody on frame 31 and crouching Cody on frame 32. I’ll add this info to the doc tho.<br><br>Mhh raw I got 13F start up yesterday, via reversal 14F this time around for Zeus. or maybe I made an error yesterday. Anyway, changed it.<br><br>Btw, are you able to cancel HP Dynamic Entry into Silent Entry? Haven’t been able to pull it off yet. :3<br><br>Leave my Paul alone :< /sadpanda I’m using Cody now as dummy.<br><blockquote class=“Quote”>
<div class=“QuoteAuthor”><a href="/profile/16564/street11">street11</a> said:</div>
<div class=“QuoteText”>Thanks to Eternal’s charge time data, I was successfully able to add Cody’s upper body invincible frames in google doc. I posted Lars changes above, so if you could test those it would be great.<br>
<br>
Also, I was wondering if you could test Cody’s Knife cr.HP using recording & frame counting:<br>
- Knife cr.HP -> jump<br>
- Knife cr.HP -> any one of Knife cr.MP/cr.MK/cr.HK (it must combo)<br>
<br>
Recording these two, then comparing the frames of Knife cr.HP would let us verify the theory.<br>
<br>
I wouldn’t be able to confirm anything before Monday, so b4 I have to go I’ll delve into Sakura’s frame data… a bit.</div>
</blockquote>
Good stuff. Can you find out which of the reversals for the characters have invulnerabillity as well? Mainly for the SRK like moves. Doesn’t need to be frame specific, just if it has any.<br><br>I actually recorded both, but forgot that I did and deleted the video. gonna record it again. brb<br><br><br>
@zUkUu<br><br>Actually, I didn’t have much time and tested cl.st.LP’s frame advantage using link method instead of hit-jump. I was able to link cl.st.LP into cl.st.MP, which has 6f start-up. I confirmed the link on Lars and Ryu. This link was possible in Vanilla as well.<br><br>Hmm. It’s certainly odd for Avalanche Stomp. I whiffed a lot against empty jumps due to the jump’s lower body invincibility, but I was able to hit the dummy at an earlier frame when the dummy did whiff normal jump.<br><br>About HP Dynamic Entry, it wasn’t cancellable into Silent Entry in vanilla, and in 2013, it is HK Silent Entry that is able to cancel into another Silent Entry, not Dynamic Entry. So no, you cannot cancel into Silent Entry from Dynamic. You can mash HK Silent Entry tho to cancel into another one, then again, and again, and again.<br><br><br>I’ll check the invincibility on Monday (I only have pre-2013 version b4 then)<br>
<br><blockquote class=“Quote”>
<div class=“QuoteAuthor”><a href="/profile/16564/street11">street11</a> said:</div>
<div class=“QuoteText”>@zUkUu<br><br>Actually, I didn’t have much time and tested cl.st.LP’s frame advantage using link method instead of hit-jump. I was able to link cl.st.LP into cl.st.MP, which has 6f start-up. I confirmed the link on Lars and Ryu. This link was possible in Vanilla as well.<br><br>Hmm. It’s certainly odd for Avalanche Stomp. I whiffed a lot against empty jumps due to the jump’s lower body invincibility, but I was able to hit the dummy at an earlier frame when the dummy did whiff normal jump.<br><br>About HP Dynamic Entry, it wasn’t cancellable into Silent Entry in vanilla, and in 2013, it is HK Silent Entry that is able to cancel into another Silent Entry, not Dynamic Entry. So no, you cannot cancel into Silent Entry from Dynamic. You can mash HK Silent Entry tho to cancel into another one, then again, and again, and again.<br><br><br>I’ll check the invincibility on Monday (I only have pre-2013 version b4 then)<br></div>
</blockquote>
Lol then I have that for the wrong move xD gotta fix it.<br><br>I was able to do cl.lp, c.mp xx silent entry as well. might be another case of this link-cancel (is that an appropriate name?) <br><br>I’ll record both right away.<br><br>
Thanks for confirmation. You checked it twice, so it must be +5 on hit, and you were able to verify that cl.st.MP was linkable after.<br>I so hate this. It screwed me up with Lei frame data in the first place, and if
it wasn’t for how Lei’s stance normal worked, I wouldn’t have been able
to realize it in the first place. I wonder what is the problem to begin with.<br>
<blockquote class=“Quote”>
<div class=“QuoteAuthor”><a href="/profile/16564/street11">street11</a> said:</div>
<div class=“QuoteText”>Thanks for confirmation. You checked it twice, so it must be +5 on hit, and you were able to verify that cl.st.MP was linkable after.<br>I so hate this. It screwed me up with Lei frame data in the first place, and if
it wasn’t for how Lei’s stance normal worked, I wouldn’t have been able
to realize it in the first place. I wonder what is the problem to begin with.<br></div>
</blockquote>
Okay I have both right next to each other (I swear Premiere is a mess, I want my Sony Vegas back :<).<br>I can’t make out a difference. The only difference I can see is some hit effects, but the hit frames off the dummies look EXACTLY identical. first jump frame and first c.hk start up frame are on the same frame, meaning link-canceling is <b>NOT </b>the case. On the 7th frame of hit advantage, cody has his first jumping frame where as when comboing into c.hk cody gets hit.<br><br>Just for you, I forward framewise.<br>http://youtu.be/XqBt15u4C58<br><br>I might do a full video about it (showing that c.hk has 7F start up, knife yields +6 etc)<br><br><br>
TYVM for testing.<br><br>I saw something.<br><br>In between 0:46 and 0:47, there is a frame where only Cody on the left side turns his body by just a bit, while Cody on the right side stays idle. This is the frame before the right side Cody goes into pre-jump frames. Do you think this is frame 1 of his sweep?<br>I’m saying this because I remember that back in SSF4 hitbox videos, the dummy would flinch at the 5th frame against a move that starts up in 4f. Therefore, if the engine is similar enough, the Cody in blue/white suit would flinch at the 8th frame against cr.HK. If you do not include the “slight turn” frame, Cody flinches at exactly the 7th frame. Of course, there’s no guarantee that it works the same way in this game.<br>
<blockquote class=“Quote”>
<div class=“QuoteAuthor”><a href="/profile/16564/street11">street11</a> said:</div>
<div class=“QuoteText”>TYVM for testing.<br><br>I saw something.<br><br>In between 0:46 and 0:47, there is a frame where only Cody on the left side turns his body by just a bit, while Cody on the right side stays idle. This is the frame before the right side Cody goes into pre-jump frames. Do you think this is frame 1 of his sweep?<br>I’m saying this because I remember that back in SSF4 hitbox videos, the dummy would flinch at the 5th frame against a move that starts up in 4f. Therefore, if the engine is similar enough, the Cody in blue/white suit would flinch at the 8th frame against cr.HK. If you do not include the “slight turn” frame, Cody flinches at exactly the 7th frame. Of course, there’s no guarantee that it works the same way in this game.<br></div>
</blockquote>
Holy shit batman, you’re right, he DOES flinch. But that would make it 8F start up if that was really his 1st frame. You have sometimes an additional frame of animation that appears before the input is actually relevant to the game. Let’s say you input “between frames”, then you have already a frame of animation of the move, but the input appears the frame afterwards on the screen, but <b>THEN </b>starts the actual start up, regardless of the additional frame of animation. That’s why I’m sometimes confused with reversal timing specials when watching them via recording, since the input is actually way earlier (buffered), I still often have one additional frame of animation that does not belong to the start up. For EX move that’s a frame where he doesn’t glow for instance, or for super/cross arts the cinematic freeze occurs during that frame. E.G Raze is -9, that means the additional frame of animation occurs on the last “block stun” frame (10F before Paul recovers).<br><br>I’m quite clueless, I’ll do the same for Lars later. Maybe @Eternal has an idea.<br>
Well, I’ll give 2 topics because I’m not sure which one you are talking about out of the two things I was mentioning:<br>1. Dummy flinching; in SSF4 hitbox videos against a move with 4f start-up the hit animation (the dummy flinching) would occur at the 5th frame (1f right after the move becomes active), so if SFxT works the same way, Cody with blue/white stripe should get hit animation at 8th frame since cr.HK has 7f start-up, and the “turning” part of the bumblebee Cody in fact becomes the 1st frame of start-up if you count it frame by frame.<br><br>2. Now about the turning part. I am not sure on this one, but we have to take into consideration that Cody on the right side wasn’t doing anything when he could’ve gone into pre-jump frames, but the one with cr.HK flinched. I’m talking about this difference. A sure-fire way to know whether this is the 1st frame or not would be
recording cr.HK on dummy record to “hold back” so that we would know
which part of frame causes blocking motion. If the “flinching” part causes block motion, it means this is the 1st frame, which confirms the theory that the last frame of recovery is cancellable, if it doesn’t, it means this doesn’t have anything to do with cr.HK, which means things get more complicated.<br><br>Since you did such hard work to test this, I’m not going to make any major requests, but if you WANT to do this, then feel free to:<br>- Doing the same thing with cr.HP -> link, but this time with Knife cr.HP -> cr.MK and Knife cr.HP -> Knife cr.MP (both of them combos as well), then comparing the “turning” part of the bumblebee Cody with the Knife cr.HP -> cr.HK. If the motions are all different in three of them at the “turning frame”, it proves our theory. If not, it most likely disapproves.<br><br>I’ll find a recording software in PC and do this if I can.<br>
<blockquote class=“Quote”>
<div class=“QuoteAuthor”><a href="/profile/16564/street11">street11</a> said:</div>
<div class=“QuoteText”>Well, I’ll give 2 topics because I’m not sure which one you are talking about out of the two things I was mentioning:<br>1. Dummy flinching; in SSF4 hitbox videos against a move with 4f start-up the hit animation (the dummy flinching) would occur at the 5th frame (1f right after the move becomes active), so if SFxT works the same way, Cody with blue/white stripe should get hit animation at 8th frame since cr.HK has 7f start-up, and the “turning” part of the bumblebee Cody in fact becomes the 1st frame of start-up if you count it frame by frame.<br><br>2. Now about the turning part. I am not sure on this one, but we have to take into consideration that Cody on the right side wasn’t doing anything when he could’ve gone into pre-jump frames, but the one with cr.HK flinched. I’m talking about this difference. A sure-fire way to know whether this is the 1st frame or not would be
recording cr.HK on dummy record to “hold back” so that we would know
which part of frame causes blocking motion. If the “flinching” part causes block motion, it means this is the 1st frame, which confirms the theory that the last frame of recovery is cancellable, if it doesn’t, it means this doesn’t have anything to do with cr.HK, which means things get more complicated.<br><br>Since you did such hard work to test this, I’m not going to make any major requests, but if you WANT to do this, then feel free to:<br>- Doing the same thing with cr.HP -> link, but this time with Knife cr.HP -> cr.MK and Knife cr.HP -> Knife cr.MP (both of them combos as well), then comparing the “turning” part of the bumblebee Cody with the Knife cr.HP -> cr.HK. If the motions are all different in three of them at the “turning frame”, it proves our theory. If not, it most likely disapproves.<br><br>I’ll find a recording software in PC and do this if I can.<br></div>
</blockquote>
I was speaking about bumblebee cody. didn’t noticed a difference for the dummies.<br><br>It isn’t that much hard work.<br><br>Is it move specific btw? I was trying to combo cl.lp into c.hk to no avail (didn’t tried it for long tho). <br><br>I can send you fraps if you don’t know a way to get it “somewhere else” ;)<br>