The KOF 98 UM Frame Data Project

[AcidicEnema]

Unfortunately, the site doesn't have any frame data on Yashiro, only his hitbox data. Nothing on Heidern.

[AtTheGates]

Its possible that they might have nerfed Ryo's crouch C in UM. There's a simple way to test this.

In 98, when I map player 1 (Kyo's) C button on the same button as Player 2 (Ryo's) C button and make Ryo do his crouch C and Kyo do his close s.C together right in front of each other, Ryo's crouch C would win every single time. This is consistent with the information from Dune's page which states that Ryo's crouch C is 1 frame faster than Kyo's close s.C.

I'm not sure if its possible to test this the same way on the PS2, but if it is, you could try and see if Ryo's crouch C still beats Kyo's stand C. If it does, that means there is a problem with your method. If not, then it shows that Ryo's crouch C was indeed nerfed.

2C is 5f fast after my methodology. New fastest move so far. Also wins against ryos 5C(c) when done at the same time =)


edit: shit i just realised you meant crouching C with c.C. yeah, it's 5 frames fast, so the method looks good.
does the site actually say that ryo's 2C is 1f fast? 1 + 4 = 5. that would be ridiculous.

[AcidicEnema]

Yep. That's what I meant when I said "Btw, according to this relationship, Ryo's c.C should take 5 frames to activate, according to your methodology... Correct?"

Yep the website says that Ryo's crouch C is '1, 2+3, 23'. This means, 1 frame to activate, 2 attacking frames for the first part of the move, 3 attacking frames for the second part of the move, 23 frames to reocver.

[AtTheGates]

hey, this is slowly beginning to make sense to me.

the 4 additonal frames on my data are the frames that happen between pushing the button and the moment the attack actually gets triggered.

this is the frame window i was talking about, in which you can still change directions (standing A, immediately crouch will give you crouching A).

i have a feeling the data is accurate =)

by the way, about subtracting those 4 frames - i guess it's a matter of taste - after all, if we take the+4 frames, this the actual time that passes between the button press and the attack, and not the acceptance of the move and the attack. For now, i will not subtract the 4 frames. Maybe i'll do it in the end, once everything is finished. Also, we have no idea about whether we need to subtract those 4 frames for special moves as well.

[AcidicEnema]

If that's the case, I'l copy all the frame data from Dune's webpage onto this thread in the next few days. Will be good to compare against findings on the PS2.

[AtTheGates]

excellent. Once we have like 95% accuracy, we can finally judge if any moves changed etc.



============
RYO
===========

5A(c) - 7f
5A(f) - 7f

5B(c) - 7f
5B(f) - 9f

5C(c) - 6f
5C(f) - 11f

5D(c) - 8f
5D(f) - 16f

6A - 19f

CD - 9f

2A - 7f
2B - 7f
2C - 5f (we have a winner!)
2D - 11f

j.A - 10f
j.B - 7f
j.C - 10f
j.D - 11f
j.CD - 15f
--------------------
623A - 3f (i guess this contradicts the 4f rule - but my gut feeling tells me specials dont need the subtraction anyway.)
623C - 5f

214A - 13f
214C - 20f

236B - 28f
236D - 31f

63214B - 7f
63214D - 10f (could be 9, not sure anymore)




=======
SAISYU
=======


5A(c) - 6f
5A(f) - 8f

5B(c) - 8f
5B(f) - 10f

5C(c) - 5f (!)
5C(f) - 13f

5D - 11f

6A - 21f
6B - 24f

CD - 14f

2A - 6f
2B - 6f
2C - 9f
2D - 14f

j.A - 8f
j.B - 8f
j.C - 11f
jump up, j.C - 13f
j.D - 11f
j.CD - 12f
--------------------

623A - 4f
623C - 4f

214A - 12f
214C - 18f

623B - 10f
623D - 18f

63214B - 18f
63214D - 30f

===============
O.YASHIRO
===============

see yashiro.

=============
SHERMIE
=============


5A - 9f

5B[1] - 9f

5C(c) - 8f
5C(f) - 13f

5D - 15f

6B - 19f

CD - 11f

2A - 9f
2B - 9f
2C - 8f
2D - 14f

j.A - 10f
j.B - 10f
j.C - 10f
j.D - 13f
jump up, j.D - 10f
j.CD - 13f

[Gunsmith]

so Ryo's C is the fastest dwn C? and hasn't been nerfed for UM?

IieKyo showed me how fast it was on GGPO - yikes - and then I started abusing it lol great stuff

keep up the great work guys!

[AtTheGates]

=======
IORI
=======


5A(c) - 8f
5A(f) - 10f

5B(c) - 8f
5B(f) - 10f

5C(c) - 6f
5C(f) - 12f

5D(c) - 9f
5D(f) - 14f

6A - 12f
6B - 28f

CD - 11f

2A - 9f
2B - 9f
2C - 11f
2D - 9f

j.A - 8f
j.B - 9f
j.4B - 8f
j.C - 13f
j.D - 10f
j.CD - 13f

---------------------
623A - 6f
623C - 6f

214A - 9f
214C - 12f

623B - 25f
623D - 31f

[AcidicEnema]

Alright. Here's the 98 frame data from the Dune webpage for: Kyo, Ryo, Benimaru, Goro

I'll first be posting the information, in the following format:

(Move Command) - (Starting frame, active frame, recovery frame). (AtTheGate's Data for comparison)

If there are more than 3 numbers in the frame data I provide, this means that the attacking move has more than 1 active component. E.g., 4, 5, 5, 6 means 4 frames to turn active, 5 frames for the first active portion, 5 frames for the second active portion, 6 frames to recover. I won't be posting special and DM moves yet, since the special/DM move data is fairly complicated and I haven't figured it out yet.

CD attacks will 2 different sets of frame data. I'm assuming that the 2nd set refers to when the CD is used in a GCCD situation.

Finally, for those where AtTheGate has already provided data, I'll be posting at the top of each character the moves where there are discrepencies between my data and AtTheGate's data. This means that either 1) one of the two data sources is incorrect or 2) there was a change between KOF 98 and 98UM. Further verification will be needed.

===================
KYO
===================

Discrepencies- 5D(f), CD, j.2C, j.CD

5A(c) - 3,4,5. 7f
5A(f) - 3,4,5. 7f

5B(c) - 5,4,8 9f
5B(f) - 7,3,13 11f

5C(c) - 2,3+9,9 6f
5C(f) - 12,2,17 16f

5D(c) - 4,3,22 8f
5D(f) - 14,5,12 14f

6B - 11,7,3,7,18 15f

3D - 2,3,12,4,24 6f (!)

CD - 19,4,16. 3[M],4[M],18. 19f

2A - 3,4,5 7f
2B - 3,4,5 7f
2C - 4,3+3,20 8f
2D - 11,3,19 15f

j.A - 5,8 9f
j.B - 5,9 9f
jump up, j.B - 5,9 9f
j.C - 6,4 10f
j.2C - 6,5 13f
jump up,j.D - 7,8 11f
j.D - 9,5 13f
j.CD - 17,4 17f


============
RYO
===========

Discrepencies: 6A, CD, j.D, j.CD

5A(c) - 3,4,9 7f
5A(f) - 3,3,8 7f

5B(c) - 3,7,8 7f
5B(f) - 5,2,12 9f

5C(c) - 2,10,13 6f
5C(f) - 7,6,24 11f

5D(c) - 4,9,16 8f
5D(f) - 12,4+4,16 16f

6A - 17,9,11 19f

CD - 9,8,18. 1[M],8[M],18. 9f

2A - 3,5,5 7f
2B - 3,5,5 7f
2C - 1,2+3,23 5f
2D - 7,5,18 11f

j.A - 6,7 10f
j.B - 3,7 7f
j.C - 6,4 10f
j.D - 5,4 11f
j.CD - 15,4 15f

============
BENIMARU
===========

5A(c) - 3,3,9
5A(f) - 3,4,9

5B(c) - 3,5,7
5B(f) - 3,7,9

5C(c) - 4,6(3+3),15
5C(f) - 5,6,18

5D(c) - 2,6(2+4),16
5D(f) - 11,5,16

6B - 19,3,27

CD - 17,4,20. 2[M],4[M],20

2A - 3,4,5
2B - 2,4,7
2C - 4,6(3+3),24
2D - 5,6,23

j.A - 6,10
j.B - 4,11
j.C - 6,3+7
j.D - 6,8
jump up D- 3,10
j.CD - 15,5

jump 2+D- 13,1,2,(2,2)*N,19

============
GORO
===========

5A(c) - 4,4,5
5A(f) - 4,5,6

5B(c) - 7,4,9
5B(f) - 4,5,10

5C(c) - 4,4,13
5C(f) - 9,9,16

5D(c) - 4,8,14
5D(f) - 5,6,16

6A - 20,2[T]+4[&T],26
3C- 3,4+9,16

CD - 15,4,26. 0,1[M],20. Does that mean that Goro's s.CD is a 0 frame move on counter? That can't be right...

2A - 3,5,5
2B - 4,5,6
2C - 5,4,19
2D - 4,7,12

j.A - 5,13
j.B - 8,17
j.C - 9,8
j.D - 7,7
j.CD - 15,4

[AtTheGates]

good morning,

i'll double check the contradictions, then post frame data for chris, goro, beni today.

i'll edit this post in a while.

UPDATE: it seems like all CD moves differ by 4 frames - i guess the 4 frame rule does not apply to CD moves (just like with special moves). if ALL CD moves are off by 4 frames (and if none are changed, of course), then this new rule holds true.

KYO:

5D(f) - changed from kof 98 classic! faster. You can see it with your bare eye. Also reflects the changelist i saw (http://efil4zaggin.com/guides/arcade/kof98um.txt), which says "Far C and Far D are faster.", although far C does not seem faster at all. I also tested both in classic & um vs. kyo's 236C, which is 18f fast. I got trades in classic, but in um, the 5D(f) beat 236C every single time.
CD - trades with shermie's 6B both in UM and classic. no change, no 4f difference -> see new rule.
j.CD - double checked, still 17 frames, and hard to test whether there was any change. I might try it anyway later.

j.2C - i made a mistake here! startup is 10f in UM.


RYO:
changelist says far B and standing CD is slower, seems wrong for far B.

CD - double checked, still the 9f i counted before in UM. In Classic, trades with Kyo's 5B, which is also 5, which means, classic should be 5, UM is 9 ->> CD got slower?
6A - double checked, still 19f in UM. Wins against Saisyu's 6A every time in classic, which *should* have the same speed. This would mean the classic data is wrong (provided that there was no change to saisyu#s 6A).

j.D - double checked, still 11f
j.CD - double checked, still 15f


=======
CHRIS
=======

5A - 8f

5B(c) - 8f
5B(f) - 11f

5C(C) - 7f
5C(f) - 11f

5D(c) - 7f
5D(f) - 12f

6A - 14f
6B - 14f

CD - 15f

2A - 8f
2B - 7f
2C - 6f
2D - 12f

3B - 12f

j.A - 8f
j.B - 9f
jump up, j.B - 10f
j.C - 10f
j.D - 10f
j.CD - 18f

-------------------
236A - 15f
236C - 23f

623B - 11f
623D - 13f

63214B/D - 19f




============
BENIMARU
===========

5A(c) - 7f
5A(f) - 7f

5B(c) - 7f
5B(f) - 7f

5C(c) - 8f
5C(f) - 9f

5D(c) - 6f
5D(f) - 15f

6B - 23f

CD - 17f (difference!)

2A - 7f
2B - 6f
2C - 8f
2D - 9f

j.A - 10f
j.B - 8f
j.C - 10f
j.D - 10f
jump up, j.D - 7f
j.CD - 15f (difference!)

j.2D - 17f




============
GORO
===========

5A(c) - 8f
5A(f) - 8f

5B(c) - 11f
5B(f) - 8f

5C(c) - 8f
5C(f) - 13f

5D(c) - 8f
5D(f) - 9f

6A - 24f
3C- 7f

CD - 15f (difference!)

2A - 7f
2B - 8f
2C - 9f
2D - 8f

j.A - 9f
j.B - 12f
j.C - 13f
j.D - 11f
j.CD - 15f (difference!)

[AcidicEnema]

- I don't think you should rely on that changelist. The person who did it is notorious for just stealing random stuff of the web and doing bad translations. I think when we're done we'll definately be able to come up with a REAL list of changes, as compared to what that FAQ writer pretended to do.

- Here are the frame data for Terry, Andy, Joe, Robert and Ralf from the Dune site. As before, i will not be copying the frame data for special moves and DMs yet since those are quite complicated.

============
TERRY
===========

5A(c) - 3,4,5
5A(f) - 3,4,5

5B(c) - 4,5,7
5B(f) - 5,3,11

5C(c) - 3,3,0,7,15
5C(f) - 5,8,17

5D(c) - 3,10,16. 2[M],10[M],16 (second set of data is probably for his CD counter)
5D(f) - 13,3,22

6A - 11,3,3,3,7+19
3C - 3,3+3,24

CD - 23,3,23

2A - 3,2,3+5
2B - 3,3,8
2C - 5,8,14
2D - 10,8,22

j.A - 4,7
j.B - 3,8
j.C - 4,9
j.D - 4,7
j.CD - 16,4

=======
ANDY
=======

5A(c) - 3,4,7
5A(f) - 5,5,10

5B(c) - 3,5,5
5B(f) - 5,5,11

5C(C) - 2,4,0,4,20
5C(f) - 7,8,17

5D(c) - 8,5,24
5D(f) - 14,3+9,18

3A - 7,2,0,5,16+6
6B - 16,4,0,4,12+9

CD - 16,5,33. 2[M],5[M],33.

2A - 3,5,5
2B - 4,5,6
2C - 5,7,18
2D - 9,4,24

3B - 12f

j.A - 4,8
j.B - 6,11
jump up, j.B - 5,11
j.C - 6,5
j.D - 5,9
jump up, j.D - 5,9
j.CD - 10,4

============
JOE
===========

5A(c) - 3,3,9
5A(f) - 3,2,8

5B - 7,3,7. 0,5[M],23. (Second one is as a CD counter)

5C(c) - 2,2,20
5C(f) - 7,2,16

5D - 6,3,28

6B - 13,2,25
3B - 13,4,24

CD - 20,2,25

2A - 3,3,6
2B - 3,3,7
2C - 4,3,19
2D - 8,4,9+14

j.A - 3,7
j.B - 6,12
j.C - 4,9
j.D - 4,12
j.CD - 11,5

j============
ROBERT
===========

5A - 3,6,5

5B(c) - 3,6,6
5B(f) - 5,7,7

5C(c) - 4,10,15
5C(f) - 5,10,21

5D(c) - 7,8+6,13
5D(f) - 8,4,25

6A - 9+7,5,4+8
6B - 16,3,30

CD - 11,4,23. 0,2[M],23.

2A - 4,6,6
2B - 3,6,5
2C - 3,2+3,25
2D - 7,5,24

j.A - 4,11
j.B - 4,9
j.C - 4,9
j.D - 8,7
j.CD - 11,6

============
RALF
===========

5A(c) - 3,4,5

5B(c) - 4,4,5
5B(f) - 4,4,6

5C(c) - 3,3,22
5C(f) - 8,5,17

5D(c) - 11,4,17
5D(f) - 9,7,21

CD - 19,4,21. 4[M],4[M],21.

2A - 3,4,5
2B - 4,5,6
2C - 4,8,9
2D - 7,9,16

j.A - 3,7
jump up, j.A- 8,- Yes. It stays active for as long as he's in the air, which is one of the main reasons why its so overpowered.
j.B - 6,6
jump up, j.B- 4,6
j.C - 6,9
jump up, j.C- 6,13
j.D - 8,5
jump up, j.D - 12,3
j.CD - 14,4

[AtTheGates]

i'll do terry and andy now. unfortunately, i got a lot of work to do, so i guess more updates will have to wait until wednesday.

[will be edited soon]



========
TERRY
========

5A(c) - 7f
5A(f) - 7f

5B(c) - 8f
5B(f) - 9f

5C(c) - 7f
5C(f) - 9f

5D(c) - 7f
5D(f) - 18f <--- 1f difference. double checked my data.

6A - 15f
3C - 7f

CD - 23f

2A - 7f
2B - 7f
2C - 9f
2D - 14f

j.A - 8f
j.B - 7f
j.C - 8f
j.D - 8f
j.CD - 16f

-------------------
623A - 6f
623C - 6f

214A - 14f
214C - 21f

623B - 7f
623D - 10f


214B[1] - 10f
214B[only 2nd hit] - 16f
214D[1] - 18f
214D [only 2nd hit] 28f

41236B - 12f
41236D - 17f

[AtTheGates]

=======
ANDY
=======

5A(c) - 7f
5A(f) - 9f

5B(c) - 7f
5B(f) - 9f

5C(C) - 6f
5C(f) - 11f

5D(c) - 12f
5D(f) - 20f - 2f difference, should be faster

3A - 11f
6B - 24f - 4f difference, should be faster

CD - 16,5,33. 2[M],5[M],33.

2A - 7f
2B - 8f
2C - 9f
2D - 12f - should be 1f slower

j.A - 8f
j.B - 10f
jump up, j.B - 9f
j.C - 10f
j.D - 9f
jump up, j.D - 9f
j.CD - 10f

-------------------------
1,6,A - 7f
1,6,C - 9f

623A - 6f
623C - 7f

41236A -
41236C -

41236B - 12f
41236D - 18f




============
JOE
===========

5A(c) - 7f
5A(f) - 7f

5B - 11f

5C(c) - 6f
5C(f) - 11f

5D(c) - 10f
5D(f) - 12f (note: far 5D is exactly the same move as close 5d, but if you aren't within close range,
it hits 2 frames later. So one could argue that it is a distinct move.)




[more coming soon]

[AcidicEnema]

- Here's the frame data for Clark, Chizuru, Kim and Shermie from the Dune site.

- That's it for the frame data available from that site. All the other characters don't have their frame data on that site.

- I finally got the PS2 version so I'll be starting on the UM frame data using AtTheGate's method. Will be starting with Krauser, Rugal, Brian, Lucky and Heavy D!

===========
CLARK
===========

Discrepencies- CD, 2D, j.CD

5A - 4,4,5

5B(c) - 2,5,5
5B(f) - 3,5,5

5C(c) - 3,4,0,7,16
5C(f) - 11,6,21

5D(c) - 9,3,21
5D(f) - 8,4,16

6B - 13,4,22

CD - 22,5,16. 6[M], 5[M], 16.

2A - 4,4,5
2B - 4,5,6
2C - 5.9.9
2D - 7,7,17

j.A - 3,7
j.B - 3,5
jump up, j.B - 3,5
j.C - 6,5
j.D - 4,7
jump up, j.D - 11,3
j.CD - 13,4

=======
CHIZURU
=======

5A - 4,2,7

5B(c) - 7,4,9
5B(f) - 5,6,7

5C(C) - 2,4,14
5C(f) - 5,2,2,3,13

5D(c) - 3,4,22
5D(f) - 12,4,20

6A - 20,3,24. 14,3,24. Second set of data probably refers to when the move is cancelled into.
6B- 7,12,2 4,24
3B - 9,13,12

CD - 27,2+6,18. 0,2[M],6[M],18

2A - 3,2,7
2B - 3,3,8
2C - 3,2,2,2,25
2D - 7,5,17

j.A - 5,7
jump up, j.A - 5,6
j.B - 6,8. 9,8. Not sure why there are 2 sets of data.
jump up, j.B - 7,7
j.C - 4,4,4,4
jump up, j.C - 5,11
j.D - 7,4
jump up, j.D - 7,9
j.CD - 15,4

===========
KIM
===========

5A(c) - 4,4,5
5A(f) - 4,5,7

5B(c) - 5,5,7
5B(f) - 5,7,7

5C(c) - 2,3,3,8,19
5C(f) - 13,13,15

5D(c) - 3,9,15
5D(f) - 7,5,18

6A - 12,8,8,13
6B - 21,3,0,3,16

CD - 10,10,15

2A - 4,4,5
2B - 4,4,5
2C - 10,7,15
2D - 7,7,20

j.A - 4,7
j.B - 3,8
j.C - 8,5
jump up, j.C - 3,10
j.D - 6,9
jump up, j.D - 8,8
j.CD - 13,6

=============
SHERMIE
=============

Discrepencies- CD, j.CD

- Far C is a definately faster in UM. Based on this data, it looks like it's 5 frames faster.

5A - 5,2,6

5B[1] - 5,3,5,9,5

5C(c) - 4,6,16
5C(f) - 14,2,0,3,27

5D - 11,5,26

6B - 15,4,4,12,4,4,4,11

CD - 11,6,17

2A - 5,2,6
2B - 5,5,9
2C - 4,3,7,17
2D - 10,8,19

j.A - 6,8
j.B - 6,6
j.C - 6,8
j.D - 9,8
jump up, j.D - 6,10
j.CD - 13,4

[AtTheGates]

yo,
good to see you helping out! i'll continue gathering data tonight, too (needed a break).