Sidestep Tier List

Page Splits
Share This Topic
Share
Subscribe/Jump Subscribe This Topic
< >
noodalls
Moderator
Joined: May 2003
Posts: 2404
From: Australia
PSN: noodalls
#1 “Quote” Edit Post
http://img718.imageshack.us/img718/2808/sidesteppingtierlist.png
http://img718.imageshack.us/img718/...ingtierlist.png

The question has been asked who has the best sidestep in Tekken 6. While there were some very educated guesses in that thread, I felt that this probably deserves careful investigation. I set about testing with my programmable stick for 38/40 of the characters in the game.

Paul was set to either press LP (left punch) or RP (right punch) on F(rame)10, which will connect on F20. The character being tested was set up to press d on a single specified frame. The two characters were walked together to prevent distance or allignment playing a role.
Where results were inconsistent, the first result to occur three times consecutively was used (e.g. if on a specified frame sometimes the punch was avoided, and sometimes it connected, whichever of hit or avoided occurred three times in a row was used.)

Eddy and Christie were not tested as most people just mash with them, so the chances of getting a sidestep without a button being pressed first were too low to consider them for testing.

Column A ? name
Columns B-E ? Sidestepping LP done on F10 (hits on F20)
Column B ? Earliest frame the character is able to avoid a LP by sidestepping towards it (corresponds to length of sidestep) = pressing d on 1P side
Column C ? Latest frame the character is able to avoid a LP by sidestepping towards it (corresponds to speed of sidestep) = pressing d on 1P side
Column D ? Earliest frame the character is able to avoid a LP by sidestepping away from it (corresponds to length of sidestep) = pressing d on 2P side
Column E ? Latest frame the character is able to avoid a LP by sidestepping away from it (corresponds to speed of sidestep) = pressing d on 2P side

Column F ? Earliest frame the character is able to avoid a RP by sidestepping away from it (corresponds to length of sidestep) = pressing d on 1P side
Column G ? Latest frame the character is able to avoid a RP by sidestepping away from it (corresponds to speed of sidestep) = pressing d on 1P side
Column H = Earliest frame the character is able to avoid a RP by sidestepping towards it (corresponds to length of sidestep) = pressing d on 2P side
Column I = Latest frame the character is able to avoid a RP by sidestepping towards it (corresponds to speed of sidestep) = pressing d on 2P side
Column J = sum of frames where character is able to sidestep a LP
Column K = sum of frames where a character is able to sidestep a RP
Column L = sum of columns J and K, i.e. the total number of frames where a character can sidestep and avoid either a LP or RP. This provides the order for the list.


I feel this has pretty much answered the question. The only remaining questions would be, how do moves such as Steve's sways, Julia's 3+4 etc. fit into this, which I may come back to answer in the future. The testing takes a lot of time (partially because of long loading times, and also because sometimes it takes a while to get consistent results) so that is something for later.

Post any questions here, but please try to understand the data before asking questions, the explanation should be adequate.
noodalls
Moderator
Joined: May 2003
Posts: 2404
From: Australia
PSN: noodalls
#2 “Quote” Edit Post
Reserved
Image doesn't show up in this forum, so will summarise the main results.
The number is the total number of frames on which a given character can sidestep towards and away from a LP and a RP. This number reflect how fast and how long their sidestep is, and so how good they are at sidestepping.

51 - Lili
50 - Alisa
47 - Nina
45 - Ling
44 - Leo, Zaf, Bruce
43 - Julia, Steve, Anna, Lee, Auka, Lars, Miguel, Raven, Feng, Heihachi.
42 - Mokujin, Roger, DevilJin, Paul, Lei, Kazuya, Armor King, Dragunov, Bryan
41 - King, Baek, Jin, Law, Yoshi, Wang
40 - Ganryu, Bob, Hwoarang
38 - Marduk
37 - Kuma
33 - Jack 6

Sidestepping LP
25-Lili
24-Alisa
23-Ling, Zaf, Nina
22-Leo, Julia, Anna, Lee, Roger, Auska
21-DevilJin, Paul, Lars, Miguel, Bruce, Steve, Mokujin, Lei, Raven, Feng, Kaz, Jin, Law, Hwo, Hei, Yoshi, Wang
20-Baek, Dragunov, Bryan
19-Kuma, Bob, King, Aking, Ganryu
17-Marduk
13-Jack6

Sidestepping RP
26 - Lili, Alisa,
24 - Nina
23 - Bruce, Armor King
22 - Ling, Leo, Steve, Lars, Miguel, Raven, Feng, Heihachi, Dragunov, Bryan, King
21- Zafina, Julia, Anna, Lee, Asuka, Mokujin, DevilJin, Paul, Lei, Kazuya, Baek, Ganryu, Bob, Marduk
20 - Roger, Jin, Law, Yoshi, Wang, Jack6
19 - Hwoarang
18 - Kuma

Last edited by noodalls on Jun 5th, 2010 at 23:05

Abigan2K
Iron Fist God
Joined: Jan 2001
Posts: 3850
From: Philippines
#3 “Quote” Edit Post
whos other bob/? hwoarang?

as expected..lili has the best ss in the game
White Angel
Iron Fist God
Joined: May 2006
Posts: 1091
From: United States
PSN: xWhiteAngelus
#4 “Quote” Edit Post
I'm amazed that Nina is in the top 3 on the sidestepping tier list.
NeoX
Tekken Lord Black
Joined: Jun 2004
Posts: 984
From: USA Texas
PSN: NeoX2634
XBL: TheGreatNeoX
#5 “Quote” Edit Post
Originally posted by White Angel
I'm amazed that Nina is in the top 3 on the sidestepping tier list.

I'm not... ss+2
Signature = Stupid
= The King of Iron Fist
= Houston Special
Blonde Hair = Top Tier
tyler2k
____________________
Joined: Jun 2006
Posts: 2794
From: USA California
PSN: tyler2k1
#6 “Quote” Edit Post
Interesting, though not surprising, to see a Gaussian distribution. Props to noodalls.
Signature My fighting game website - NOW BACK!!! | twitch.tv/tyler2k - My twitch.tv channel |
Tutorial Series: iWS tutorial video | CC and CDc tutorial video | Hit Confirm tutorial video
Beginner Series: Throw System (Part 1, Part 2)
aris
The Beard Signal
Joined: Jan 2004
Posts: 864
From: USA California
#7 “Quote” Edit Post
Awsome thread. Hopefully you can make a backdash tier list in the future as well.
0kills
Iron Fist God
Joined: Jan 2006
Posts: 4936
From: Philippines
noodalls
Moderator
Joined: May 2003
Posts: 2404
From: Australia
PSN: noodalls
#9 “Quote” Edit Post
Originally posted by aris
Awsome thread. Hopefully you can make a backdash tier list in the future as well.


That was my thought as well. Any thoughts on the best way to go about it?

My thoughts were as follows - the backdash has two components, the startup and duration (just like the sidestep). To test this would require two different components. First, sidestep (back for two frames, neutral for one, back for two frames) and have the other player (Paul again probably) press LP on later and later frames until the jab doesn't hit the character at the start of their backdash.

To check the length of the backdash, let the backdash complete fully and then press F for a set number of frames followed by LP. Obviously, the fewer frames holding F before the jab connects the shorter the backdash.

Any other ideas how to approach this?
Tysan
Iron Fist God
Joined: Jan 2000
Posts: 1040
PSN: TrueTysan
#10 “Quote” Edit Post
I think that standing stance hitbox is also related to how things whiff when diferent characters backdash Alisa and Lili have slim standing stances that might help with backdash evasion, also I was playing yesterday a Kuma player and it was terribly difficult to get low hits with Kuma backdashing (its also very hard to score a low clean hit like Annas db+3+4 in range 0...), Kuma main hitbox is the huge body but his legs are quite farther so that should be why his backdash its so effective against lows.

So maybe a set of standard moves from X character should be used to test different backdash evasion against high/mid/low/throw attempts.

PD.- Really awesome work with the SS tier list

Last edited by Tysan on Jun 7th, 2010 at 13:43

Signature <SSnake> Why are you calling me Brother !!??
<SSnake> Who the Hell are you !!??
<LSnake> Im You...
<LSnake> ...Im Your Shadow !!
tekken6bryanfur
Iron Fist God
Joined: Jan 2010
Posts: 1234
From: Brazil
XBL: JSmith007
#11 “Quote” Edit Post
Originally posted by noodalls
I feel this has pretty much answered the question. The only remaining questions would be, how do moves such as Steve's sways, Julia's 3+4 etc. fit into this, which I may come back to answer in the future. The testing takes a lot of time (partially because of long loading times, and also because sometimes it takes a while to get consistent results) so that is something for later.

when you do, don't forget bryanfury has a humongous spin sidestep. his b2~f1~ssL is a just frame double punch cancel sidestep(one of the games hardest moves), with your program pad you could do it easily.
Royal_Flush21
Expert
Joined: May 2010
Posts: 134
From: New Zealand
#12 “Quote” Edit Post
basicly the smaller u are the better the ss. but what about beaks sss
noodalls
Moderator
Joined: May 2003
Posts: 2404
From: Australia
PSN: noodalls
The Invincible
Iron Fist God
Joined: Mar 2004
Posts: 1089
From: Philippines
#14 “Quote” Edit Post
Originally posted by tekken6bryanfur
when you do, don't forget bryanfury has a humongous spin sidestep. his b2~f1~ssL is a just frame double punch cancel sidestep(one of the games hardest moves), with your program pad you could do it easily.



Huh??!!


@noodalls: Thanks for the hard work you put into this. Your findings are awesome as usual.
Signature By the way, I consider myself as a cold and ruthless warrior... always in search of someone to slay... I live and believe in myself alone! Uninhibited by foolish emotions...

d/b+1 is your fate...
BlueBolt
Expert
Joined: Sep 2010
Posts: 140
From: Serbia and Montenegro
#15 “Quote” Edit Post
so... when im playing lars... who can ss a LP at earliest 3... does that mean i can ss a lp with max. -7 ?
and how should i understand the latest frames... i think i got something wrong :S
Clarity Ratios
1st Dan
Joined: Feb 2011
Posts: 10
From: USA California
#16 “Quote” Edit Post
Where results were inconsistent, the first result to occur three times consecutively was used


Do you happen to recall any specific characters being more inconsistent than others?
noodalls
Moderator
Joined: May 2003
Posts: 2404
From: Australia
PSN: noodalls
eX3Me
Virtuoso
Joined: Jul 2010
Posts: 221
From: Philippines
#18 “Quote” Edit Post
cool.. a sidestep tier list.. good thing my alisa is second to the top but jack is at the bottom.. T_T
Signature / ..Experience is the purest form of practice.. /
Ei8hTy SiX
Drunken Master
Joined: Sep 2011
Posts: 157
From: USA Iowa
XBL: Ei8hTy SiX
#19 “Quote” Edit Post
Today I found myself wondering about side-steps.

I really like the tests you conducted against Paul's jabs. I've found that information extremely useful. I also like the idea you have to test side-steps against Kazuya's strings. I'm very curious to hear the results of that one. It sounds like it will take a great deal of time to do for almost 40 characters. Good luck! I wish I could help.
Your initial test against jabs helps us begin to understand the different qualities a side-step has. It attempted to measure the various "speeds" and "length" that each character's side-step possesses. The speed was to be demonstrated by how quickly each character is able to avoid being punched. The length was to be demonstrated by how early each character can anticipate the jab and still avoid the punch. However, I don't believe the test against jabs has given an accurate measure of either speed or length. Let me add though that your type of specific "in-game" scenario testing may actually be the best sort of test to measure the overall effectiveness of each character's side-step.

You pointed out in your back-dash study that you believe back-dashing probably isn't character specific. It seemed that each character can back-dash at equal speeds and equal distances. The evasive results depend more on the character's size and stance. The back-dash tests may have actually been measuring the depth of each character's hit-box. It also brought about new questions, like whether or not the relative distance between characters has influence on the back-dash distance. I can understand how all this may have felt overwhelming and possibly caused you to throw up your hands and say, "That was a waste of time! I'm not going to get that deep into the testing."

It makes me wonder if the side-step test has resulted in a similar paradox. Does the test give a better measure of each character's "horizontal hit-box" (or width of there character) rather than the measure of each character's side-step? It probably measured both, and by measuring both, it didn't accurately measure either one. This is all assuming that there is in fact a difference between the different character's side-steps. Regardless of what your side-step test may have measured, the results and information it gave us has proven to be useful. I'm now able to somewhat confidently say, "This character only needs this many frames of an advantage to successfully side-step a left/right jab". Of course this info will only hold true when our characters are as they were in the test - at point blank range.

All of this has got me thinking. We want to know which characters have the best side-step and back-dash. I'm having ideas on how to more accurately measure a character's side-step and back-dash. I'm thinking about the best ways to test each of the individual qualities of a back-dash and side-step separately (e.g. distance, speed, duration). Let me explain a few ideas:

I first set out to measure the actual length of each character's side-step. I thought it would be interesting to see how many consecutive side-steps it would take for various characters to completely circle around their opponent. I chose to do this test in Manji Valley so that the snow would help me determine when the circle had been completed. I began with characters that lack any sort of sway or movement in there stance to prevent that from being a factor. This was especially important for the character who was being circled. I found that the average number of side-steps required to make a circle was around 13. This was a problem. With the number being so small, it wouldn't vary much between the characters. I also found that different characters take different lines or paths around. Some seem to make perfect circles in the snow. Others tend to inch forward with each side-step, so the line they created looked more like a spiral than a circle. By ending in a position that was closer to the opponent, they essentially didn't travel as far, but rather took a sort of short-cut around. This created a problem and was another major flaw in the test.
That forced me to think of a test idea that would eliminate the "inching forward" problem as well as get a larger value for the measurement.
My idea now is to put the character who's side-step is to be tested versus themselves. I think that the test to determine the length of each character's side-step needs to be done in a straight line rather than in a circle. To accomplish this, we can walk the two characters together so that they are touching each other nose to nose. Then we can have them perform identical synchronized side-steps into the background or foreground. To exaggerate and give better results, the test should be done across the largest possible stage. It will be necessary to use walls (and possibly corners) as the starting and end points. We can measure how many side-steps it takes for the "twins" to make it from one wall across to the other. It will help if the stage's floor has lines or if the stage's walls have markers to insure each pair is traveling along the same path. I've tried to do this with two controllers at the same time and continually fail an input at some point. This test will get much more consistent results with a programmable joystick. I believe this test would give us different results and rearrange the side-step rankings in regards to the LENGTH of each character's side-step.

A similar test could be done to more accurately measure the length of each character's back-dash. By putting a character behind themselves, with one in BT, you could have them perform identical synchronized back-dashes across the stage. We can then measure how many back-dashes it takes for the twins to make it from one wall across to the other. To be more accurate, and eliminate the size of the character from being a factor, we can measure how many back-dashes it takes for the character's foot (or toes) to get from one line on the floor to another, etc.

Another bit of information to test that I think would be useful, would be to determine the useful duration of a sidestep. Let me try to explain what I mean by "useful duration".
"One sidestep on F1 allows for a second SS to be input on F15."
"Sidestep attacks are accepted from F3-47."

I'm guessing that it's somewhere around F35 that the actual side-step animation ends. I'm guessing that it's probably closer to F25 or so that the effectiveness of the sidestep's lateral movement ends. So that is what I mean. I'm wondering when it is that the sidestep can be cancelled, yet still gain all of it's useful lateral movement frames. I haven't spent much time thinking of a method for this test. Now that I think about it... your values for sidestep "length" in the test versus jabs may have done just that. But it may be necessary to test this versus a linear move that takes longer to connect than a i10 jab. By using a move with more initiation frames, you should get more exaggerated results, similar to using a bigger stage in the other test I suggested.

These are just ideas that are popping in my head that I thought I'd share with anyone else who may be interested in testing the game's mechanics. Hopefully I'll have access to a programmable joystick someday, so I can just go ahead and do this crap and post my results instead of having to post my ideas.

Adios, my wifey Mary Jane is calling my name.
Signature Rip - "So... does Dragunov have any stances?" Aris - "No. He's heterosexual."

Steve - "Don't be jealous." Lei - "Know how Sergei got that scar on his mouth? Haha!"

LUYG Ep.7 24:40-25:12

 All times are GMT. The time now is 10:34

Page Splits
Moderator Tools
Forum Jump