Sunday, July 13, 2008

Super Diagram Turbo - Version 1

For the chart below I contacted SSF2T players from all over the US and asked them for their take on their favorite characters' match-ups. A '1' means their character has a big disadvantage. A '9' means they've got a big advantage. '5' means even.

The values in the +/- column indicate the difference between the sum of all the numbers in the row and the baseline value of 80.

There are some notes that players sent me and additional discussion in the post below this.

Thank you very much to everyone who helped out with this project! I think people will find it really interesting.

Ryu-6776756367664436+9John Choi
-6574544356453336-7David Doyle
-5586663377776437+10Jesse Howard
Ken4.5-5.564.5664365.56555.546+2.5Wes Truelson
O.Ken55-87772288874627+13Jesse Howard
48-96747396517719+13Justin Wong
E.Honda332-4674677725546-2Bill Gould
Chun-Li4677-8245236855440Chris Doyle
5658449-277765245+6Axel Kelly
Dhalsim66.58564.587-86775676+23Jason Cole
6.567554.577-75.5764.5576+16Arturo Sanchez
T.Hawk553433.5574-5561.5444-11Henry Cen
5.55-45.554.543.5-3.54.5445.535.5-8Jason Wilson
Cammy3550447279-523624-12James Chen
Fei-Long44.5-3.54.5464465.5-44444-9Jason Wilson
DeeJay4.55.56857.575.5477.56-564.56+15Afro Legends
Boxer66.575564.5656.5765-55.55.5+11.5Afro Legends
5447578446994-627+11Julien Beasley
6657675868.5776-566+21.5Henry Cen
Claw66636897579964-43+18Julien Beasley
54575676587766-66+16Julian Robinson
O.Sagat777474663677654-4+10John Choi
688788883887852-5+27Jason Nelson
999757791995742-7+26Justin Wong
Dictator5443566546774371--3Julien Beasley
55-4445.56545.56544.54.5--3Jason Wilson

UltraDavid posted this in the comments below:
Here's some analysis of this data: [click for full size]

It's a set of tier rankings, each compiled in a different way and split into 4 tiers as reasonably as possible. The results are also compared to a couple of Japanese tier lists (and the reason I picked 4 tiers was just because the Tougeki one came separated into 4 already). It's interesting to see how the lists are mostly the same but how a couple of characters are valued pretty differently by our respective scenes.


  1. This show how much new sagat is considered in the US

  2. Thanks for putting this together! I was actually looking for a character ranking system for ST. Much appreciated!

  3. axel kelly needs to play jodim's zangief. then he'll change his mind about putting that 9.

  4. Wow- great work, I like the method and I could see this being referred to often in the future. Very interesting. Juilien's scheme seems a nice complement to this (actual tournament data). Have you ever read The Wisdom of Crowds by James Surowiecki?

  5. Spinalblood - New Sagat is pretty rare no matter where you go. Even yaya has supposedly switched to Guile.

    919 - I asked XTG for his opinion on Zangief (both him and Axel are NW US) but haven't heard back from him yet.

    David - I don't think that Julien's program keeps track of character usage.
    The Wisdom of Crowds
    I'll check this out.

  6. Are there really no DeeJay players? If you just go by Afro's matchups, Deejay is top-tier.

  7. excuse the ignorance since I don't play tournaments:

    Is Akuma a soft banned character or is he just low tiered so people don't play him.

  8. 957 - Afro Legends wasn't the only person I contacted for DJ. Version 2 will hopefully have more numbers for that section.

    shimo - Akuma is strictly banned from tournaments everywhere.

  9. O Ken vs Dic : 9-1 according to Justin W. wtf ? Is that a typo ? Or maybe he only played versus Dic players who can't punish whiffed lp dp's.

    I rate this match 5-5.

    His Ken vs O Ken rating is also rather strange to say the least...

  10. o.sagat vs blanka 4-6

    i wouldn't put this in blanka favour...

    boxer vs o.sagat 2-8
    ryu vs hawk 9-1
    o.sagat vs ryu 9-1

    These values are ridicolous

  11. Tried to say it before, wasnt posted though - some numbers are really off. people need to stop and think before they say a matchup is 9-1. a 9-1 is absolutely unwinnable.
    also, gief winning over chun, fei, dhalsim (!), dj is just ridiculous.

    i'd also love to know why according to wong o.sagat vs. fei is a 5:5. honest question.

  12. I don't know about the other guys, but in posting my matchups I didn't look at the numbers as being how many matches a character would win out of 10, I looked at them as being relative ratings of how difficult a match was (which I think made sense given the instructions and some private messages Nohoho and I traded). We were also told only to use whole numbers, so I stuck with that. I ended up rating Honda v Gief as a 9 for Honda and a 1 for Gief because in my opinion it's Honda's easiest fight and Gief's hardest and among the easiest/hardest matchups in the game.

    I'd redo some of my matchup numbers a bit if they were supposed to be how many matches I think the character would win out of 10.

    That said, yeah, I totally agree that some of these numbers are way off base.


  13. I looked at them as being relative ratings of how difficult a match was

    This was how the Melty Blood diagram that inspired this project worked. The numbers were supposed to indicate the difficulty of any isolated match. I didn't bother to stipulate this per my "make chart first, ask questions later" philosophy.

    Arcadia went on to discuss how the theoretical values in the chart differed from actual (high level tourney) match results. Some of the MB characters do better than theory/popular opinion dictates others do worse.

    Like the one girl with the quick overhead tends to win yea more in practice than in theory because people fail to block it. And the other girl with the really fierce offense does worse in practice than in theory because her rushdown gets parried by top guys.

    re: O.Sagat-Fei
    Justin played against Jason Wilson @ Fei at MWC. I was distracted and didn't get to watch the whole match but I had watched dude earlier in casuals and it was some of the most impressive ST playing I've ever seen.

    gief winning over chun
    Should we take matches vs. Jodim into account or not? Axel didn't. Chris did.

  14. Here's some analysis of this data:

    It's a set of tier rankings, each compiled in a different way and split into 4 tiers as reasonably as possible. The results are also compared to a couple of Japanese tier lists (and the reason I picked 4 tiers was just because the Tougeki one came separated into 4 already). It's interesting to see how the lists are mostly the same but how a couple of characters are valued pretty differently by our respective scenes.


  15. well, about taking into account jodim matches - this is not the way it works, at least in my book. you can't just say "i win x amount of matches against pro player y", because that would suggest you both are at an equal skill level (and that is something we don't know). also, since this was from gief's perspective, it makes even less sense, since we don't know the quality of the chuns jodim played against.

  16. Dang.. I feel that I was too hard on poor Blanka after seeing Nohoho's views.

    I used many of my games agaist Cole, Giga, Afro, DGV, Ultra, Jason Nelson and a other of hand full of players. But, maybe I am just not good enough with him! LOL!

    Sorry Nohoho! I know Blanka is your boy!


  17. I could modify my ranking program to start taking into account character selection. This could be interesting, as we could then see how tournament results pan out.

  18. Ultra - Very cool. I'll add that to the main post later.

    AtTheGates - Yeah I guess I was teasing. Though I feel like the whole theory vs. practice thing is kind of personal. I mean everybody has to make up their own mind. Whether they want to factor in Jodim matches or not etc. heh. Zass has it right in his comments that its hard to sum up one's take on a matchup in a single variable.

    fatboy - Man..... except for Blanka-Guile we really only differ by degrees. Your blanka match breakdowns are always on target.

    Zass - It seems like there isn't enough data out there for that. People don't keep detailed records let alone post them up online.

  19. Have the Japanese posted a complete matchup chart anywhere? I find it strange that 3S has a full matchup chart, but st only has a tier list (and it's not even ordered within tiers?).

  20. Gian has been writing comments in Tougeki Spirits to go with that tier list but they've been common sense stuff so I haven't reposted them. Part of the reason they've drifted away from charts is that characters with few weaknesses (Chun) can wind up too far below characters that steamroll but have achilles heels (O.Sagat.)

    I think the chart that Spence mentioned (from T.Akiba's joint) is the most recent one. Like he says it's pretty old.

  21. IC. That's cool. Is that also why they didn't bother ordering within the tiers as well?

    I like T. Akiba's chart, especially since it includes all O.chars. Too bad it's old. Good thing I still have NKI to tell me my matchups [I <3 Chun].