Avoiding the Frame Data Trap

Frame data can shackle your creativity. If you don’t understand it completely – and nobody does – access to frame data can limit what you think is possible. There’s a whole range of fundamental tactics that top players have been using since Hyper Fighting, which are just now being explained in terms of frame data. Who knows how much longer that development path would’ve taken had frame data gotten in everyone’s way from day one?

Unless you can visualize how fighting game engines run better than i can, be careful about restricting all your ideas to frame data. Even the most experienced combo video makers commit mistakes when interpreting and applying that breadth of information.

Not to mention, published frame data neglects or misrepresents a broad spectrum of elements including jumping attack properties, cancelability windows, elusive hitboxes, effective ranges, pushback distances, and a whole slew of projectile characteristics. If you think about it, the sheer amount of critical information regularly left out of frame data tables is simply staggering.

The question is, how much of your game do you want to put on the line for some oversimplified numbers you read on a chart? In truth, the way to learn fighting games hasn’t changed since the first generation. You simply browse through the available cast, narrow down your choices to a handful of characters which appeal to you, and choose one to start with. Spend an hour or two learning their moves, and then it’s on to matches – either against the CPU or preferably against human competition.

Obviously you shouldn’t pass up any major opportunities to augment your knowledge by reading strategy articles, discussing matchups on SRK forums, or using frame data as a backup option. However, your core understanding of the game should always come from firsthand experience with your own senses and digits.

ComboVid.com - Fighting Game Combos, Tutorials, Matches, Screenshots, and Strategy

Frame data works much better as a quick reference than as a foundation. At the end of the day, nobody processes games numerically in real time. Even people like me, who cite frame data when answering questions, never treat live matches like a two-player math quiz.

Fighting game mastery is all about “feel” and it’s a serious mistake to drift away from that principle. For example, if you need a quick attack to counter a specific rushdown tactic, it helps to glance through frame data for the most promising candidates. At this point, anyone who’s had any experience with frame data will tell you that trusting the smallest startup number is a losing bet more often than not. Fighting games are usually balanced in such a way that the most impressive attack on paper tends to have the weakest hitbox properties as compensation.

The golden rule is: You never know until you test it. Most people have a much easier time coming up with new tricks and solutions when freely messing around in Training Mode, even though – or possibly because – it requires more work than basic arithmetic. You have to admit, playing the game is certainly more conducive to creativity than staring at a number sheet.

Equally important is understanding that fighting game strategy evolves in direct relation to players actively trying to bend and break the established rules of the engine. Frame data can still be helpful in explaining these deviations, but it’s nearly useless in deriving them. The point being, frame data will only lead you so far. Don’t stop there.

18 thoughts on “Avoiding the Frame Data Trap

  1. error1

    Reminds me of the first time I looked at Blanka’s frame data and noticed how bad his c.mp looked. I considered it his best normal move for it’s huge range and ability to dodge high and jumping moves and I spam it all the time at the right range.
    You would never notice it looking at the frame data tho. It has a 7 frame start up, it’s -5 on block, it’s not special cancel-able, and it can be blocked high. Still it consistently wins at footsies no matter how many of ken’s 4 frame c.mk the opponent has

  2. CPS2

    While I agree that there is a whole lot more to a game than just what they show in frame data, I feel lost without it =P. Tried playing Tekken 6 for a while, had no idea what my fastest moves where, had no idea what was unsafe, and didn’t know if any of the basic patterns I came up with were any good. So I sort of prefer to have it there so I can check stuff if I want, but yeah it’s useless for a lot of things, air attacks don’t make much sense with frame data, and taking advantage of hitting early or late during a juggle for different purposes is something you have to figure out another way. At least for playing safely, finding basic combos and knowing what you can punish, it does help you level up pretty quickly. Maybe it’s more essential in 3D fighters? Hard to say…

  3. Maj Post author

    Oh it’s definitely more important in 3D fighters because they’re a lot more mathy and RPSy in general. There’s no range game and no jump game, so it’s all mid/close range poking and counterpoking. The other critical difference is pushback, which is virtually nonexistent in 3D games. So the mindgame in Tekken ends up being way more straightforward, which makes frame data that much more central.

    You gotta remember, Tekken and VF started this whole frame data craze. Nobody cared about that shit before like Tekken 3, which was 1996. Street Fighter didn’t have frame data until Alpha 3, which was 1998. SF3 didn’t have frame data until 3S in 2000. Even CvS didn’t have it until CvS2 in 2002.

  4. Suirad

    Even in 3D fighters over focusing on frame data often obscures options such as movement which tends not to be referred to as much in the charts. Instances of moving hitboxes where attacks have different data depending on distance until contact are also fairly common.

    It seems like in SFIV in particular that the frames won’t add up when performing block strings since outside of chaining and special cancels you aren’t able to buffer one attack into another and so your timing has to be spot on to make attacks come out at full speed.

    I come from a 3D background myself, but I’ve had little success applying the frames I’ve seen for Street Fighter to actually playing the game.

  5. jamheald

    Can you explain to me rose’s spirals? I just checked out the frame data and the heavy one is safer.

  6. Rufus

    “If you think about it, the sheer amount of critical information regularly left out of frame data tables is simply staggering.”

    This is only too true. Of course, what’s even worse is that, absent the frame data, there really isn’t that much out there in measurable and testable information. Figuring out how you’re doing something wrong or whether something will work is staggeringly difficult for fighting games because there really is no context.

    Nobody – or at least close to nobody – even mentions fundamental stuff like Street Fighter 3 not locking people into blockstun.

  7. error1

    jamheald :
    Can you explain to me rose’s spirals? I just checked out the frame data and the heavy one is safer.

    lk has a four frame quicker startup then hk and is four frames slower on block. Similar to how sould spark works

  8. jamheald

    @error1
    Ahh i see. Well it’s useless cos i never use them if they are going to be bocked unless they’re at a distance which only the lk one is safe from set-up-able distances.

  9. Maj Post author

    There’s an excellent post written by VirtuaFighterFour on SRK about how VF player Minami revolutionized the way VF4 was played by taking an overlooked move with mediocre frame data and showing its true potential. Really cool story, definitely worth reading.

    I just think that our brains process an insane amount of spatial information when playing fighting games, and it’s easy to fall into the trap of undervaluing that. When someone hands you a frame data guide, it’s easy to say, “Oh i can get all that information by skimming through this chart” but you really, really can’t.

    So that’s why i wrote this article. I noticed too many people giving frame data too much clout over their gameplay decisions, when we all know how misleading those numbers can be. There are countless examples of frame data essentially lying about which move is best poke, or the best punisher, or the safest panic button. There are simply too many hitbox-related properties not covered by frame data, so there’s no substitute for actually playing the damn game and developing an intuitive grasp of your character and the combat engine in general.

  10. zero

    published frame data can not describe the game fully. it leaves most mechanics untold. source code can tell almost everything. but that’s generally inaccessible. emulators provide a window to look into the inner part of games. especially for cps1 games. it is not too difficult to trace CE using mame debugger. but for new complicated games like cvs and sf4, trial and error could be a better way.

  11. Icege

    @Maj
    On the contrary, spacing plays a vital role in 3D fighters. Whiff punishment is essential, and there are several moves that push opponents back (some are punishable, but depending on circumstances, you can make them safe). Sidewalking also plays a huge role, and while it isn’t the face-to-face spacing typical of 2D, can be thought of somewhat as a “footsies” tool of sorts. Korean Backdashing also allows you to create situations where you can create space and bait whiffs.

  12. Maj Post author

    Icege: I’d rather not get sucked into that old 2D vs 3D debate, because honestly i don’t really care about 3D mechanics. All of that stuff exists in 2D games, with the exception of sidestepping.

    The way 3D games are designed, it’s basically like a zoomed-in microcosm of the 2D playing field. Jumping basically doesn’t exist because the number one reason for jumping in 2D games is projectiles which don’t exist in 3D games.

    Also pushback is so rare in 3D games that it’s almost treated like a special move property. Anyway my point was that there’s a lot more “i block that, i punish with this” in 3D games, so frame data is more necessary and more useful at face value. I’m sure you could come up with even more reasons for why that’s the case than i could.

  13. Kareeem

    Simply moving from 3D games (VF was my game) to 2D games taught me to not rely too much of theoretical data. I still believe in eventually breaking everything down to numbers and hitboxes to understand the game (but thats simply the way my mind works). But I’ve had to learn the hard way to respect tactics that ‘shouldn’t work’ on paper but definitely do in matches.

    Not sure if this is the same thing you were trying to convey but realizing that made me break away from my theory fighter antics. Good post Maj.

  14. Kareeem

    Oh and Maj, it’s virtua, no L. I hear the most OG of players refer to it as VirtualFighter, still don’t know why lol.

  15. Maj Post author

    Oh sorry, typo. Fixed it, since it was someone’s username and i don’t want to be disrespectful. I think most people know it’s Virtua and not virtual, but “virtua” isn’t a word so unless you’re a VF player, it’s unnatural to stop before the “L” when typing fast. I basically have to delete it every time and sometimes forget.

  16. Kareeem

    Ah, virtua not being a word completely eluded me all these years (english isn’t my first language), that explains why so many people call it virtualfighter. Another mystery solved.

  17. Icege

    @Maj

    What I said has absolutely nothing to do with the 2D vs 3D debate. It has to deal with the inaccuracy of the statement you made that there was no range game in 3D games. There is, very much so. While it isn’t controlled with fireballs and meaty hit boxes, there is very much an area that a character can control (ie: Mishima’s mid-range with EWGF).

    Jumping exists also, though it performs a different function. While in Street Fighter, it is utilized to set-up cross-ups after a knockdown or avoid a fireball, it also serves as a reason to make people block high. 3D has mid attacks that serve that function. Also, jumping attacks in 3D will avoid lows. While they might not serve the same exact purpose (or function the same exact way) as 2D, the fact of the matter is that they are important and they do exist.

    You are correct though in that frames do play a bit more of a vital role, though the “block and punish” mechanic is still very much alive in Street Fighter. Being able to consistently punish Blanka Ball, for instance, is quite important to the match up.

  18. Maj Post author

    I’m sorry sir, but you’re reaching. When i said range game, i obviously meant far range. Of course there’s a mid-range game in Tekken, but it doesn’t go beyond that.

    And no, the 2D definition of jumping doesn’t exist in 3D games. What you’re describing is a manual hopkick.

    “Block and punish” definitely exists in 2D games because like i said, almost every major 3D game concept exists in 2D games too, but it’s less prevalent. Blanka ball isn’t a good example though because it’s got nothing to do with frame advantage. Did you know that move is -15 on hit and -24 on block? The only reason it’s good is because he bounces so far away. In fact if you put Blanka ball (with block damage) in Tekken or VF, half the cast would lose 8:2 to that move alone because there’s no range game. Especially on a stage with no walls.

    A real example of “block and punish” in 2D games is SF4 Gen’s KKK:c.LK which is -8 on block but sets up a free juggle when it connects, so you can kinda use it in a throw mixup if you give the opponent a valid reason to stand up. Maybe using Gen’s KKK:s.MP overhead to confuse them or whatever. Unfortunately throws can be teched low in SF4 which contributes to why people call it a defensive game.

Leave a Reply