Thursday, March 5, 2015

Frame Advantage in Neutral

So I know I’ve been primarily talking about sheik, but I wanted to take some time to discuss a more general melee concept (fret not sheik players, I’ll talk more about her soon enough!). This is a concept that I think is fairly intuitive ultimately, yet I have never really seen it written nor talked about; after a recent trip to Orlando and throughout the course of a 7 hour car ride, I managed to formalize it a bit and decided to write about it…it being the concept of frame advantage, but in the context of neutral.

Let me take a step back, just in case the typical usage of frame advantage isn’t clear to anyone. Frame advantage traditionally is used in reference to a situation in which an opposing character either gets hit by, or blocks an attack. For example, if fox uptilts a shield his frame advantage is -11, which means that his opponent is actionable for 11 frames before he is. If peach does a float canceled aerial (well, except down air) at the lowest height possible, she is +4 on shield; again, this means that she can act 4 frames before the opponent can. In traditional fighters, frame advantage on hit is something that is talked about, though not so much in melee so I’ll skip that (though the concept is similar, shouldn’t be too hard to connect some dots).

Alright, so what does this have to do with the neutral game? Before I get to that, I’ll suggest that rather than viewing the goal as landing a hit on the opponent, we can view the game in terms that can be better generalized: the objective is to gain a significant enough frame advantage that the risk/reward in any given situation is skewed in one’s favor. An example of applying this very general (and I believe, more useful) definition to a specific situation is the case of baiting out an extremely laggy move…say, a marth forward smash. If marth’s forward smash can be successfully baited, then you have maneuvered the situation into one where you have a significant enough frame advantage that you have a guaranteed hit. This is a very simple example, and seemingly makes my redefinition of the game rather unnecessary. Once I discuss how this can be applied to the neutral game, I hope that it will seem more beneficial, even necessary to understand the game in this manner.

Let’s talk about falcon vs. sheik, as this is the situation I was originally thinking about when I formalized this concept; in particular, I want to talk about the falcon neutral air vs. sheik forward tilt interaction. Falcon’s nair is an amazing anti air and has great utility in that it protects him/his landing while also threatening the opponent on the ground, but if he wants to use this tool he needs to find ways to interact positively with sheik’s ftilt; this is because sheik primarily wants to stay grounded in the matchup, and ftilt is her best grounded anti air option that hits in front of her. Sheik’s ftilt hits on frame 5, but the part of the move that will realistically hit falcon isn’t until frame 6. Falcon’s nair, on the other hand, hits on frame 7; he has 4 frames of jump squat, so it effectively has a 10 frame startup (and hits on frame 11). Comparing the two leads to the conclusion that for falcon to nair without losing to ftilt, he needs to be “+6” in neutral (11-6 = 5, so with a 5 frame advantage nair would hit on the same frame that the relevant part of ftilt hits, and trade).

This frame advantage can be achieved through many means, and how much advantage you have determines whether or not it is a good idea to challenge the opponent in any given situation. Let’s say sheik whiffs an ftilt while falcon does not take any action that puts him in lag (either he already finished a wavedash/just landed from jumping while her ftilt is recovering, or perhaps he was just dashing) but he isn’t in range to whiff punish with grab. Falcon may not be able to directly punish the ftilt, but he has still gained some sort of frame advantage: he can act before sheik can. If he is able to position himself close enough to sheik after she whiffs the ftilt and gets there with at least 6 frames to spare, suddenly his nair actually *beats* the ftilt option!

Of course, nobody sees this game frame by frame, but the concept of “how much time” we have to execute an action in a situation is intuitive to all of us. Extrapolating this concept to the neutral game as a whole isn’t that difficult; every movement option chosen when players are “feeling each other out” is just a marginal risk taken to gauge what sort of advantage can be gained. Each player is trying to find a way to reposition relative to their opponent’s actions that gives them the highest chance of winning an ensuing interaction, and if it seems like the “timing” of the situation will not be advantageous, then there is an attempt to disengage and reestablish favorable positioning. If it is likely that an opponent will try to wavedash forward and that can be predicted, then that entirely changes the nature of potential interactions; even if this happens from very far away, a wavedash still gives the opponent ~14 frames of time to do whatever they want. Again, I’m not telling anyone that we can see the game frame by frame, but this is what many high level players intuitively do based on feel: they get a read on someone’s movement patterns, and although they might not think “this person will wavedash at this exact time”, there is still an instinctive action taken that allows them to gain an advantage while the other person repositions.

Honestly I wish I had more to say on the topic, but the specifics of it vary a great deal from matchup to matchup and interaction to interaction. What I really want everyone reading this to take away is a different way of thinking about how the game works. Each game state in melee can be fully described by variables of space and time: the positions of the two players, and their relative frame advantage. When thinking about and analyzing matchups it is important to think about both how interactions play out, given various amounts of frame advantage, as well as how effectively each character can influence the frame advantage from a neutral state.


  1. Amazing post.

    Kind of weird to flesh it out, because I do think that any 'competent' player really does think in this way. Which is a sort bash on myself because only recently have I realized to think in this way (as Falcon in the Fox MU I was getting out speed'd, and couldn't think of why until I realized that I was putting myself in frame-negative situations).

    I don't know how you could expand on this, but would love to see more from you on this :)

  2. Thanks :)

    Not sure how to expand more, but we'll see what I come up with in the future!