Smash science

[forward]

I recently made a post in a character discussion thread in which I referred smash as a science. Here's a piece of that post.

Virtually all the priorities in the game are known, or can be known, with 100% certainty. There is no randomness to this. If attack A and attack B are performed or collide at time T and at range R you will get the same result, every time (except phantom hits? I'm not sure if those can be produced or are actually coding glitches, which even then must have a reason for existing.). At percentage X with DI Y and hit by move A at range R your character will be hit in the same manner with the same stun every time. Though no one actually thinks about the game in this manner, they actually do on a subconscious level. We are always searching for the range and time to use the right move to out prioritize the opponent, and we are aware of the variables that effect the outcome of the move.

So I was wondering what other people's opinion on this matter is. Can we turn smash into a science? Can we express this game in mathematical formulas?

One thing I've wanted to do but don't have the knowledge base for is statistical analysis. I think this could be used to understand metagame and people's decision making process.

 

[Pi]

fairly certain ones brain would explode trying to process a real time match this way

 

[Seikend]

I recently made a post in a character discussion thread in which I referred smash as a science. Here's a piece of that post.



So I was wondering what other people's opinion on this matter is. Can we turn smash into a science? Can we express this game in mathematical formulas?

One thing I've wanted to do but don't have the knowledge base for is statistical analysis. I think this could be used to understand metagame and people's decision making process.

As far as I understand, phantom hits are consistent, it's when hitboxes are at a tangent with each other, i.e. just touch.

But as far as turning smash into a science I don't see it happening.

Take chess. Chess is a game with 100% certainty. Every turn a player is limited to a number of moves (limited number of pieces, where they can go etc.). To take a wild guess of the average number of possible moves each turn, let's just say 20. Your opponent takes their turn, with another 20 possibilities. From just one round of turns you're already at 400 possible outcomes. Chess has so many possible outcomes that to create a formula would be entirely unpractical.

In the context of Melee, the case only becomes more absurd. Chess has only 64 "positions" to be in. Melee has a myriad of different positions on the screen. In chess things can only occur on each players move, games average at around 40 moves. Melee has 60 frames per second, each game lasts for several minutes, and crucial moments last for far more than 40 "moves".

Like in chess, players can't work out every possibility. It's all about looking at the appropriate possibilities. Ruling out ineffective options (for your opponent and yourself) in your head, planning out your options, limiting your opponents options and reading what their intentions are.

 

[Dark Sonic]

Phantom hits occur when a hitbox is tangent with a hurtbox

Priority the way we think of it (disjointedness really) would be a ***** to calculate precisely because we lack a unit of measurement for distance.

Trajectory is easier to calculate mathematically, though I still think we're missing some key components.

1. We need a formula for how far a character is launched based on a certain launch power based on launch angle, character weight, and character fall speed (and this still requires a unit for distance for our result)
2. We need a formula for calculating launch power based on base knockback and knockback growth.
3. We need to know how much DI can affect the launch angle of an attack

I believe we already know #3 (well, Magus knows ) and we already have the information that we'd plug into numbers 1 and 2 (minus launch power obviously)



Also guys, use your heads. He's obviously not saying that players would be able to do these precise calculations during a match. He's talking about using such information for research purposes, gathering knowledge that may be applicable to situations during a match later. It's similar to using frame data to figure out if moves are fast enough to be punishers in certain scenarios.


It's an attempt to substitute the "just try out a lot of things and see if they work" approach by saying "the data shows that these things work when done correctly and have this margin of error," which will have you practicing with a clearer goal in mind. Though it's debatable whether one approach is better than the other (I personally prefer the second approach).

 

[omgwtfToph]

Phantom hits can be controlled, yes.

anyway, the highest and most practical level we can make of "turning the game into a science" is really more like "turning the game into a game of statistics." You have to account for human reaction time, so no, you can't win every encounter. But better players are better at noticing (subconsciously, for most players) how to approach/how to edgeguard/how to techchase a certain opponent. You can develop statistics for this sort of thing, sure. But it's going to vary from opponent to opponent.

 

[INSANE CARZY GUY]

The thing is nothing is broken other than movement camping where they can't caugh you esle everything can be countered in one way or another when you have the option to use that move or movement.

If I expect a f-tilt from sheik i could take it and CC to a grab or something I could jump away I could try to wait to beat it or use a move lie marth's f-smash to try and beat it, DD grab or whatever. but sheik grabs causing win or lose

 

[Metal Reeper]

Im sure Magus thinks like this in game and is accurate, and lol'd at this thread.

 

[Zodiac]

Forward, this is exactly how I think about the game and try to play it that way always. for instance,I know the exact position I have to be in to sweetspot with all my characters. This should be common knowledge for all smashers but it is part of smash science. You press up B while being at position A and sweetspot the ledge to where you cannot be hit down past the ledge from most attacks, and if you would like to go so far afte rthat as to add the ledgetech then go for it. You press L at time T and that covers any ledge teching you are able to do, In short it looks something like this.

Falling to position A, Take into account the time it takes to press up b and subtract it from the time it takes to get to position A.(This is essentially the real position A) You press at that time and then I pick a frame to look for do press R at T and hit it. I flushed it out a bit more as usually I only think about this kind of stuff inside my own head. But when I follow that I can confidently say I never press my buttons at a different point in time during these operations.

 

[t3h Icy]

Go here for what's possible when you can play with the math and science of video games.

 

[SSBMLahti]

Smash is a lifestyle, not a science

 

[FoxLisk]

\
One thing I've wanted to do but don't have the knowledge base for is statistical analysis. I think this could be used to understand metagame and people's decision making process.

statistical analysis of what?

Trajectory is easier to calculate mathematically, though I still think we're missing some key components.

1. We need a formula for how far a character is launched based on a certain launch power based on launch angle, character weight, and character fall speed (and this still requires a unit for distance for our result)
2. We need a formula for calculating launch power based on base knockback and knockback growth.
3. We need to know how much DI can affect the launch angle of an attack

I believe we already know #3 (well, Magus knows ) and we already have the information that we'd plug into numbers 1 and 2 (minus launch power obviously)

has anyone ever just run some basic tests to see if DI follows some relatively simple vector operations? it looks to me like it's just vector addition and some scaling, but no one has ever checked this (afaik)

 

[vZakat]

I think the phrase "smash as a science" doesn't totally fit. I thing it's more of analyzing smash scientifically. It seems like for the most part when we look at smash, we analyze it philosophically and psychologically (i.e.mindgames). Crunching numbers is an important part of smash too.

Edit - Also regarding DarkSonic's post, Doesn't the game already have some sort of distance in it? Is there some reason why we couldn't just use the same measurements for the distance traveled and flight distance as they do?

 

[ChivalRuse]

Smash in its physical form can be condensed within mathematics and naturalized to an engine that operates upon an array of predefined rules and a set of functions that mechanizes based on parameters that are limited and predictable. But the relationship between humans and the game isn't succinct at all. When we play the game, we engage in a constant decision making process. We use our past knowledge of the game, or our understanding of the metagame, in this process. Tactics, spacing, and tech skill can all be determined formulaically. Mindgames are far too abstract, however. They depend on way too many circumstantial factors. Our abilities to perceive these factors and react to them accordingly is variant, and such an ability can't be formalized in scientific terms.

So I think we can conceptualize a pseudo-science out of the tangible and unwavering components of Smash, but its philosophy can't be tied down in a textbook, gamer's manual format.

 

[Winston]

Can we turn smash into a science? Can we express this game in mathematical formulas?

I think the short answer is yes.

It depends on what you mean though. It's certainly possible to describe frame perfect play mathematically, but that's not very interesting. It would also be viable to create mathematical models that describe how smash works fairly accurately. I don't think it would be viable to solve smash exactly for human play though, since it's too hard to quantify human reaction time, the limits of human technical precision, and how fast you process visual cues.

One thing I've wanted to do but don't have the knowledge base for is statistical analysis. I think this could be used to understand metagame and people's decision making process.

Statistical analysis of common responses in certain situations would be doable. If you had a ton of data on one player, you could figure out how frequently they choose certain options in similar situations, or what factors cause them to respond in certain ways.

I think it'd just be really tedious, and not very useful since most people have different sets of habits.

Mindgames are far too abstract, however. They depend on way too many circumstantial factors. Our abilities to perceive these factors and react to them accordingly is variant, and such an ability can't be formalized in scientific terms.

So I think we can conceptualize a pseudo-science out of the tangible and unwavering components of Smash, but its philosophy can't be tied down in a textbook, gamer's manual format.

I don't think this is completely accurate. Most situations where "mindgames" apply are either based on conditioning responses, or are situations that are described well by a payoff matrix.

The major aspect of "mindgames" that can't really be described mathematically is something similar to bluffing in poker or something. Reading or influencing your opponent based on "body language", rather than the situation and the options available/options they usually choose.

 

[Violence]

Indeed, I think we can easily develop formulas for DI, SDI, launch trajectory, launch distance, percentage increase, stale move decay, crouch cancel effectiveness, lag time, etc.

The most important thing we lack though, is a unit of measurement of distance, and a method of measuring that distance accurately.

 

[forward]

I was also aware that we don't have a unit of measurement for distance. I thought that we could perhaps find the smallest distance known, perhaps the circumference or radius of the circular hitboxes, and go from there.

I have no experience with programming though, so how people find these things and apply them are beyond me.

The statistical analysis would be, as Gustav said, used to find common responses in certain situations. Although I agree it would be tedious to find, I disagree that it would not be useful. Knowing how your opponent responds is a huge factor in player skill, especially once you get to the level of technical consistency. I've always believed that good players break down their opponent play because they pick up on habits, and they pick up on the most obscure habits that you may not think matters. How you run and wavedash gives me information on how you attack. How you attack gives me information on how you evade. How you evade gives me information on how you counter, and so on.

In my mind, I'm thinking something like this. "He just did X, so I know that means he is likely to do Y or Z. If he does Y then he will do V or U, and if he does Z then he will do M or N." Usually, I can break down every decision as good or bad, winning or losing, and from then I know how well the opponent knows how to win in other situations. For example, if a person doesn't know how to sweet spot, they don't know how to win in this situation. Sweet spotting is a basic skill necessary for efficient and high level play, if they cannot do this basic aspect of high level play then it is likely that they cannot L cancel consistently, punish rolls, follow up on combos, etc.

I may be wrong, but I believe this has been done to create artificial intelligence in, at least, Virtua Fighter 4. They programmed their AI to play based off of the best players in Japan. If this were true, how did they do that? What information was needed from high level players?

If we could program a high level AI in smash, what would we tell it to do?

 

[tarheeljks]

almost anything can be quantified to some degree; aspects of melee have been quantified in various ways (e.g. frame data, weight, running speed etc). that said i don't believe anyone will ever be able to describe/solve/whatever this game (or most any game really) primarily w/#'s, so it will always be more of an art than a science imo


edit:

One thing I've wanted to do but don't have the knowledge base for is statistical analysis. I think this could be used to understand metagame and people's decision making process.

this is certainly possible. statistical analysts have begun to pop up on payrolls in major sporting leagues (has been this way for a while in baseball and it has spread to the nba). you can conduct statistical analysis of any process provided that there is enough data and there certainly is in smash given the amount of video content available. it would not be a small feat though

 

[FoxLisk]

but you either get crap answers or require HUGE data sets. i mean, what kiond of questions are you going to ask? When a player is shielding at the edge, how often does he roll in?

The answer to that is going to be "usually, if that player is bad and playing from a certain set of characters, unless he has developed another habit in that situation or his training partners have caught on to it, and it depends what the other player is doing." You can go ahead and get a number, like 63% of the time they roll in within 60 frames or smoething, but that's not really useful information. it's not going to be easy to quantify the kind of answers you actually want and the factors going into them (e.g. player skill, etc), and even if you can do that you'd need on the order of a hundred instances of the situation being set up exactly the same. unfortunately, that basically means you'd need many players of almost identical skill playing the same character in the same matchup on the same stage to determine anything.


edit: additionally complicated: melee isn't like chess, where the beginning of a game affects the end. so you can't even say, 'rolling in at the edge tends to lead to lost games,' because a player could easily roll in from the edge the first two times - once to get away, a second time to condition his opponent - and then stop for the rest of the match and win. obviously that action itself wasn't what lead to the win, it was his manipulation of his opponent, which is going to be impossible to quantify or study statistically.

 

[tarheeljks]

but you either get crap answers or require HUGE data sets.

hence me saying it wouldn't be a small feat. he asked if it was possible, the answer is yes to the extent that you are trying to identify trends and not do something like "solve the game."* whether it is a worthwhile endeavor is another issue. i'd say probably not as it requires a significant amount of legwork and the payoff isn't nearly large enough to take on such a project on a large scale imo. it might be something worth attempting on small scale-- for example it could provide a great deal of insight to matchups

unfortunately, that basically means you'd need many players of almost identical skill playing the same character in the same matchup on the same stage to determine anything.

that said, i disagree with this. that is the ideal scenario for obvious reasons, but that doesn't mean you can't obtain meaningful information otherwise. i'll point to sports again-- people get paid to try to "quantify" sports like baseball, basketball, and football. are they able to obtain formulas that encapsulate these games perfectly? of course not. there are too many variables to isolate and conditions are too dynamic ( and more dynamic than in a video game); however, the analysis has still yielded results that have advanced the way these sports are played/experienced/analyzed. you don't have to be able to breakdown a process into a neat formula in order for the results to be meaningful. however, there is clear incentive for these resources to be used in the sporting world as it generates massive amounts of revenue. smash, not so much

where the beginning of a game affects the end.

i don't see how it's possible for the beginning of the game not to affect the end. you're basically saying that actions performed within the game are independent, i.e. players have no recollection of moves they/their opponent performed and that they didn't adjust. i guess some players do play this way where they just sort of randomly throw out moves and if they work great, if not oh well, but i think we can agree that that's not the best way to play



cliffs: you could conduct a large scale melee stats project and reasonably expect to obtain useful results (provided it was a well designed study of course) but it's highly likely not worth the time/effort



*meaning. is this possible

Can we turn smash into a science? Can we express this game in mathematical formulas?

nah man

is this

One thing I've wanted to do but don't have the knowledge base for is statistical analysis. I think this could be used to understand metagame and people's decision making process.

sure

 

[FoxLisk]

that said, i disagree with this. that is the ideal scenario for obvious reasons, but that doesn't mean you can't obtain meaningful information otherwise. i'll point to sports again-- people get paid to try to "quantify" sports like baseball, basketball, and football. are they able to obtain formulas that encapsulate these games perfectly? of course not. there are too many variables to isolate and conditions are too dynamic ( and more dynamic than in a video game)

also there's a huge pool of professional athletes in real sports. we have, like, what, 20 people at the top of the metagame? you'd end up with a **** ton of stuff that works against mid-level players and almost no information about what matters to the top players.

i don't see how it's possible for the beginning of the game not to affect the end. you're basically saying that actions performed within the game are independent, i.e. players have no recollection of moves they/their opponent performed and that they didn't adjust. i guess some players do play this way where they just sort of randomly throw out moves and if they work great, if not oh well, but i think we can agree that that's not the best way to play

that is not at all what i said, i dont think you read my post very well.

what i mean is this: in chess, for example, once i trade off a few pawns, I can never get those pawns back. In melee, what I do on my first stock does not at all affect my options on the next stock. it has only a psychological effect on how my opponent and I play. I can play exactly the same throughout a game of smash in a way that I cannot in chess.

Or in Starcraft, once I harvest a mineral patch, it's gone forever. When I deplete my main, I have to expand, which requires a change in play. Choices in the beginning distinctly change the options you have open to you later in the game.

that's part of what makes a fighting game like Melee so hard to study. In starcraft you can say, "when playing terran vs zerg, the terran wins 55% of the games he opens with 1-rax FE, but only 49% of the games he opens with BBS, so 1-rax FE is statistically superior." that kind of analysis doesn't apply to smash.

 

[Bacon Isai]

I like to think of smash as an art.

 

[tarheeljks]

also there's a huge pool of professional athletes in real sports. we have, like, what, 20 people at the top of the metagame? you'd end up with a **** ton of stuff that works against mid-level players and almost no information about what matters to the top players.

when they run stats in sports they don't just do games where a couple of top 10 nba players are playing against each other. taking info from national tournaments or strong regionals would produce a pretty large data set given the amount of videos online

that is not at all what i said, i dont think you read my post very well.

what i mean is this: in chess, for example, once i trade off a few pawns, I can never get those pawns back. In melee, what I do on my first stock does not at all affect my options on the next stock. it has only a psychological effect on how my opponent and I play. I can play exactly the same throughout a game of smash in a way that I cannot in chess.

it doesn't literally change the available options, no but that is trivially true. if it affects the way you think/play then it affects your options for all intents and purposes. it's kind of pointless to say "you're options don't change" when people actively strive to adapt on the fly and read opponents.

Or in Starcraft, once I harvest a mineral patch, it's gone forever. When I deplete my main, I have to expand, which requires a change in play. Choices in the beginning distinctly change the options you have open to you later in the game.

that's part of what makes a fighting game like Melee so hard to study. In starcraft you can say, "when playing terran vs zerg, the terran wins 55% of the games he opens with 1-rax FE, but only 49% of the games he opens with BBS, so 1-rax FE is statistically superior." that kind of analysis doesn't apply to smash.

and i didn't say anywhere that you could do this. in fact i explicitly said that you probably cannot break smash down into a formula and that it is more of an art than a science. however, your view of what constitutes useful results with regards to statistical analysis is too narrow. being able to say "action x results in victory y % of the time" would interesting, but that is not the only benefit that can be derived.



edit: this notion of "smash is complex" and consequently can't be analyzed doesn't resonate with me. there are FAR more complex systems that have benefitted from quantitative analysis but again it is an issue of whether it would be productive as much as whether it is feasibe


edit2: also wrt the huge pool of professional athletes-- i'll give you that it is easier to get your hands on stats from sporting events, but sample size is an even larger issue for them b/c circumstances are even more variable. e.g. lineups in basketball-- the 5 people playing together changes often, and the competition changes even more often as you only play the same team a small # of times. beyond that the same 5 people will interact in a variety of ways, so despite having all this data of them playing together, you still have a small # of trials of a particular event (which is one of the limitations of statistical analysis in sports. one issue in the nba is that you need like 4 seasons worth of data to cut the error in half). smash, esp when it's 1v1 offers a large # of trials, and short trials at that, w/circumstances that are very similar even if relative levels of competition aren't completely static. the nature of smash is more conducive to this sort of analysis than sports; fewer moving parts

 

[Rappster]

I like to think of smash as an art.

the difference between art and science is that you don't need background knowledge to create beauty in art.

 

[FoxLisk]

gah you dont understand what im saying or what sorts of things statistical analysis can be used on, conversation over.

 

[Pengie]

This sounds like an amazing idea. This would involve delving into all of the values for different moves (i.e. knockback, how much it is affected by DI -if DI does even change from move to move-, how strong DI is, etc). I do have some ideas on how to go about putting together certain "formulas" for smash, but it'd be necessary to have all of the values to make sure these ideas are actually accurate. The biggest one that sticks out to me is how DI seems to be an additional force that is being applied to your character. With that knowledge, it seems possible to mathematically find out the most optimal way to DI through adding together force vectors. If anyone who is well versed in Physics/Calculus would like to chime in on this it would be greatly appreciated

 

[Dark Sonic]

DI doesn't change from move to move.

DI can affect the angle of a move by a maximum of 18 degrees (thanks Magus)

The optimal survival DI would be the DI that changes the trajectory of the move so that you are sent towards the upper corner of the screen (as blastzones in melee are rectangular, this would give you the greatest possible distance to travel)

Well, it's slightly more complicated due to the constant force of gravity making your character travel in an arc rather than a line, but for in game applications this information is more than enough. Going any further would pretty much be strictly to satisfy curiosity.

 

[tarheeljks]

gah you dont understand what im saying or what sorts of things statistical analysis can be used on, conversation over.

i have a background in stats/math, deal with them everyday, and i'm comfortable w/everything i've said so far. would like to hear why you disagree w/anything i've said outside of "it's hard to do" i haven't even disagreed with that; analyzing anything statistically is "hard"


edit: do you see that based on your description of the circumstances necessary to analyze situations statistically, it is impossible to analyze things statistically?

edit2: and yet people do it everyday in systems that are considerably more complicated than smash (easiest example is the stock market). smash is not unique in terms of its complexity. practically every worthwhile statistical study has examined a dynamic, complex process. what would the point be if it were easy to see what was happening?

also, i meant to say that your assertion that earlier portions of the game do not affect the end would make analyzing outcomes easier not harder

 

[Pengie]

All right, so after screwing around in GIMP for a bit I came up with a graphical representation of how I think knockback works given the information that I know.

This graph gives the basic force vector of a move using the moves knockback as the vectors magnitude and the angle relative to the character as theta. With this it can then be broken down into its x and y components so that it can then be added to...

this, to account for DI. The value of the DI would be the magnitude and the angle relative to the character hitting them would be theta. Once again it can now be broken down into x and y components and the two vectors can be added to create the new vector, which would give a rough estimation of what direction you would go in.

Please tell me if there is something that I am overlooking in this, as I may be oversimplifying how DI works.


In all honesty, this seems somewhat useless though. Most of the "science" would just come to conclusions that most people already know. That's one of the wonderful things about Melee; every match is like a big experiment and the observable outcomes are there for everyone to see. After you have enough experience these things seem to become second nature, and the "formulas" and "calculations" would be more or less useless. I do, however, see quite a bit to gain from statistical analysis. It would probably be a pretty extensive project, but it could help move the community forwards in terms of picking up on/breaking bad habits more easily.

 

[jugfingers]

As far as I understand, phantom hits are consistent, it's when hitboxes are at a tangent with each other, i.e. just touch.

But as far as turning smash into a science I don't see it happening.

Take chess. Chess is a game with 100% certainty. Every turn a player is limited to a number of moves (limited number of pieces, where they can go etc.). To take a wild guess of the average number of possible moves each turn, let's just say 20. Your opponent takes their turn, with another 20 possibilities. From just one round of turns you're already at 400 possible outcomes. Chess has so many possible outcomes that to create a formula would be entirely unpractical.

In the context of Melee, the case only becomes more absurd. Chess has only 64 "positions" to be in. Melee has a myriad of different positions on the screen. In chess things can only occur on each players move, games average at around 40 moves. Melee has 60 frames per second, each game lasts for several minutes, and crucial moments last for far more than 40 "moves".

Like in chess, players can't work out every possibility. It's all about looking at the appropriate possibilities. Ruling out ineffective options (for your opponent and yourself) in your head, planning out your options, limiting your opponents options and reading what their intentions are.

actually the estimated amount of game-tree complexity in chess is at least 10^123

to put that in perspective the total estimated number of atoms in the known universe is 10^81

Im not really sure what this thread is about though...

 

[Esper]

Turning Smash into a Science would be even more interesting if we found a way to alter/recode the CPU-AI. For now it's more practical to have statistics about moves which cover most/all options in certain situations (depending on position/matchup/percentages).

 

[BigD!!!]

i think that smash can be scientifically analyzed

ive considered that, in recent years, as people understood more about the game and m2k displayed how well he could do with his style, people moved more towards treating this game as a science whereas before it was more of an art

you can still see some players, shiz comes to mind as the most successful player still embodying this in my opinion, who plays in a statistically unsound way that seems mainly based simply on feel, ie as an art. its like hes dancing all over them

dr peepee, however, plays way more on risk vs reward analysis and makes safe, logical choices. this seems to me to represent a more scientific approach to the game, maximizing odds of success and whatnot

 

[Signia]

I think the most interesting application of "science" as you guys are using the word is the use of game theory, which is usually used in business or war to (omg) videogames. The concept of mixed strategies in nash equilibrium allow you find the optimal strategy, which involves a specific degree of variation of certain actions or moves in a way such that no matter what your opponent does, he/she cannot improve the "odds," even if they know their strategy. This would be something like drill-shining 90% of the time, and drill-grabbing 10% of the time.

Put simply, you keep your solid options common and your surprise options rare. Exactly how often you do each thing is determined by expected value calculations. A problem with this though, is quantifying value. Damage is easy enough to quantify. But how valuable is knockback in comparision? That could be estimated, but what about positioning? Even if all these could be assigned accurate values, only a computer could get close to being able to execute a mixed strategy, because it involves varying your choices randomly by specific ratios.

It would be much easier in other genres and simpler FGs to come up with useful calculations. Check out these RTS mathematics: http://finalascension.org/viewtopic.php?p=329#p329

I could probably "solve" Soul Calibur IV convincingly if I studied more game theory but Smash has quite a few factors at work at any given time and while it would only require more calculations to predict some things like knockback or DI, to actually recognize a situation properly in smash is very difficult. You need to be able to tell which hitbox of the move you got hit with, how hard your shield is, how many times the move was used in the last 10 moves for the "stale" effect, how much shieldstun the move causes at that stale level, how much landing lag the move has, about how many frames between the hit and landing, and specific spacing of the character's hurtbox and grab hitbox in order to determine whether your shield grab will get them if they dash away at the first possible frame.

Oh and I'm just curious, do you guys even know what you're saying when you call something a science or art? Be more specific, "science" is commonly misused and no one can agree what "art" is.

 

[KirbyKaze]

We have a standard unit of measurement it's called a "Bowser Roll".

 

[Grim Tuesday]

I think we should get a list down of EVERY variable in Smash. I'll edit this post and add some later.

EDIT: I mean every variable that isn't related to the players. So, things like falling speed, frame data for attacks, etc... Rather than "how many frames have passed since they last used this attack in this situation".

 

[LLDL]

There are like. Thousands of variables. In melee lol. But that'd be a great addition to smashboards.

 

[Kanelol]

fairly certain ones brain would explode trying to process a real time match this way

This.

Can we turn an average smash game into a mathematical expression? Yes. Should we? Absolutely not never ever ever. Unless you're some kind of savant. that's so awesome they actually have that smiley face now

 

[AxelSlam]

so if I post in this thread will I be unable to see anything??

sorry to use this thread, but your threads are great forward.


anyway, this is brookman on a friend's account. I am posting to inquire the details regarding the reason for the melee forums being private on my own account. everything from melee discussion to the zelda character forum is marked as private. I can't even look at my WARNINGs.


If anyone can explain this to me, it would be appreciated. just dont do it in this thread, since I can't actually access it. (aka: PM me!)

 

[Grim Tuesday]

There are like. Thousands of variables. In melee lol. But that'd be a great addition to smashboards.

Well, I'm working on it now.
Once I have some substantial stuff, I'll post it.

 

[forward]

I think we should get a list down of EVERY variable in Smash. I'll edit this post and add some later.

EDIT: I mean every variable that isn't related to the players. So, things like falling speed, frame data for attacks, etc... Rather than "how many frames have passed since they last used this attack in this situation".

Well, falling speed, weight, frame data on attacks, those are constants, not variables. A 7 frame jab will always be 7 frames.

% is a variable, it changes when players get hit.

Position is a variable. It's also important to define position as it relates to something else. If your reference frame is your character then they are never moving. Position has to be distance between two objects. Maybe that's obvious though, I don't know. I emphasize it though because position in relation to the distance between two characters is very important and something that players are always consciously judging.

Is priority a variable? I'm not sure about this one. Priorities are constant, marth's uptilt will always out prioritize Fox's dair because a disjointed hitbox, but that priority's existence is dependent on other constants, such as frames for up tilt to come out. So, priority's existence is a formula dependent on constant such as frames and fall speed, but also based on the variable of position.

Just some thoughts...

 

[P.C. Jona]

idk why you guys are trying to make this game more complicated then it already is