Official Metaknight Discussion

[etecoon]

actually I've seen several video's along those lines concerning meta knight, especially concerning his stalling techniques

 

[fkacyan]

Thio: these have been addressed, read the thread.

I have been reading, and they've barely been addressed. At all. If you think I'm wrong, read through hundreds of pages and repost them for me.

The only two people to ever make decent headway for anti-ban have been OS and Crow!, and they only have made data posts in that regard.

 

[Kewkky]

Why is he better? What makes him better?

Before you state aspects of his character, broken ATs, etc....

Factor in that he has better people playing him, more people playing him, and more people dedicated to making him better / breaking him to ban him, and has for quite some time, and thus his metagame, techniques, and matchup knowledge by definition will be farther along than the rest of the cast.

That's your opinion, and as you can see a majority of the community (the whole of pro-ban) disagrees with it. Do you honestly believe that your argument wasn't brought up a year ago? Unless you're saying that MK's metagame didn't peak back then, we gave no chance for other characters' metagames to increase as fast as MK's, and that during the whole year that the argument was originally presented the incredibly vast majority of the Brawl community still can't beat MKs consistently EVEN MORE SO considering they've had 2 years to practice against every MK player around, is still something worth overlooking?

 

[etecoon]

The only two people to ever make decent headway for anti-ban have been OS and Crow!, and they only have made data posts in that regard.

as opposed to...no one on the anti ban side, lol.

 

[Tien2500]

Why is he better? What makes him better?

Before you state aspects of his character, broken ATs, etc....

Range, speed of attack, multiple jumps, priority, recovery, gimping.

Factor in that he has better people playing him,

No evidence to back this up. And we see the MK effect at all levels across the board. For this to be the case nearly every MK main would have to be nearly better than all Snake and Diddy mains. Why would this be at all likely?

more people playing him,

Crow's data controls for popularity.

and more people dedicated to making him better
/ breaking him to ban him, and has for quite some time
, and thus his metagame, techniques, and matchup knowledge by definition will be farther along than the rest of the cast.

Pretty much all of that is entirely speculation. The one thing we really know is that MK mains almost always have LESS matchup knowledge then their opponents due to the opponents fighting so many MKs.

I guess what I'm saying is that because you can't make a video like the ones above...

...there's a good chance that the character isn't broken and thus fits within the game and its cast.

How do you justify banning a character who is not broken?

I thought overcentralization was the standard. Now they have to be as broken as characters in (mostly) horribly programmed fighting games?

 

[Tommy_G]

Very slow aerial speed, Technically no priority, no projectile to force approaches.

getting gimped means youre recovering wrong...and you must recover...l2recover.

btw zss not being tournament viable? One of the top 2-3 players in florida is a zss main

 

[Omni]

I thought overcentralization was the standard. Now they have to be as broken as characters in (mostly) horribly programmed fighting games?

Street Fighter 2 and HD Remix aren't horribly programmed fighting games.

Can you name a fighting game where there is a clear best character that doesn't overcentralize?

 

[HeroMystic]

Very slow aerial speed, Technically no priority, no projectile to force approaches.

getting gimped means youre recovering wrong...and you must recover...l2recover.

btw zss not being tournament viable? One of the top 2-3 players in florida is a zss main

Technically no priority? If by that you mean MK can't clank with attacks then you'd be right, but how is that a disadvantage? That practically means MK can space till content.

Here is MK's only true exploitable disadvantage: He can't stop projectiles unless he shields or approaches with Tornado (which is usually seen as an opening).

 

[Omni]

Hero, what gives you the impression that tornado beats projectiles?

 

[Kewkky]

Very slow aerial speed, Technically no priority, no projectile to force approaches.

Congratulations. The only thing you mentioned that MK can't work around is "no projectile to force approaches".

Now, help me type this in: "pseudo-stalling tactics that are too nebulous to limit properly in a tourney's ruleset"... We can add this in with his pro's and cons, right?

getting gimped means youre recovering wrong...and you must recover...l2recover.

You know you're being obnoxious when you keep repeating the same argument over and over, when your argument literally has NO base at all. Are you telling me that there's a skill level we can reach with each viable character, where when we're forced offstage, we're at no risk of getting gimped? Because that is exactly what you're saying. If we get gimped we're recovering wrong? There are too many characters in the roster that have no choice but to get gimped, even more so when the MK opponent has knowledge over the mix-ups that the player can do. There is NO way you can back this statement up. Everyone gets gimped, even the top players do, and saying "l2recover" makes people want to disagree with you more and just ignore you outright.

Do you even understand why people are so frustrated when they argue against you? It's because you feel that everyone's such a stupid player that they can't even recover properly. That's how I see it whenever you re-state similar 'arguments': l2recover, learn the MU, don't get grabbed/hit... That's no argument.

btw zss not being tournament viable? One of the top 2-3 players in florida is a zss main

I'm a ZSS mainer, top player in PR. I may not have used my ZSS against the top-players in Pound4, but I can assure you that i know how she works, and I know her weaknesses... And I have to agree that ZSS leans more towards non-tourney viable than tourney-viable. Please, point out to a tourney outside of Florida where ZSS is placing well, better than Kirby even (bottom of the borderline-viables).

 

[Crow!]

Here we go.

Those of the anti-ban side smart enough to understand what is important in a debate but who don't happen to have the intuition for statistics to see that their question has already been answered, have been asking for quantitative proof that MK's dominance is because of the character himself. Thiocyanide has gone so far as to say that:

Where is the evidence that the MK mains simply aren't just better than the rest of the players? And where is the evidence that the switch would make the other players as good?

There are my fundamental issues with your data. You convince me of those two points, I'm essentially pro-ban.

When Thiocyanide (and people like him) are done reading this post, they either will be pro-ban or will show that they are liars.


Let's link to data from Omni (the de facto anti-ban spokesperson).
http://www.smashboards.com/showpost.php?p=9479597&postcount=2284
The key here is that, even though the numbers he emphasizes are worthless (for reasons I explain in the links below), he also (either accidentally or incidentally) provides some information that IS useful.
http://www.smashboards.com/showpost.php?p=9481002&postcount=2402
http://www.smashboards.com/showpost.php?p=9481331&postcount=2419


Omni identifies that, among the top players (defined by those who score at least a given amount of points), among those who play either MK or Snake, 13 of them are MKs and 2 of them are Snakes.

Let us suppose for a minute that MK is no better a character than Snake, i.e. that for a MK main to perform better than a Snake main (as is observed) he actually needs to be a better player.

This is our "null hypothesis." In Statistics, it is common practice to reject the null hypothesis if it would mean that the odds of the observed data occuring were less than 5% or so.

So, our question is the following: what are the odds that among the top 15 players who play those two characters, at least 13 just happen to be MKs?

This is the standard form of a binomial probability problem. The formula for exactly 13 is:
0.5^15 * (15 choose 13). Sum this over 13, 14, and 15 MKs.

Answer: 0.00369...
That's less than half a percent, it's one chance in about 271, it's more than 10 times more unlikely than the probability threshold usually used to quantitatively reject the null hypothesis that would require the observation to become so unlikely.

MK is a better character. That case is closed.

-------------

Thiocyanide's second question is, does the difference in the goodness of the character itself account for the amount by which MK is dominating? To answer this, let us consider the character's learning curve.. that is, their ability to perform in tournaments as a function of the player skill. Let us be generous to MK and say that the overall shape of the curve is the same as for Snake (it has been argued that MK is easy to play, i.e. his skill curve gets high earlier than, say, Snake's does.. this would only make the picture even more ridiculous, as the analysis below would then require that Snake would be being outperformed by as much as they are by MKs of lesser skill so I'll be nice and ignore that possibility.)

The learning curve is obviously a mononotically increasing function with skill, but aside from that we don't really know its shape. But looking at the tournament results of the players, we can figure some stuff out:
Link to data: http://www.smashboards.com/showpost.php?p=9585089&postcount=5511

The top couple players of each main are somewhere in the "really really skilled" area and it's not obvious how well you can compare one to the next, although it does appear that MK is better.

The next bunch of players are interesting though: there's a plateau. What does this mean? This is what this means, in the BEST case scenario for anti-ban:

As we scatter into the game Snake and MK mains of various skill levels, there is a certain level of tournament performance a lot of them fell into. This means the skill levels of these plateaus are comparable!
"But wait," you might say, "what if we're actually comparing two different plateaus?" No, what's what we proved ISN'T the case in the previous section of this post.

The answer to the exact question Thiocyanide asked, then, is yes, given the Snake mains' skill levels, if they were instead MKs, they could perform as well as MK is. Or possibly better (at, say, M2K level) if the "MK is easy" theory is right.

So, how much better does MK do at comparable skill levels? On average, within this plateau, MK mains are scoring 2.54 times as many points as Snake mains.
The overall outperformance we're trying to account for is, from Ankoku, 1.99 times as many points.

So, the amount by which MK, the character, is better than Snake is MORE THAN ENOUGH to explain MK's tournament performance.
That is, the data suggest that MK is, overall, an unusually UNPOPULAR a character given how good he is!

 

[MarKO X]

Street Fighter 2 and HD Remix aren't horribly programmed fighting games.

i think he was referring to the broken characters themselves, but maybe that's just speculation.

Can you name a fighting game where there is a clear best character that doesn't overcentralize?

SUPER SMAAAAAAASH BROTHERS!!!
although that's arguable because people didn't really go in with Pikachu like that and then there was Isai.

There's also American SF4. lolol.

Hero, what gives you the impression that tornado beats projectiles?

try it. i'm sure it beats some of the smaller projectiles, but it's unwieldy because of its cooldown.

and lol @ another Crow post.

 

[Tien2500]

Street Fighter 2 and HD Remix aren't horribly programmed fighting games.

Can you name a fighting game where there is a clear best character that doesn't overcentralize?

Thats why I said mostly. Street Fighter is unrelated but at least is an established competitive game. But bringing in Power Ranger cash ins is pretty silly.

I'm really not an expert on other fighting games so I'm not going to comment on them.

 

[HeroMystic]

Hero, what gives you the impression that tornado beats projectiles?

In some, maybe most cases Tornado does beat projectiles. Not grounded, but aerial. It doesn't beat Falco's lasers or Snake's grenades, but it does beat projectiles because Aerial tornado does clank with other attacks and doesn't cancel like grounded tornado does.

 

[DoonKoon]

@Budget Player Cadjet_

What do you mean when you say, "We have to work a lot harder"?

Yeah, obviously, every other character has to work a lot harder than Metaknight, but who does this include exactly? S to A tier characters only? And what about the character who have to work harder than them? A Ganon will have to work at least 3 times harder than a Snake to win a tournament, and a Jigglypuff will have to work harder than a Wario. (Jigglypuff & Ganon aren't good examples since they're non viable anyways, but I'm just making a point.)

Still waiting Budget <___<

 

[Chuee]

There's PLENTY of projectiles that beat out nado.

 

[DoonKoon]

There's PLENTY of projectiles that beat out nado.

That doesn't stop it from being broken.

 

[Tien2500]

Crow Post.

Win. 10 wins.

 

[Omni]

@Crow: Why did you limit your data to only MK and Snake in that example?

@Hero: MK's tornado loses to Falco's lasers, Snake's up-smash, grenade, and C4, IC's blizzard, and Diddy's bananas pretty hardcore.

@DoonKun: The fact that tornado gets beaten by an array of moves does mean it is not broken.

@Tien: Don't quote long posts like that. Mic has already given people warnings for that.

 

[MarKO X]

@Crow: Why did you limit your data to only MK and Snake in that example?

probably because those are the top 2 characters with the top 2 amount of wins... of course, more speculation.

@Hero: MK's tornado loses to Falco's lasers, Snake's up-smash, grenade, and C4, IC's blizzard, and Diddy's bananas pretty hardcore.

well there u go. nado loses to the projectiles that matter.

edit: c4 is a projectile? hmmmm...
(talking about the actual definition, not how it works in terms of brawl itself)

edit 2: ok, so its not a projectile per se, but neither is Sonic's spring. *nods* continue.

 

[Humpy Thrashabout]

I think that people saying that anti-ban doesn't have to make any counter-arguments have to realize that the people far on the either side of the fence are NEVER going to agree. At no point will Omni agree MK is broken, and at no point will Overswarm agree he isn't.

The only thing this debate is trying to do is convince OTHERS to think one way or another. If anti-ban stays as relatively quiet as they are now they wont do too well if it were to ever come down to a vote.

 

[Chuee]

That doesn't stop it from being broken.

If that doesn't then the shield button does.

 

[Crow!]

@Crow: Why did you limit your data to only MK and Snake in that example?

I was being nice to you. But if you don't want me to be that nice, I don't have to be.

Let's suppose that MK, Snake, AND Diddy Kong are all of equal skill level. Then the formula becomes:
(1/3)^13 * (2/3)^(18-13) * (18 choose 13). Sum this over 13 through 18 MKs.

The probability of at least 13 MKs out of 18 possible out of 3 characters is 0.000853.

That's less than a tenth of a percent, that's one chance in 1173, that's absolutely absurd to suggest.

As you add more characters, it becomes even more stupid.


Anything else for which you'd like me to abandon my "Assume the best possible case for the anti-ban argument" policy?

 

[GunmasterLombardi]

@Crow: Why did you limit your data to only MK and Snake in that example?

@Hero: MK's tornado loses to Falco's lasers, Snake's up-smash, grenade, and C4, IC's blizzard, and Diddy's bananas pretty hardcore.

@DoonKun: The fact that tornado gets beaten by an array of moves does mean it is not broken.

@Tien: Don't quote long posts like that. Mic has already given people warnings for that.

See there's this thing called "predictability".

Figure it out.

 

[rvkevin]

Can you name a fighting game where there is a clear best character that doesn't overcentralize?

What is your criteria for over centralization?

 

[Budget Player Cadet_]

Still waiting Budget <___<

You missed the post where I said "**** I don't know how to counter this argument" and passed it on to someone else.

 

[Delvro]

This is our "null hypothesis."

So, our question is the following: what are the odds that among the top 15 players who play those two characters, at least 13 just happen to be MKs?

This is the standard form of a binomial probability problem. The formula for exactly 13 is:
0.5^15 * (15 choose 13). Sum this over 13, 14, and 15 MKs.

Answer: 0.00369...

Actually, the probability is even lower than that, because as you've proven below, the playerbase for MK is smaller than Snake as shown here... thus the sampling distribution will have a mean proportion of top 15s shifted in favor of snakes rather than half of each (under your theoretical assumption that the characters are equal)

So, how much better does MK do at comparable skill levels? On average, within this plateau, MK mains are scoring 2.54 times as many points as Snake mains.
The overall outperformance we're trying to account for is, from Ankoku, 1.99 times as many points.

That is, the data suggest that MK is, overall, an unusually UNPOPULAR a character given how good he is!

Nitpicky I know, but it drives the point in that null hypothesis is rejected even at the 1% level of significance.

 

[HeroMystic]

@Hero: MK's tornado loses to Falco's lasers, Snake's up-smash, grenade, and C4, IC's blizzard, and Diddy's bananas pretty hardcore.

You're only strengthening my point. <.<

 

[Espy Rose]

Wow @ Crow post.

Get ***** everyone.

 

[Omni]

I was being nice to you. But if you don't want me to be that nice, I don't have to be.

Let's suppose that MK, Snake, AND Diddy Kong are all of equal skill level. Then the formula becomes:
(1/3)^13 * (2/3)^(18-13) * (18 choose 13). Sum this over 13 through 18 MKs.

The probability of 13 MKs out of 18 possible out of 3 characters is 0.000853.

That's less than a tenth of a percent, that's one chance in 1173, that's absolutely absurd to suggest.

As you add more characters, if becomes even more stupid.


Anything else for which you'd like me to abandon my "Assume the best possible case for the anti-ban argument" policy?

Keep it civil, Crow. You don't have to be a douche bag to make a point.

Isn't the data still supported by tournament results? Also, what is the earliest tournament results that are used to reflect information from this data?

The reason why I ask is because a metagame grows. What was a trend or results a year ago is not always the same in the present. Also, point contribution deals a lot with repetition of tournament attendance. A person who consistently places 5th in tournaments but goes to a tournament every weekend will have more points than a player who places 2nd consistently but goes to a tournament once a month. The person who consistently places 2nd is still the better player. The results become a bit skewed in this manner.

This isn't to say the data is invalid because certain players using certain characters do not show up on a consistent basis; that would be silly but it is a point to note.

*fixed*

Can you answer those questions? I'm not really good when it comes to statistics so most of my questions will be an attempt to gain clarity on the information being presented.

Edit: Error on my part. More than 7 exist in the recent metagame. What was the population sample used in your chart above, Crow?

 

[Mortimer]

I was being nice to you. But if you don't want me to be that nice, I don't have to be.

Let's suppose that MK, Snake, AND Diddy Kong are all of equal skill level. Then the formula becomes:
(1/3)^13 * (2/3)^(18-13) * (18 choose 13). Sum this over 13 through 18 MKs.

The probability of at least 13 MKs out of 18 possible out of 3 characters is 0.000853.

That's less than a tenth of a percent, that's one chance in 1173, that's absolutely absurd to suggest.

As you add more characters, it becomes even more stupid.


Anything else for which you'd like me to abandon my "Assume the best possible case for the anti-ban argument" policy?

You've demonstrated that there's a reason that people play MK, certainly, but I don't think you've proven that the reason is that he's better. This shows that people perceive him as a better character, and thus they choose him, but that's not proof that he's actually better.

 

[Delvro]

You've demonstrated that there's a reason that people play MK, certainly, but I don't think you've proven that the reason is that he's better. This shows that people perceive him as a better character, and thus they choose him, but that's not proof that he's actually better.

Yes, he has shown that MK's dominance extremely likely NOT due to popularity.

So, how much better does MK do at comparable skill levels? On average, within this plateau, MK mains are scoring 2.54 times as many points as Snake mains.
The overall outperformance we're trying to account for is, from Ankoku, 1.99 times as many points.

That is, the data suggest that MK is, overall, an unusually UNPOPULAR a character given how good he is!

 

[Tien2500]

You've demonstrated that there's a reason that people play MK, certainly, but I don't think you've proven that the reason is that he's better. This shows that people perceive him as a better character, and thus they choose him, but that's not proof that he's actually better.

Ok. So lets say people perceive MK as better. Why does this explain him actually getting better results? If he isn't actually a better character why are his results twice as good when controlling for popularity?

 

[Omni]

I'm not sure what Mortimer is saying. Metaknight is clearly the best character. o_O

 

[salaboB]

Ok. So lets say people perceive MK as better. Why does this explain him actually getting better results? If he isn't actually a better character why are his results twice as good when controlling for popularity?

The most competitive players believe he's the best and so switch to him, elevating his results above where his actual "best character" point would put him.

I don't think this is the case, mind you, but it's a possible explanation.

 

[Tien2500]

The most competitive players believe he's the best and so switch to him, elevating his results above where his actual "best character" point would put him.

I don't think this is the case, mind you, but it's a possible explanation.

Then that still leaves you with the same problem. For this to be true MK players have to be about twice as good as Snake mains at every possible level. Its not very likely.

 

[Omni]

The most competitive players believe he's the best and so switch to him, elevating his results above where his actual "best character" point would put him.

I don't think this is the case, mind you, but it's a possible explanation.

Mew2King used DDD before he switched to MK.
DSF used Toon Link, Snake, Wario, and DDD before he stayed consistently on MK.
Havok used Marth before he switched to MK.
Seibrik used DDD before he switched to MK.
Overswarm originally used ROB before switching to MK.

Those are just some of the MK's I can think of at the moment.

Metaknight's overall performance and view from the public does give him more attention in the retrospect that switching to him will give better results since he's deemed as the best character. I don't see how this couldn't be taken into account as to a reason why MK's point distribution is inflated.

 

[salaboB]

Then that still leaves you with the same problem. For this to be true MK players have to be about twice as good as Snake mains at every possible level. Its not very likely.

If a significant majority of the most competitive players (So the ones who sink the most time into practicing) pick MK because they believe he'll give them the best chances of winning, they could be better than the Snake mains (Combined with a slight boost due to MK being the best character.)

 

[rvkevin]

I'm not sure what Mortimer is saying. Metaknight is clearly the best character. o_O

At what point would you say that MK over centralizes the meta game?

 

[etecoon]

Then that still leaves you with the same problem. For this to be true MK players have to be about twice as good as Snake mains at every possible level. Its not very likely.

http://www.youtube.com/watch?v=AKIxu669Y_A

It's plausible.

lol bad trolling