Meta Knight's Winnings Data Is Wrong

SaveMeJebus · Aug 25, 2011

http://img542.imageshack.us/img542/5411/week33.png

If you look at it, they take all of the percentages from the "Total Cash Won in Tourney" which means that the (No Split) and Average percentages go over 100%

Budget Player Cadet_ · Aug 25, 2011

Question: is it seriously wrong, or is it just like 1-2% that are odd?

John12346 · Aug 25, 2011

Jebus, we explained this to you many times over. We gave you detailed explanations as to why the percentages add up to above 100%, and why it is completely correct, and each and every ****ing time, you just simply took a **** in our face and said "You're wrong" without giving us any legitimate explanation. You just continued to insist your unsupported theories over and ****ing over again. No one thinks you're right on this point. Just... stop already.

John12346 said:
"Total Cash Won in Tourney" is actually supposed to read "Cash Won(Full Split)." I just haven't gotten around to renaming it yet... >___>;

Splitting completely doesn't give the characters enough money.
No splitting at all gives the characters too much money.

The average is the closest thing we've got to the true value.

John12346 said:
The totals for the latter two don't really mean anything. The percentages for "Full Split," "No Split," and "Average" are all based off the total sum of all characters found in "Full Split," because that is the true value of how much money has circulated 2011 Brawl tournaments. As a result, this causes the percentages in "No Split" and "Average" to not add up to 100%, but instead go over. There is a reasoning behind this, mind you.

Drop all of your assumptions for now, and assume that only one Brawl Tournament has occurred in all of 2011, and it paid out $50 to first place only.

Assume a MK/Snake users wins this tournament. For "Full Split," MK and Snake would both have $25 each, and the total sum of their overall money would add up to $50, which is the total amount of money that has circulated 2011 Brawl so far. In this case, both characters would have 50% in this category.

For "No Split," both characters would receive $50. Now, the total sum of the money in this section would add up to $100, and that $100 doesn't represent anything! Only $50 has circulated 2011 Brawl so far, so this hundred dollar figure doesn't hold any value to us whatsoever. By basing the percentages on the total amount of money that has circulated 2011 Brawl, we see that MK and Snake both have a value of 100% in this section, which interprets to, "MK users have won 100% of all available tournament money, and Snake users have won 100% of all available tournament money." The statements are completely true, even though they add up to 200%. The total sum of the percentages in "No Split" adds up to some random arbitrary figure that doesn't really represent anything either.

And of course, as far as "Average," we're averaging something that adds up to 100%, and something that goes over 100%, so of course the average will also exceed 100%. Everything stated for "No Split" applies here for the most part.

John12346 said:
Okay, let's work under the assumption there has only been one 2011 Brawl tournament, and it paid out $50 to a MK/Snake user.

It would look like this. The percentages in the latter two categories add up to above 100%, but in no way does that prove that the data is wrong. The fact there are percentages that exist over 100% in this case is due to the fact that players use more than one character in tourney, it's that simple.

C.J. said:
You can NOT add up all of the totals in the no split total. That is more money than has been put into circulation.

The amount of money that has been put into circulation via brawl tournaments, is $80863.37. You find this number at the bottom of "Cash won in tournament" column. This column should be named "Cash won in tournament (full split)." In this column, for any player who uses a character, their winnings get divided evenly among the characters they have used during the course of the tournament. As a result, this is the ONLY column that has an ACCURATE total of the amount of money put into circulation via brawl tournaments (excluding MMs, crew battles, Brawl mods, MK banned tournies, and any other exception you could POSSIBLY think of in order to try to refute any of that).

Now, the next thing you do, is you go to the "No Split" column. This shows much much money PLAYERS who have USED MK have won. The amount that MK has in this column is $44692.41. NOW, COMPARE THE TWO NUMBERS WE HAVE.

Total money from Brawl tournaments- $80863.37
Money players have won while using MK- $44692.41

DIVISION TIME!!!
44692.41/80863.37
maybe you can do this mentally, I'm going to use a calculator.
It comes out to ~55.27%
HEY, that is over 50%!
Now, this is the players who have used MK, you're right. That's not completely fair. However, for the exact same reason that this isn't fair, you can't use the column titled, "Percentage of Each Dollar Won" which shows ~38.79%.
Now, since both of those are unfair for the exact same reason, what is the best, reasonable solution? Use the average column! So, looking at that column we come up with 47.03%.

Hey, you're right, it's not over 50%. However, 50% is a completely arbitrary number, and seems like a VERY high expectation for a game with 40 characters in it. NOW, you're going to argue that there's only 10 or 15 or 20 viable characters or whatever. And you're going to say that in that case, it's not that high of an expectation! COOL! Let's do this now! Subtract all of the winnings from every character under the top 10, or 15, or 20. I'm lazy and not going to do this. Eyeballing it, it looks like you're taking about $5000 out of the equation. While I'm not sure, if I had to guess, MK's % of money won would go up, to probably right about 50%!!!

The sums of the columns are not relevant. The percentages of anything except arguably the "No Split" and "Average" column are not relevant. QUIT REFERENCING THEM.

For those of you who are unaware as to why I'm abnormally saltier than usual, read this, starting from post 717 all the way to the end, where the thread gets locked.

SaveMeJebus · Aug 25, 2011

You obviously don't know percentages if you are getting more than 100%

@BPC, it is off by at least 10%

UltiMario · Aug 25, 2011

How about this?

From what I gather, If they chopped off everything but the Average column, the chart would mean the exact same thing as what the chart says right now. The two other columns are minimums and maximums earned because due to multiple mains, exact numbers aren't exactly available. The first column is merely mislabeled, just a small mistake. There are no percentages that go over 100%, as the second cash column is not a percentage of the first one.

If I got anything wrong, someone please correct me, but from what I see, there's nothing wrong here.

Bizkit047 · Aug 25, 2011

I wonder if Jeebus qualifies for being banned for excessive trolling.

John12346 · Aug 25, 2011

Almost as if to reinforce what I said, you choose to not listen and repeat how I'm wrong.

::sigh::

I'll just have to bring a real world example into this. Let's say we have a spinner, labelled as such:

One could argue that 1 appears twice(50%) or thrice(75%) on this spinner; the same applies to 2, of course.

With the argument of both numbers appearing twice on the spinner, we parallel the Full Split scenario, where equal dues are given to any collaborating numbers. This interpretation is perfectly legitimate, and also has the probabilities add up to 100%.

However, with the argument of both numbers appearing thrice on the spinner each, we parallel the No Split scenario, where all selected numbers gain full credit for being selected, instead of splitting in the previous scenario. This interpretation is also legitimate, and has the probabilities add up to 150%.

A controversy exists for which interpretation of the spinner is the correct one. Either could be right; it's completely subjective. The most objective way to approach it would be to take the average of the two. While this applies really awkwardly to the spinner example, it suits the intent of my charts perfectly fine and reaches the most compromised and agreeable solution.

theunabletable · Aug 25, 2011

copied from my AiB post:

The no split part goes over 100% because it gives each character the full amount of money that place won.

Like say we have a tournament where first place wins $1,000, and the person winning used MK, Falco, and Diddy. afaik no split means that they add 1,000 to MK, Falco, and Diddy each. Split means that MK would get 333.3, Falco would get 333.3, and Diddy would get 333.3.

I'm not completely sure if that's what it means, but I always thought that's what the split thing meant, and if that's what it means that'd explain why no split shows more money.

Although I'd like to note that when using data (to everyone here, as any time I've seen someone use data, they leave something really important unmentioned), the raw data by itself doesn't help anything, there's a lot that has to be taken into account to make data like this actually useful.

Shockna · Aug 25, 2011

SaveMeJebus said:
You obviously don't know percentages if you are getting more than 100%

@BPC, it is off by at least 10%

Never took anything above High School Statistics, did you?

SaveMeJebus · Aug 25, 2011

What if a player only used MK for certain stages or certain MUs? What if the player used his main more than he used MK. Is it fair to give both characters an even amount of money? If MK was truly a problem, He would have a higher percentage in "Full Split" because players wouldn't need to main any other character besides MK which means all the money would go to MK.

Deleted member · Aug 25, 2011

This thread is completely unnecessary. Go argue your point in the Character Success/Popularity thread please.

Player-1 · Aug 26, 2011

Jebus you do realize this is what you do every single time, right? You put up some ******** argument then you get proven wrong and then you go ask some "what if" question that just derails from the original point, over the time of the 2nd argument you're bringing up it circles back to the first argument THAT WAS ALREADY PROVEN WRONG pretty much ignoring it from the first time.

SaveMeJebus · Aug 26, 2011

If MK was really a problem, He would have a higher percentage in "Full Split" because players wouldn't need to main any other character besides MK which would mean that all the money the players made would go to MK

Player-1 · Aug 26, 2011

Jebus...before we move on to another topic can you just at least admit you were wrong with your math?

John12346 · Aug 26, 2011

Jebus, that's some flimsy logic right there. It can be worked in the reverse direction.

If I were to use MK in 28 of my matches at a national, Lucario in 1 match, and Ganondorf in one other match, and I won $1000 total, how fair could it possibly be that all three characters only get $333.33 despite seeing leagues of different usage from me?

The proper interpretation of my charts would probably have to be: "(Character) has won between (Full Split)% and (No Split)% of all tournament money." Since we don't have an accurate indicator of how much usage each character has seen from each player, we can just go ahead and use the Average for our best guess.

SaveMeJebus · Aug 26, 2011

All this data shows is that there seems to be more of a problem with players that main MK and other characters than there is with players who just main MK.

John12346 · Aug 26, 2011

Not really, in my above scenario, MK is receiving about $600 less than what he should be receiving for "Full Split."

Who's to say these kinds of cases don't occur more than the ones you're describing? Maybe MK's "Full Split" value should be a lot higher for all we know?

SaveMeJebus · Aug 26, 2011

No split assumes that the players that won the tournament did it all with one character since you give every character the full amount of money. This doesn't happen with "Full Split" which is why MK has a lower percentage there

John12346 · Aug 26, 2011

That's... uh... you're kinda stating the POINT of my chart now.

When we consider a full split for all characters, we observe the lowest possible amount of money any given character has won, because every character receives their fair share in that scenario. And of course, when we consider no split, we observe the highest possible amount of money any given character could have possibly won, because every character just flat out receives the full amount for their respective tournament. A better way to observe this would be to check the Spinner post that I had made earlier in this thread, because it almost accurately parallels the phenomenon we're seeing here.

What you need to recognize is that these are the two extremes of the statistics, and it holds true for every character. It's also very unlikely(actually, nearly impossible) for any character to have won the amount of money defined by either of the two extremes, but rather some amount in between.

Since we don't actually have a full fledged way of reliably determining everyone's ratio of character usage per tournament, the only way we can accurately interpret the data is either by using the Average((Full Split + No Split)/2) or a Range Theorem that simply hits the two extremes(Between (Full Split) and (No Split)).

SaveMeJebus · Aug 26, 2011

Let me give you an example:

MK main makes 50$
MK main makes 50$

Diddy main makes 50$
Diddy/MK main makes 50$

Snake main makes 50$
Snake/MK main makes 50$

Falco main makes 50$
Falco/MK main makes 50$

Your results would make it seem like MK is the problem when the real problem is the players that co main MK

John12346 · Aug 26, 2011

Or maybe MK just has a lower percentage because people choose to use other characters ONCE or TWICE during a high-payout tournament, thereby depriving MK of at least half of those big winnings.

It's just as likely as your scenario, and equally as unprovable.

SaveMeJebus · Aug 26, 2011

Then why should we take your data as fact? People are using your data as evidence for the MK ban

John12346 · Aug 26, 2011

I don't think you're aware of what your assertion actually proves...

Let's just assume you're correct, for a moment, and that MK really does have to attribute ALL of his tournament winnings to the usage of other characters. In that case, MK's true percentage of tournament winnings is (Full Split).

However, if the reverse is true(due to scenarios like the one I stated above), and MK cannot attribute any of his winnings to any other sub-characters the player may have used, then MK's true percentage of tournament winnings is (No Split).

Both of our assertions are true to some extent, but neither is 100% true, meaning that MK's true percentage of tournament winnings is neither (Full Split) nor (No Split), but rather something in between. As such, we can refer to whatever value that may be as "(Average)" or "Between (Full Split) and (No Split)."

Your choice.

SaveMeJebus · Aug 26, 2011

"However, if the reverse is true(due to scenarios like the one I stated above), and MK cannot attribute any of his winnings to any other sub-characters the player may have used, then MK's true percentage of tournament winnings is (No Split)."

How would this one ever be true? It's obvious that they used both characters to win

John12346 · Aug 26, 2011

I use MK in 28 out of 30 of my sets at a national, Lucario in the remaining 2. I win $1000, what now?

SaveMeJebus · Aug 26, 2011

John12346 said:
I use MK in 28 out of 30 of my sets at a national, Lucario in 2 others. I win $1000, what now?

You said it yourself, you don't count stuff like that. That's why you guys didn't include M2K's ganon in the winnings

Gardex · Aug 26, 2011

Because he only used him one match.
In pools.
And lost.

John12346 · Aug 26, 2011

Basically that. Only the super extremes that are obvious outliers(or requested out by the player) are the removed data. I don't really go crazy nuts with removing data; only if there is an IMMENSE problem with it and it's pointed out to me.

Look, the main point is that the value of (Full Split) that you're trying to argue to be the truth just isn't so, just like how (No Split) is not the correct value either. It's something in between, and it's as simple as that. It could be higher than (Average), or it could be lower, but we can say to some degree of certainty that it's near (Average).

SaveMeJebus · Aug 26, 2011

Ok then. The winnings from all of the MKs (including people who second and third main him) is somewhere between 38% and 55%

John12346 · Aug 26, 2011

SaveMeJebus said:
Ok then. The winnings from all of the MKs (including people who second and third main him) is somewhere between 38% and 55%

Or, better yet, now that I think about it, this does take into account all tournaments, correct? What if we wanted to emulate things down to the higher levels of play? Say... tournaments with at least 30 entrants? Well, luckily, someone did request that I do such a thing, so I have been faithfully updating that, all the while. This chart only takes into account tournaments that had an attendance of at least 30 people, meaning that in order to win any of the tournaments listed, you'd need some serious skills in order to win.

Or perhaps, rather than skill, a certain character...?

How does... 43.52% to 60.64% sound?

SaveMeJebus · Aug 26, 2011

You could have also made one that includes OOR results but you decided to leave those out didn't you?

Kantrip · Aug 26, 2011

Well damn.

Jebus, please stop strawmanning. What does making a chart for OOR results have to do with anything?

SaveMeJebus · Aug 26, 2011

CrimsonFeint said:
Well damn.

Jebus, please stop strawmanning. What does making a chart for OOR results have to do with anything?

What does making a chart for $ winnings in tournaments with 30 players or more have to do with anything? More players does not equal higher level play

Spelt · Aug 26, 2011

I believe he said to stop strawmanning.

SaveMeJebus · Aug 26, 2011

I didn't bring that up. John12346 did. I am just responding to his post

ElDominio · Aug 26, 2011

John12346 · Aug 26, 2011

Tournaments with at least 30 entrants does not PERFECTLY reflect what the higher levels of play are, but what it does do is remove all tournaments with an otherwise lower base of players, AKA where players with lower skill levels can earn money with just a few amount of wins. With this, the data becomes more influenced by nationals and regionals, which suddenly become more commonplace. This is unlike the data that takes into account all tournaments, because any 10 random locals, where nonskilled players are able to win more easily, can add up to the monetary strength of a regional or national.

Y'know, now that I think about it, maybe I'll have another look, and make a chart that only compiles the REALLY BIG tournaments, and see what happens then...

SaveMeJebus · Aug 26, 2011

John12346 said:
Tournaments with at least 30 entrants does not PERFECTLY reflect what the higher levels of play are, but what it does do is remove all tournaments with an otherwise lower base of players, AKA where players with lower skill levels can earn money with just a few amount of wins. With this, the data becomes more influenced by nationals and regionals, which suddenly become more commonplace when only considering tournaments with at least 30 entrants. This is unlike the data that takes into account all tournaments, where any 10 random locals, where nonskilled players are able to win more easily, can add up to the monetary strength of a regional or national.

Y'know, now that I think about it, maybe I'll have another look, and make a chart that only compiles the REALLY BIG tournaments, and see what happens then...

You should be gathering separate data for tournaments with a certain amount of ranked players instead

Player-1 · Aug 26, 2011

ElDominio said:

i ****ing died.

Player-3 · Aug 26, 2011

http://pigtales.minecraft.net/8007553067ce4f82972268a4893f6642

Meta Knight's Winnings Data Is Wrong

Smash Master

Smash Hero

Smash Master

Smash Master

Out of Obscurity

Smash Lord

Smash Master

Smash Lord

Smash Cadet

Smash Master

Deleted member

Guest

Smash Legend

Smash Master

Smash Legend

Smash Master

Smash Master

Smash Master

Smash Master

Smash Master

Smash Master

Smash Master

Smash Master

Smash Master

Smash Master

Smash Master

Smash Master

Smash Lord

Smash Master

Smash Master

Smash Master

Smash Master

Kantplay

Smash Master

BRoomer

Smash Master

Smash Journeyman

Smash Master

Smash Master

Smash Legend

Smash Hero

Information

Network