N.A.G.A.C.E
Smash Champion
^ ok sorry i just felt i was being attacked for no reason
-
Welcome to Smashboards, the world's largest Super Smash Brothers community! Over 250,000 Smash Bros. fans from around the world have come to discuss these great games in over 19 million posts!
You are currently viewing our boards as a visitor. Click here to sign up right now and start on your path in the Smash community!
Character Rankings List - Post-March 2010
- Thread starter Zankoku
- Start date
- Status
Hylian didn't play Ice Climbers in Melee, last I checked.
memphischains
Smash hhkj'
try wobbles and chu dat
FishkeeperTimmay!
Smash Ace
*Facepalms*
I've been mixing up Hylian with other pros for years. I don't know why.
At any rate, it doesn't ruin my argument.
Hey Ankou, I'm almost done with my model. I'll post the completed model in probably a few days. For the most part all I was having trouble with was the "fee" part of the model. Since generally fee translates into pot size relative to the number of people playing it works. However the problem in controlling the variable lies in the fact that you might have a high cost entry tournament with crappy turn-out, or just a cheap tournament with massive turnout. So using the fee as a direct method of modifying the point system is inherently flawed. However, while your system has little problem with that it's issue lies in the fact that you always consider the "top 8" when if you were to look at it purely from a statistician's point of view that won't work with different base populations. Also your data points will skew much more if you just have a basic divisor to control the point value increases. You need to use a top %, and higher math functions to describe the fee/population relationship. Here's what I have so far...
% Ranked
The % Ranked of a given tournament population (P) is 12%. If a number of similarly ranked players exist that are equivalent to players within the top 12% that comprise of 50% or more of that ranking level then they will all be counted. If not that entire ranking level will be discarded. All point values will be assigned the following points based on % divisions, the greater value the player could receive is always the one received.
#1 Placement: 8
Top 25% of % Ranked: 6
Top 50% of % Ranked: 4
Remaining % Ranked: 2
These point values will be divided by the RANKED POPULATION TOTAL (#R).
The highest level of points is always assigned unless multiple occurrences of a character exist. If multiple character appearances occur then the included data will consist of the top number of players equivalent to 3% of the INITIAL TOURNAMENT POPULATION(P).*
For example: In a tournament of 64 people the top 8 (12%=7.68 rounded up to 8) are...
1. Snake
2. Metaknight
3. Lucario
4. Marth
5. Luigi
5. Metaknight
7. Snake
7. Snake
Point assignments would be:
Snake = ([8+2]/8) = 1.25
Metaknight = ([6+2]/8) = 1
Lucario = ([4]/8) = .5
Marth = ([4]/8) = .5
Luigi = ([2]/8) = .25
*The alternative to this is to apply the following equation which will also limit point gain, but the first part does it in a much simpler manner....
ALTERNATIVE POINT GAIN LIMITATIONS
{[A(.6)^(X-1)]+ [B(.6)^(X-1)]+[C(.6)^(X-1)]+...[n(.6)^(n-1)]}
#R
Occurrences of Snake and Metaknight in the above example:
Snake
[A(.6)^(X-1)]+ [B(.6)^(X-1)]+[C(.6)^(X-1)]
[8(.6)^(1-1)]+[2(.6)^(2-1)]+[2(.6)^(3-1)]
[8(.6)^(0)]+[2(.6)^(1)]+[2(.6)^(2)]
[8(1)]+[2(.6)]+[2(.36)]
8+1.2+.72 = 9.92 divided by 8= 1.24
Metaknight
[6(.6)^(0)]+[2(.6)^(1)]
6+1.2= 7.2 divided by = .9
This method ends up with all recurrences but at an exponentially decreased value. Thus, they contribute to the data but do not skew it so drastically in favor of their recurrences but from their ability to place high. Popular characters will still maintain a lead, but their popularity will not place them in an insurmountable lead.
The Fee Factor
Now we reach the real issue with this model is the fact that while population and entrance fee are both important variables in establishing how a tournament should be weighted, the extreme values created by some potential tournaments could throw data off course.
The average tournament fee tends to come out to about $7.84 from what I've compiled from both your list and other sites worth of information. If someone comes up with a different number let me know.
The average tournament population tends to come out to about 31 (30.85) people. If someone comes up with a different number let me know.
Population will take care of itself for the most part, since the bigger the population the more potential points to be distributed. HOWEVER, population also needs to be included with the fee to properly weigh the prize pool. Since most of you do not POST PRIZES, I need to spend time compiling data on the average ratios between population, prizes, and entry fees; since all of those are a factor.
Let me know what you think Ankou, I'll get back to you with the information on fee/population/prize payout. Hopefully fee+payout stay consistent with each other that way I can use them as a constant ratio instead of worrying about them as a variable.
P.S. I know this is a bit technical, but every good mathematician knows: If your equation outputs data that isn't in a line or a curve, you're not trying hard enough.
% Ranked
The % Ranked of a given tournament population (P) is 12%. If a number of similarly ranked players exist that are equivalent to players within the top 12% that comprise of 50% or more of that ranking level then they will all be counted. If not that entire ranking level will be discarded. All point values will be assigned the following points based on % divisions, the greater value the player could receive is always the one received.
#1 Placement: 8
Top 25% of % Ranked: 6
Top 50% of % Ranked: 4
Remaining % Ranked: 2
These point values will be divided by the RANKED POPULATION TOTAL (#R).
The highest level of points is always assigned unless multiple occurrences of a character exist. If multiple character appearances occur then the included data will consist of the top number of players equivalent to 3% of the INITIAL TOURNAMENT POPULATION(P).*
For example: In a tournament of 64 people the top 8 (12%=7.68 rounded up to 8) are...
1. Snake
2. Metaknight
3. Lucario
4. Marth
5. Luigi
5. Metaknight
7. Snake
7. Snake
Point assignments would be:
Snake = ([8+2]/8) = 1.25
Metaknight = ([6+2]/8) = 1
Lucario = ([4]/8) = .5
Marth = ([4]/8) = .5
Luigi = ([2]/8) = .25
*The alternative to this is to apply the following equation which will also limit point gain, but the first part does it in a much simpler manner....
ALTERNATIVE POINT GAIN LIMITATIONS
{[A(.6)^(X-1)]+ [B(.6)^(X-1)]+[C(.6)^(X-1)]+...[n(.6)^(n-1)]}
#R
Occurrences of Snake and Metaknight in the above example:
Snake
[A(.6)^(X-1)]+ [B(.6)^(X-1)]+[C(.6)^(X-1)]
[8(.6)^(1-1)]+[2(.6)^(2-1)]+[2(.6)^(3-1)]
[8(.6)^(0)]+[2(.6)^(1)]+[2(.6)^(2)]
[8(1)]+[2(.6)]+[2(.36)]
8+1.2+.72 = 9.92 divided by 8= 1.24
Metaknight
[6(.6)^(0)]+[2(.6)^(1)]
6+1.2= 7.2 divided by = .9
This method ends up with all recurrences but at an exponentially decreased value. Thus, they contribute to the data but do not skew it so drastically in favor of their recurrences but from their ability to place high. Popular characters will still maintain a lead, but their popularity will not place them in an insurmountable lead.
The Fee Factor
Now we reach the real issue with this model is the fact that while population and entrance fee are both important variables in establishing how a tournament should be weighted, the extreme values created by some potential tournaments could throw data off course.
The average tournament fee tends to come out to about $7.84 from what I've compiled from both your list and other sites worth of information. If someone comes up with a different number let me know.
The average tournament population tends to come out to about 31 (30.85) people. If someone comes up with a different number let me know.
Population will take care of itself for the most part, since the bigger the population the more potential points to be distributed. HOWEVER, population also needs to be included with the fee to properly weigh the prize pool. Since most of you do not POST PRIZES, I need to spend time compiling data on the average ratios between population, prizes, and entry fees; since all of those are a factor.
Let me know what you think Ankou, I'll get back to you with the information on fee/population/prize payout. Hopefully fee+payout stay consistent with each other that way I can use them as a constant ratio instead of worrying about them as a variable.
P.S. I know this is a bit technical, but every good mathematician knows: If your equation outputs data that isn't in a line or a curve, you're not trying hard enough.
Yoshi Kirishima
Smash Lord
Wow very well done. I've also seen some of those flaws and such. This reminds me a lot about the stuff we learned in math
Hey I'm Canadian too
Falconv1.0
Smash Master
Yeah, it would either be A.) A person or B.) A stray cat.
@Bouse:Get that math away from me and keep it to PM's, my head hurts.
Anyways, once again for my staple leaving comment.
****ING USE SHEIK ****IT
Pikaville
Pikaville returns 10 years later.
That maths stuff just remind me of how good it is not to have to do maths anymore.I wanna see pika move up that list.We have tournamnets every week/second week in Dublin if you want to take those into consideration.
Yoshi Kirishima
Smash Lord
And now I call you grateful and loving, AND rude and obnoxious, so now you're grateful and loving, rude and obnoxious, and grateful and loving and rude and obnoxious XD. jk
and btw, why does the Fee have to be accounted into Character Rankings? Im guessing its like the more you have to pay the more it should weigh cus the more you pay the more serious you get?
FishkeeperTimmay!
Smash Ace
Thats exactly the idea. Random tournaments with nearly no pot don't attract player from abroad. When 1000+ dollars is on the line, people will travel for the chance at that. Its to help keep big tournaments more meaningful than small ones. =)
Yoshi Kirishima
Smash Lord
Ok, I see thanks
Bouse, what do you intend me to output with that formula? I kinda figured I'd just have a simple formula with which to compile a list where both popularity and performance had a weight on it, seeing as how I intended to gauge how the average tournament scene might look in terms of character population/possible prediction on which characters might win.
Your list definitely puts a very large weight on popularity. I'm not saying it's not a factor, but to just directly input data like that will mean that even MONTHS from now when say Character X grows in popularity the amount of points attained by Snake will grant him an insurmountable lead and even though Character X is currently played the most it will take months upon months for your ranking list to reflect that data. This is the problem.
It's fine for now if nothing ever changes, if nothing new is discovered, or if no new play-styles are built. However, since your method of calculating data SUPPORTS popularity in a raw format then that's what it will primarily be influenced by. To leave it as an unfettered variable will just perpetuate that ranking chart's information. "Snake is most popular? He must be most popular because he's the best. I should play Snake."
My opinion is that tournament PLACEMENT is what is important. If people just flood tournaments with Snake's which is probably what will most likely happen sooner or later, if he wins those tournaments he'll still be on top. Hell he'll probably maintain a lead if he's at least ranks in top 4.
The point being that limiting the growth of a character due to popularity is reasonable, because then your list will only exhibit that. Metaknight could end up placing relatively similar in a tournament, with my example of having two Metaknight's who both did well, and there was a .25 difference between the two for that tournament. The difference is much smaller than if you had counted all data equally (.5 difference).
Yes, on a scale this low a change of .25 is relatively large, add together a hundred tournaments with fees and population considered and that could account for anywhere from 50-200 points on a ranking system.
To be honest, I'm just offering a method I see as being more accurate, if you don't want to use it then it's no skin off my back. I'm just trying to do my part as a member of this community to help make it as accurate as possible. If you don't want a method to enter data that will require a little extra work on your part that's alright by me, and it's completely understandable. I don't even know if I'll finish it because of how much random data I'm getting for the prize-fee ratio which I had hoped would turn out a bit better.
Another issue is that you do NOT take into consideration the prize-fee ratio. I could see this as an invaluable piece of information. Potential Gain against Potential Loss. I hope to the Flying Spaghetti Monster that someone had brought this up before I did, otherwise I'd feel sad all over.
It's fine for now if nothing ever changes, if nothing new is discovered, or if no new play-styles are built. However, since your method of calculating data SUPPORTS popularity in a raw format then that's what it will primarily be influenced by. To leave it as an unfettered variable will just perpetuate that ranking chart's information. "Snake is most popular? He must be most popular because he's the best. I should play Snake."
My opinion is that tournament PLACEMENT is what is important. If people just flood tournaments with Snake's which is probably what will most likely happen sooner or later, if he wins those tournaments he'll still be on top. Hell he'll probably maintain a lead if he's at least ranks in top 4.
The point being that limiting the growth of a character due to popularity is reasonable, because then your list will only exhibit that. Metaknight could end up placing relatively similar in a tournament, with my example of having two Metaknight's who both did well, and there was a .25 difference between the two for that tournament. The difference is much smaller than if you had counted all data equally (.5 difference).
Yes, on a scale this low a change of .25 is relatively large, add together a hundred tournaments with fees and population considered and that could account for anywhere from 50-200 points on a ranking system.
To be honest, I'm just offering a method I see as being more accurate, if you don't want to use it then it's no skin off my back. I'm just trying to do my part as a member of this community to help make it as accurate as possible. If you don't want a method to enter data that will require a little extra work on your part that's alright by me, and it's completely understandable. I don't even know if I'll finish it because of how much random data I'm getting for the prize-fee ratio which I had hoped would turn out a bit better.
Another issue is that you do NOT take into consideration the prize-fee ratio. I could see this as an invaluable piece of information. Potential Gain against Potential Loss. I hope to the Flying Spaghetti Monster that someone had brought this up before I did, otherwise I'd feel sad all over.
Squidster
Smash Ace
He resets the list every few months though, theres never an unattainable number.
Of coarse popularity will affect the results a little. I'm sure metaknight and snake aren't that high because of popularity. If popularity was such a factor ike and marth would probably be higher. Even if a character is unpopular if the character is good then people will play as them in tournaments just look at wario.
Yoshi Kirishima
Smash Lord
Bouse, may be you should consider making your own Character Rankings List? I think that would be a good idea, I mean there isn't a rule about how many dif Char Rank Lists there can be. That way there'd be more than 1 way to look at data, which would be useful.
First of all, Ike is a character who won't rank high for his obvious short comings. I'm not saying that the entire list will be drastically changed. In the A and B levels a few characters may shift places, but I doubt the S rank which change other than the fact that it won't have it's ranks separated by such outlandish a number of points.
Odds are I will try to though (to Ragnarok17), the main issue that seems to exist is that the data being recorded isn't extensive enough and the models people are willing to use are too simplistic to describe and properly weight point distribution.
sagemoon
Smash Lord
Bellevue smash tournament
entree fee 10$
entrants: 16
http://www.allisbrawl.com/event.aspx?id=1877
1st Sagemoon (Pit)
2nd Bladewise (marth/GW)
3rd Eggz (Meta)
4th Emo$ (Snake/dedede)
5th Help (mario)
5th Zero (Dedede)
7th Ben (Dedede)
7th Luigitheking (Luigi/dedede)
entree fee 10$
entrants: 16
http://www.allisbrawl.com/event.aspx?id=1877
1st Sagemoon (Pit)
2nd Bladewise (marth/GW)
3rd Eggz (Meta)
4th Emo$ (Snake/dedede)
5th Help (mario)
5th Zero (Dedede)
7th Ben (Dedede)
7th Luigitheking (Luigi/dedede)
Winston
Smash Master
Just wondering, how much of the Lucario representation in the top8/top4/topwhatever is solely Azen?
I bet it's quite a large chunk of it =)
I bet it's quite a large chunk of it =)
cj.Shark
Smash Ace
nice job sagemoon. You like killing penguins dont ya xD
GimmeAnFSharp
Smash Journeyman
What good is a past results list, when the 2-month old results aren't being edited out of the large section?
The large section, being only updated once a week, will no longer have the archaic results either at the time of the next update.
ChronoPenguin
Smash Champion
>.>...this makes me uncomfortable.
Also, I declare D3 is not a penguin, because penguins are far more uber >.>
h4rdcor3
Smash Journeyman
As a lucario main, i'm gonna say he gets a lot of credit for the larger tourneys. There still aren't very many lucario players but more and more are showing up on the boards. We may be seeing some more lucario wins in the near future.
GimmeAnFSharp
Smash Journeyman
Ah. Thanks for clarification.
memphischains
Smash hhkj'
Everyone should look out for the GRAZE combo video, coming soon
goodkid
Smash Lord
He has a large chunk, but there are some Midwest Lucario players like Bloodhawk & Mendoza that take some credit. If you want to see the names, just search Lucario on the tourney list, I bet Azen has about 5 placements up there.
asob4
Smash Champion
ugh my oli has been lagging ***
i'll get him back up to A class soon enough ^^
i'll get him back up to A class soon enough ^^
sagemoon
Smash Lord
Surprisingly I didnt even play 1 dedede the whole tournament. Emo$ went snake on me.
foxsmasher
Smash Cadet
I missed the meaning of having that list of "past cutoffs". Please give an explanation.
Hardcorenesser
Smash Journeyman
Tournament #1
Number of entrants: 24
Entry fee: 5 dollars
Where: Visalia, California
Top 8, mains only:
1st: Yoshi (Jon)
2nd: Lucas (Me)
3rd: Jiggs (Alan)
4th: Ness (George)
5th: Lucas (Erik)
5th: Ike (John)
7th: Pokemon Trainer (Michael)
7th: Kirby (Alex)
Tournament #2
Number of entrants: 24
Entry Fee: 5 dollars
Where: Visalia, California
Top 8, mains only:
1st: Lucas (Me)
2nd: Yoshi (Jon)
3rd: Marth (Jose)
4th: Lucas (Erik)
5th: Snake (Daniel)
5th: Jiggs (Alan)
7th: Falco (Angelo)
7th: Ness (George)
Number of entrants: 24
Entry fee: 5 dollars
Where: Visalia, California
Top 8, mains only:
1st: Yoshi (Jon)
2nd: Lucas (Me)
3rd: Jiggs (Alan)
4th: Ness (George)
5th: Lucas (Erik)
5th: Ike (John)
7th: Pokemon Trainer (Michael)
7th: Kirby (Alex)
Tournament #2
Number of entrants: 24
Entry Fee: 5 dollars
Where: Visalia, California
Top 8, mains only:
1st: Lucas (Me)
2nd: Yoshi (Jon)
3rd: Marth (Jose)
4th: Lucas (Erik)
5th: Snake (Daniel)
5th: Jiggs (Alan)
7th: Falco (Angelo)
7th: Ness (George)
asob4
Smash Champion
AVC Tournament 2up
Pacific West
57 entrants
$5
1. Teba - DeDeDe
2. Creepy Fingers - Falco
3. Lite - Snake/Olimar
4. BoA - Lucario
5. Mike Haze - Marth/DeDeDe
5. Aphro - Kirby
7. Shadow - ROB
7. Lee Puff - Game and Watch
everyone with two characters put significant time in with those characters
Pacific West
57 entrants
$5
1. Teba - DeDeDe
2. Creepy Fingers - Falco
3. Lite - Snake/Olimar
4. BoA - Lucario
5. Mike Haze - Marth/DeDeDe
5. Aphro - Kirby
7. Shadow - ROB
7. Lee Puff - Game and Watch
everyone with two characters put significant time in with those characters
You people need to remember thread links and dates. Also, Hardcorenesser posted the most unbelievable results I've seen.
Besides this of course:
Fireworks Brawl
Midwest
459 Enterants
Entry Fee: 600$$ (Although the TO took 95% of the pot to cover food costs)
Date: 7/4/08
Thread: gameFAQs thread already deleted sorry
1: Xiivi (Captain Falcon)
2: M2K (Yoshi/Sheik/ROB/Peach/Snake/Metaknight/Luigi) He seriously used them all equally.
3: Ankoku (Sheik/Zelda/SheikZelda/ZeldaSheik) Yeah, all those were used equally.
4: Ken's Mom (Zero Suit Samus/Samus)
5: I forget
5: I said I forget
7: Overswarm (Pikachu/Jigglypuff/Lucario/Mewtwo)
7: My friend Fred who doesn't have a smashboards tag (Roy)
I swear they're true.
Besides this of course:
Fireworks Brawl
Midwest
459 Enterants
Entry Fee: 600$$ (Although the TO took 95% of the pot to cover food costs)
Date: 7/4/08
Thread: gameFAQs thread already deleted sorry
1: Xiivi (Captain Falcon)
2: M2K (Yoshi/Sheik/ROB/Peach/Snake/Metaknight/Luigi) He seriously used them all equally.
3: Ankoku (Sheik/Zelda/SheikZelda/ZeldaSheik) Yeah, all those were used equally.
4: Ken's Mom (Zero Suit Samus/Samus)
5: I forget
5: I said I forget
7: Overswarm (Pikachu/Jigglypuff/Lucario/Mewtwo)
7: My friend Fred who doesn't have a smashboards tag (Roy)
I swear they're true.
Adapt
Smash Lord
Ken's Mom is top tier!
She wins all the tournaments I go to
She wins all the tournaments I go to
GUIGUI
Smash Journeyman
- Joined
- Jan 21, 2006
- Messages
- 418
I don't get it!
In the "Most up-to-date rankings, no details" list (only two days old), Ganonondorf is under U, which probably means he has never reached top 8 in a competition, and yet in the "details" list, it is written that Ganon is under D (with a top4 reaching).
If the "Most up-to-date rankings, no details" list is more recent than the "details" list, how did Ganon disappeared? How can an update suddenly decide that previous record suddenly doesn't count?
GUIGUI please learn to read. Tournaments more than 2 months old are dismissed.
gantrain05
Smash Master
hellz yeah for peach moving up again, she's gonna be top of mid tier soon, i know it.
GUIGUI
Smash Journeyman
- Joined
- Jan 21, 2006
- Messages
- 418
OK, I see it, now, tough it should be tell with the rest of the rule above on the the Original Post, rather than just be mentioned on the outdated list as if it was just a side note.
Also, directly mentioning when the "details" list was last edited on top of it would be useful.
- Status