• 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!

Record of Main - Popularity Track List For 2.5

Status
Not open for further replies.

B.W.

Smash Champion
Joined
Apr 27, 2006
Messages
2,141
Location
Darien, IL
Like that's even hard to do lol.

Updating from my last one though to follow the new chart.

B.W.
Toon Link = 8
 

Burnsy

Smash Lord
Joined
Nov 4, 2012
Messages
1,167
Location
Phoenix, AZ
Calculations are nothing, I have all the data in Excel, which does everything for me.

I'm not the best at data management, so I just need something that is reasonable to give a view of how much people use characters.
Did everyone ignore this post? The data is all in excel and 1 smart formula fixes this problem. Why bother with making corrections when plenty of people aren't going to come back to re-vote?
 

~Frozen~

Smash Apprentice
Joined
Jan 4, 2013
Messages
157
Location
NY
NNID
Frozen491
3DS FC
3909-8017-8600
Updating from old post:

~Frozen~
Lucario - 5
ZSS - 5
Wolf - 2
Peach - 2
 

Dubforce

Smash Sidius
Joined
Jan 1, 2013
Messages
212
Location
Jacksonville - UNF
I'm following the 6 + 2(x) rule, as I read it in the OP. Everyone else seems to be summing to 10, so I will also include a version that does that, just in case.

List me as "VGDev"

6+2(x):
1. Squirtle - 6
2. Toon Link - 5
3. Ike - 2
4. Luigi - 1

10:
1. Squirtle - 4
2. Toon Link - 3
3. Ike - 2
4. Luigi - 1
 

F. Blue

Smash Journeyman
Joined
Oct 11, 2011
Messages
433
F. Blue
:jigglypuff64:: 5 points
:bowsermelee:: 4 points
:pit:: 3 points
:falconmelee:: 2 points
:nessmelee:: 2 points
:snake:: 2 points
 

Badge

Smash Apprentice
Joined
Jan 3, 2013
Messages
186
How so
I can't imagine how anyone could be upset with what the system does.
I'll be blunt:
The system is:
  • completely arbitrary
  • skewed in favor of multi character mains (which was the intention)
  • impossible to properly convert back into the simple 10 point system
  • hard to convert into any different information for that matter
  • unnecessary complicated on those giving data (which in turn lowers the amount of usable data you'll get)
  • incompitable with the old system
Or in other words: An unclean solution for a subjective problem.

Just fiddling with the (raw data)->results function would have done the same but without tossing all your old data overboard or preventing you from easily switching back and wouldn't have been more work for GMaster. I already told him about what I'd suggest on that end, so I won't repeat that further.
Now that the new system was used, just switching back seems unprofessional and may confuse people even more, but it's IMO the best you can do. The old system gives good raw data without any avoidable subjective component and can be interpreted to give very similiar information to the new one. It's also less confusing and still used by some people, so now's not too late to go back on what I'd call a mistake due to a too hasty decision.
 

SpiderMad

Smash Master
Joined
May 6, 2012
Messages
4,968
Shouldn't someone from some other fighting game community know all the right systems to use for this stuff by now?
 

leelue

Smash Lord
Joined
Aug 3, 2004
Messages
1,926
Location
All up in your personal space, NY
8adge
Do you not see the tremendous weight given by single-main players to their characters as a problem? Adding fractional parts and wholes of a static pool and comparing the numbers in a way that leads people to believe that some characters are more popular than they are looks like a problem to me.


For the record, you can convert back to the old system easily, or any system by converting the scores given into percentages. You may not get round numbers, but that hardly seems relevant.

The "easy" alternative to the main problem I had with the old system would be to let people rank their own characters out of 10 apiece. Which would be even more heavily skewed towards multi-main players. To me, both "easy" solutions (my hypothetical and the original system) are bad solutions considering the way the data are presented.

I basically drew off of the US government's legistrative branch. There is a "base" number for everyone to have a semblance of equal say, but the people who represent more than one character aren't made to be "half as important". Any compromise would seem arbitrary as far as I can tell. But this one should not be confusing with proper instruction.
I don't know why the OP hasn't just posted the point-pool you have to choose from in the OP rather than let people mess up on their own in either the math or the interpretation.
 

Badge

Smash Apprentice
Joined
Jan 3, 2013
Messages
186
Do you not see the tremendous weight given by single-main players to their characters as a problem? Adding fractional parts and wholes of a static pool and comparing the numbers in a way that leads people to believe that some characters are more popular than they are looks like a problem to me.
Whether it is a problem depends on what you want to measure. Say, you define the popularity of any given character as the chance that a randomly chosen player would use that character in a match. Then just averaging across all responses in the old system gives you an ok estimate*, but the new system doesn't.

Now, the definition I gave above won't coincide with your, GMaster's or many other definitions of popularity, but that's not a problem with the old system, because the data you collected includes as few bias as possible towards any subjective definition of popularity. You can just go ahead, interpret your data cleverly and get a good estimate for how popular characters are by your definition. But you can also give an estimate for the definition above and some other stuff.
For the record, you can convert back to the old system easily, or any system by converting the scores given into percentages. You may not get round numbers, but that hardly seems relevant.
This does not work, because the new system intentionally tries to increase the numbers for secondaries and decrease those for primaries. You could just go ahead and make them into percentages, but you would get lower numbers for characters that are mained often and higher numbers for other characters than you would had the same people posted their data in the old format. E.g. you could now add each of the characters you like to play for fun at 2 without changing the numbers of the characters you actually use, etc..
The "easy" alternative to the main problem I had with the old system would be to let people rank their own characters out of 10 apiece. Which would be even more heavily skewed towards multi-main players. To me, both "easy" solutions (my hypothetical and the original system) are bad solutions considering the way the data are presented.
This actually gathers totally differnt data. The old system is something akin to the "How probable are you to use a character?" I mentioned above*, the system you propose here asks instead "How much do you like to play every individual character?".

The current system is a mixture between both and "Which characters would you play at all?" that produces difficult to interpret data with a subjective evaluation function ingrained in the data collection process. It's also confusing for people answering the survey not only because of the calculating required but also because the information asked for isn't clear.

This was a problem before, but is even more so know. On one hand you've got people that post their information mostly as they would have done before, but on the other hand there are responses that seem to just add any character the poster likes to play in friendlies sometimes, because it only gives additional points anyway. The old system made you think about which characters you really do play, the new system instead makes people about how many points exactly they are supposed to give.

What you want is here data that's easy to analyze, precise, quick to gather and as objective as possible. The evaluation process will do the rest for you.

*This isn't exact, of course. People gave more points to characters they liked to play than characters they played mostly because they are good or increase their chances to win, ignored characters below a threshold of approximately 10% etc..
 

TheKittyloaf

Smash Cadet
Joined
Apr 7, 2013
Messages
45
Location
Atlanta, Georgia
in the current format, it seems as if someone who mains more than 2 characters is weighed more heavily opinionatively in the overall system. Based solely on the fact that if someone can amass enough points by "maining X amount of characters" they could assign at least 10 between two of their characters (or like, two characters with 8), and then include values for their lesser played options. Sure there is a drawback to someone saying that they main one character and would only play that single character competitively, but if it's the truth then how does that skew the data at all. I would rather have that one person's vote count the same as someone else (or in that sense just allow everyone the same point value in total for who they would play competitively like the old system). if i mained half of the cast or saw fit to play them in tourney.. ok that's a bit unreasonable, but someone posted 7 characters for a total of 20 points. i could hypothetically put 8 points in one character, 7 in another, and 1 in the rest. and then, my personal character selection would count practically twice that of someone who just mains 1 character.
 

Rarik

Smash Journeyman
Joined
Feb 20, 2013
Messages
206
Location
Boston
I agree with 8adge. The data from a 10 point system roughly translates into how likely that character is to be played, which creates no issues from people only main-ing 1 character as they're are 100% likely to pick that character, while a person who voted say 6 for their main and 4 for another character, is saying I'm only going to play my main about 60% of the time, and then the other 40% of the time I'm playing my secondary. The current system being used I honestly have no idea what it's measuring although what 8adge said about it seems to make sense.

However, by the dictionary definition, neither system is measuring popularity.
Oxford English Dictionary said:
Popularity:
The state or condition of being liked, admired, or supported by many people.
Source
So, by that definition, the only thing that matters to how popular something is, is the number of people that like/support it and the amount they like it hardly matters. Which, if we're only tracking the popularity of Mains then any point system can measure popularity as you just take the character with the most points and that's 1 vote towards that character. If we want the popularity of a character in general, I personally feel that frequency covers that, but if we're going by the definition, you could just simply count the number of people that play that character, and completely disregard any sort of weighted system like the point systems.



Multi-point systems suck for measuring popularity, just count how many people play a character.
 

Badge

Smash Apprentice
Joined
Jan 3, 2013
Messages
186
Kittyloaf addresses an important flaw I forgot to mention: The new system was intended to increase the effect of lower singular scores compared to higher ones. What it instead does is incentivize the (alleged) usage of multiple characters by giving votes of players that use multiple characters more weight. You depend on the good will and insight of the respondents for those two to be equal and especially the latter you're guaranteed not to get from many people.
However, by the dictionary definition, neither system is measuring popularity.
The thing with dictionary definitions is that they won't change what people mean with what they say and only help understand them if you've either got no idea wath the word is supposed to mean or they are strictly followed by that very person in that specific or all instances of usage. This is why you usually define any words you want to be understood the right way before using them.

On the other hand, from what I understood GMaster is also mainly interested in how many people use every character. But for that you can just take the old data and identify every number with one before adding the seperate results, so you have it included in your data (with the minimum usage threshold being somewhere around 5-10%).
 

PMS | LEVEL 100 MAGIKARP

Hologram Summer Again
Joined
Jan 16, 2013
Messages
3,303
Location
Tri Hermes Black Land
well, first of all this system is stupid, you're giving a disproportionate amount of weight to people who main one character and occasionally play a smattering of others. The x/10 system is by far better. Multi mains may not like their mains half as much as solo mains like their mains, but we're not trying to figure out who likes what character, we're trying to figure out how often each character is played.

That being said:

6+2x system
Mario 8
Gdubs 2.5
Diddy 2.5
Pit 1

% or 10 system
Mario 65%
Gdubs 15%
Diddy 15%
Pit 5%

edit: I know my ratios aren't the same, **** swag, the % one is more accurate for me, but I'm too lazy to calculate that out for the stupid system
 

Problem2

Smash Champion
Joined
Jun 12, 2006
Messages
2,318
Location
Crowley/Fort Worth, TX
NNID
Problem0
My updated submission:

Problem0
Sonic - 8
Fox - 2

Actually, mine didn't change under the new system! Oh well, it probably needs to be recorded again anyways.
Also, I notice a slight flaw with the 8+2x system. By giving your secondaries a score of 1, you can inflate the worth of your main beyond a person maining exclusively that character.

In example

Solo Main
Fox - 8

I play lots of people!
Fox - 12
Marth - 1
Falco - 1
C. Falcon - 1
Peach - 1
 

PMS | LEVEL 100 MAGIKARP

Hologram Summer Again
Joined
Jan 16, 2013
Messages
3,303
Location
Tri Hermes Black Land
My updated submission:

Problem0
Sonic - 8
Fox - 2

Actually, mine didn't change under the new system! Oh well, it probably needs to be recorded again anyways.
Also, I notice a slight flaw with the 8+2x system. By giving your secondaries a score of 1, you can inflate the worth of your main beyond a person maining exclusively that character.

In example

Solo Main
Fox - 8

I play lots of people!
Fox - 12
Marth - 1
Falco - 1
C. Falcon - 1
Peach - 1

Yeah, that's why he instated the max of 8 pts per char rule, which is still stupid as it changes the ratio of playing, which doesn't give realistic scores. So yeah, the whole system is broken.
 

GMaster171

Smash Ace
Joined
Aug 26, 2012
Messages
676
Location
Halifax, NS
This thread is pretty much done. I've lost the file with the data somehow, and don't really feel like redoing everything. Ellondu said he would pick this up again after 2.6 comes out, so no worries.

This thread can be closed now.
 
Status
Not open for further replies.
Top Bottom