Tier List Discussion as Apex's interesting top 8 is over

[Technical_Knockout]

edit: sorry for double-post i hadn't realized this one made it to this page... requesting to delete this post.

 

[Technical_Knockout]

imho a match-up chart is imperative for:

1. consistency between the tier lists.
2. concretely ranking the characters.
3. transparently explaining placement.
4. updating the dreamland-only ruleset.

 

[Shears]

I could settle on a tier list if there was some definitive and objective measures used to explain placement. Simply averaging votes that are either unexplained and/or opinion doesn't sit well with me. I might make a thread to discuss other charts and form a tier list from a combination of factors.

 

[Olikus]

I could settle on a tier list if there was some definitive and objective measures used to explain placement. Simply averaging votes that are either unexplained and/or opinion doesn't sit well with me. I might make a thread to discuss other charts and form a tier list from a combination of factors.

I second this. koro shall have for the good effort, but as it stands now there are like 70 percent randoms with lots of weird opinions. A tier list shoudnt be determined solely on democratic voting from unexperienced people. Its ofc fun to see what the result will be, but dont put it up anywere as an official list or something.

 

[Shears]

I second this. koro shall have for the good effort, but as it stands now there are like 70 percent randoms with lots of weird opinions. A tier list shoudnt be determined solely on democratic voting from unexperienced people. Its ofc fun to see what the result will be, but dont put it up anywere as an official list or something.

I'm going to use a series of metrics and use a formula to weigh each that we can then debate whether the formula and metrics used are worthwhile. With that said, I think the democratic voting will hold value and this thread and the results of the voting will serve a purpose. Maybe like 10% or something of the tier list will come from popular vote. The weight of each metric will also be voted on. I think the first step is to gather results and data for each metric. Then weigh each metric to determine its impact in the final results. A formula using each metric and its weight will be used to normalize the composition of these metrics to a calculated 1-12 tier list.

I will be creating a thread for this sometime soon which will basically outline the steps I believe are important to creating a tier list that people so desperately care to have. Basically:

1. Determine possible measures
2. Gather data/results for measures
3. Determine weight of each measure
4. Calculate rating value for each character using measures and their respective weights
5. Bestow upon the common smasher a composite and calculated tier list for SSB64

 

[Sedda]

70% arent just randoms lol

determining rules for a tier list voting system is going to take you longer than actually voting for a tier list.

seems like a waste of time

 

[Shears]

70% arent just randoms lol

determining rules for a tier list voting system is going to take you longer than actually voting for a tier list.

seems like a waste of time

Your mom was a waste of time




http://en.wikipedia.org/wiki/List_of_burn_centers_in_the_United_States

 

[KeroKeroppi]

i agree with some other people

the majority of the players posting lists have no ****ing clue what they're talking about

their opinions should not be taken into consideration

we should have like 15-20 good / educated players vote on a list

also

we could just have javi construct a list by himself bc he's always right

ty

 

[Olikus]

70% arent just randoms lol

determining rules for a tier list voting system is going to take you longer than actually voting for a tier list.

seems like a waste of time

i prefer longer time over worse list. but as i said, we can just post the results now and see what the poeple that have posted just for the fun of it. just not make it official or anything

 

[BananaBolts]

You guys realize that Koro isn't going to count every vote and make that the definitive tier list, right?

I think he's going to compile data for just the notable players alongside the current tier list w/ randoms. All this talk about alternate ways to decide on a way to calculate a tier list is silly. People aren't just randomly thinking up a list in their heads and spewing it out onto a keyboard. Everything you're mentioning about other ways to determine a tier list is being thought out by the individuals that submit their tier list. Just take Yoshi for example. A lot of people are placing him higher this time. Why? Primarily tournament placement and notable character potential. No matter how you want to go about making a tier list, it will always be subjective. Even if you decide to make a tier list solely based on tourney results, it will still be flawed due to the decision of what tournaments affect the data.

People choose characters from a subjective view anyways. The whole idea of a quantitative, metric-based list is flawed for the intents and purposes of a tier list. Basing a subjective list on opinions formed from knowledge of playing, watching, theorizing, and discussing is probably the optimal way of making a tier list. Players are basing their tier list off of their knowledge of the game, tourney results, character potential, etc.

Shears, you claim to not care about any tier list yet you refuse to accept a tier list unless it meets your standards. ???

1. Determine possible measures
2. Gather data/results for measures
3. Determine weight of each measure
4. Calculate rating value for each character using measures and their respective weights
5. Bestow upon the common smasher a composite and calculated tier list for SSB64

1. subjective
2. objective but based on subjective restrictions
3. subjective
4. objective but based on subjective restrictions
5. objective

I recognize that every measurement has to be decided on prior to its use but you're asking for really, really subjective data/measurements. It's just not going to be any more fruitful than our current method. Besides, people will simply decide on measurements that properly represent their own selfish opinions. That thread will result in time wasted for something that's likely worse than what we're doing right now.

Koro has given people sufficient time to submit their tier list. An article will be primed to go when he's finished. The net exposure from this will be very good for us especially since this will be done before Smashcon, our projected biggest tourney of the year. The thread that you propose will only postpone an equally subjective tier list w/ no added benefits to its completion.

Edit: Another thing to note is that a list compiled by only the good players is still subjective and even in several cases, misinformed. That's like thinking the President knows best just because he's in a state of power. He's human; he's not always right.

Edit 2: A lot of players (good ones especially) have strong bias against certain characters. Luckily, the way we're compiling data it at least averages out for the most part.

 

[Sedda]

let me put it this way then:

from looking at pidge's results on the last page, whats your problem with that particular tier list? you cant throw random numbers out like "70%." if you want just ask koro not to count the votes of people who havent been playing long (and i dont think he will anyway) but youd be getting into another big issue and would have to come up with a criteria of whose vote is allowed which is also subjective in the first place.

waste of time

 

[Shears]

People aren't just randomly thinking up a list in their heads and spewing it out onto a keyboard. Everything you're mentioning about other ways to determine a tier list is being thought out by the individuals that submit their tier list

The fallacy in the current tier list design is the belief you and others have that the above statement is true. I do not think people are thinking about it as deeply as they should be. Its natural for people to overvalue certain things and undervalue others when in reality some of those undervalued things are more important and some of those overvalued things are less important. The problem is people have bias and because of that they cut corners and pick and choose (subconsciously) which factors matter. They think they made their decision based on certain measures and when they go through it in their head they think they've covered all the bases, but in reality, when things are left to peoples own judgement, they tend to overlook a lot of information. There have been several studies polling reality vs what people think reality is and almost always the actual results differ from the peoples opinions of reality. Ask and I will give you a personal example of how actual results differ from opinionated results (even if those opinions seem to be based off fact). If we break it down I truly believe people will see that their first posted tier list will differ from their calculated tier list. Either way, you do not have to waste any more time if you don't want to, this will be an endeavor of mine and anyone else who wishes to "waste" their time on it.

As far as the subjectivity of my system, yes it is not perfectly objective but I think it has a greater deal of objective information and basis than the current design to build the tier list. In fact your breakdown of my design shows some objective points while I don't see anything objective about the current system since the criteria of which to build an individuals list is not universal.

In baseball, football, basketball, and every single sport, there are statistics and metrics used to evaluate athletes. Outside of the direct measurements there are plenty of sabermetric and advanced statistics that are used and accepted as reliable and accurate measurements of an athletes rating/skill. These metrics come from a combination of objective and subjective analysis using direct and indirect measurements but still they are universally accepted and paint a greater picture than opinion polls or even the raw stats. Most player rating formulas use aggregate stats combined with agreed upon weights to form these mostly objective ratings that accurately represent player value. So why can't we take a similar approach to smash and the tier list.

I believe people try to make a personal tier list as unbiased as they can based off as much concrete evidence that they can, but because they are human and subconsciously lazy, they unknowingly make incomplete or subjective lists and I think if my suggested analysis is done with far more data apparent and definitely used, we will see a difference in the popular vote and the calculated rating. I double dog dare you to give it a shot.

 

[KoRoBeNiKi]

I would like to point out regarding this that when I have asked a "certain other player" his opinion, the only thing he posted was "lol." So it's a bit hard to get every top player's opinion, especially when several of our top players either don't post on Smashboards or frankly don't give a ****.

I don't plan on making separate lists due to two reasons. First off, what Sedda said, outside of the waste of time. There is no exact clarity in my post.

More though, it's the simple thought. that at the beginning, did I think that people would have been like, "Hey this is a cool idea. I'm a top level player. Therefore, I should contribute to this! I will clearly post on smashboards, not troll and hey, we might even get the 64 backroom back, and not just troll/spam there! Time to go get all the top Kailera people, translate things to other languages, and do everything else possible."

We are not remotely that good of a scene, especially online. We might be a relaxed bunch of fun loving jerks offline but that doesn't translate to helping online.

I'd like to see you try to get all of the top players here and not have just a ton of Komo level posts as well as arguments against tier lists (arguments against a tier list, what is this Gamefaqs 2007?)

I made this list for a simple reason, the other one is hella old and Hyrule is no longer a legit stage and it used to be THE legit stage.

 

[Shears]

I would like to point out regarding this that when I have asked a "certain other player" his opinion, the only thing he posted was "lol." So it's a bit hard to get every top player's opinion, especially when several of our top players either don't post on Smashboards or frankly don't give a ****.

I don't plan on making separate lists due to two reasons. First off, what Sedda said, outside of the waste of time. There is no exact clarity in my post.

More though, it's the simple thought. that at the beginning, did I think that people would have been like, "Hey this is a cool idea. I'm a top level player. Therefore, I should contribute to this! I will clearly post on smashboards, not troll and hey, we might even get the 64 backroom back, and not just troll/spam there! Time to go get all the top Kailera people, translate things to other languages, and do everything else possible."

We are not remotely that good of a scene, especially online. We might be a relaxed bunch of fun loving jerks offline but that doesn't translate to helping online.

I'd like to see you try to get all of the top players here and not have just a ton of Komo level posts as well as arguments against tier lists (arguments against a tier list, what is this Gamefaqs 2007?)

I made this list for a simple reason, the other one is hella old and Hyrule is no longer a legit stage and it used to be THE legit stage.

I think what you've done is great and I think it has certainly been pivotal in opening the tier list discussion and updating the perspective of this game and even taking it to the next level. But I also think its not done and there is more that can be done regarding this and so I will take it upon myself to try and go one step further.

 

[asianaussie]

can we discuss team tier lists now

 

[Capos]

Why do most people consider pikachu the unquestioned number 1 when kirby is great against two of the more common characters (falcon and yoshi). Is because he wins the kirby-pika matchup? Is he significantly better against the lower tiers? Not sure how pika-fox compares to kirby-fox.

 

[Technical_Knockout]

best recovery = best gimping potential.

edit: again, a match-up chart would provide the data for constructing a good tier list.

 

[KoRoBeNiKi]

Why do most people consider pikachu the unquestioned number 1 when kirby is great against two of the more common characters (falcon and yoshi). Is because he wins the kirby-pika matchup? Is he significantly better against the lower tiers? Not sure how pika-fox compares to kirby-fox.

Best recovery by far in the game, I mean who else can escape combos just by Up B? Pikachu does win that matchup vs. Kirby and wins every other matchup as well. The only higher tier matchup where Kirby is better is Falcon and perhaps Yoshi as you noted (but then again, I know a lot of Yoshi users who dislike the Pikachu matchup just as much) but as has been known recently, Falcon hasn't been nearly as common. I honestly think that the only reason why Kirby is more often seen is since Kirby is that much easier to use. In terms of Pikachu-Fox/Kirby-Fox, I am not positive either, both are winning for Rat and Kirby, but by how much and if there is a difference, I am not sure.

Though I do think the list is much closer than it used to be, in terms of how much better Pikachu is (as in it's not as clear.) I don't understand people putting Pikachu in it's own tier anymore. I mean it's clearly those two at the top but SS compared to S tier, nah.

 

[Capos]

Thanks for answering.

On Yoshi - during Apex, Sheer, the commentators, Kero's play made it seem like Kirby is awful for Yoshi (much more so than Pikachu), so I thought that was a consensus. Maybe I was wrong on that part.

 

[KoRoBeNiKi]

Thanks for answering.

On Yoshi - during Apex, Sheer, the commentators, Kero's play made it seem like Kirby is awful for Yoshi (much more so than Pikachu), so I thought that was a consensus. Maybe I was wrong on that part.

Well I personally do think Yoshi vs. Pikachu is pretty bad but as said Yoshi/Kirby is worse, but not by that much.

 

[Horbie]

Has anyone ever considered using the PageRank algorithm to do a tier list?

The more I think about it the more it seems like this is probably the most accurate way. It takes into account each individual matchup and takes into account how these match ups speak to the tier list as a whole. If a character has a particularly good match up it will only boost their rating significantly if it's against a character who's high on the tier list.

This is prolly the best way to go about it from a mathematical perspective in terms of consistency, and tbh it makes compiling the tier list much easier, because all you have to do is agree on individual match ups. Which is way easier to do, and even if there isn't consensus it's much easier to average those numbers without losing information than to somehow compile a ton of lists which are only ordered and for the most part ignore the disparity between different ratings (e.g. char 1 might be a wee bit better than char 2 who is cray better than char 3).

My b for saying this so far into this whole process, but in the future I think this is prolly the way to go. Plus super easy to set up a system that auto-updates a tier list based on tourny results if you do it this way.

 

[Shears]

Has anyone ever considered using the PageRank algorithm to do a tier list?

The more I think about it the more it seems like this is probably the most accurate way. It takes into account each individual matchup and takes into account how these match ups speak to the tier list as a whole. If a character has a particularly good match up it will only boost their rating significantly if it's against a character who's high on the tier list.

This is prolly the best way to go about it from a mathematical perspective in terms of consistency, and tbh it makes compiling the tier list much easier, because all you have to do is agree on individual match ups. Which is way easier to do, and even if there isn't consensus it's much easier to average those numbers without losing information than to somehow compile a ton of lists which are only ordered and for the most part ignore the disparity between different ratings (e.g. char 1 might be a wee bit better than char 2 who is cray better than char 3).

My b for saying this so far into this whole process, but in the future I think this is prolly the way to go. Plus super easy to set up a system that auto-updates a tier list based on tourny results if you do it this way.

The problem in that formula is the circular reference to who is at the top of the tier list. Thats what the debate is. That formula requires a tier list to already exist, but if we can't agree on a tier list then we have nowhere to start. There also isn't an agreed upon matchup chart either. My thread takes a step in the direction you are looking at but people here are too lazy/exhausted from this to do any more work.

 

[Horbie]

The PageRank algorithm does not in any way require a tier list to start from. It shouldn't depend on anything other than the individual match up numbers. That's why it's so useful in most of its applications.

And yah, I realize that there isn't necessarily a consensus on the match up numbers. I just think it is much simpler to find a meaningful metric by averaging the match up numbers that contributors submit than to compile a large number of lists into a single list.

 

[Shears]

The PageRank algorithm does not in any way require a tier list to start from. It shouldn't depend on anything other than the individual match up numbers. That's why it's so useful in most of its applications.

And yah, I realize that there isn't necessarily a consensus on the match up numbers. I just think it is much simpler to find a meaningful metric by averaging the match up numbers that contributors submit than to compile a large number of lists into a single list.

But you're giving weight to certain matchups over others. The way you describe the algorithm suggests that a favorable matchup against someone like Ness isn't as valuable as a favorable matchup against Pikachu. In order to determine that weight an initial tier list is needed.

 

[Horbie]

The way that PageRank is implemented is usually through some sort of iteration.

The gist is that you take the "rankings" from the previous step and use them to determine the rankings for the current step of the algorithm. You repeat this over and over until the rankings stop changing, and at this point you've arrived at the PageRank vector. And this PageRank vector is unique.

Technically you do need some sort of initial ranking for the first step. But it's been mathematically proven that no matter where you start (as long as all entries have non-zero values to begin with) you will always arrive at the same final result, hence the PageRank vector being unique.

So yah, we need somewhere to start, but it doesn't matter where cause we'll get the same answer either way. Typical starting point is to assume everyone is equal and let the algorithm build from there.


Which match ups are given weight isn't arbitrary or determined by what tier-list we start with. It's determined by the match ups themselves.

 

[Capos]

If you had a full matchup chart, iteration 1 would just be the unweighted average, no?

 

[Horbie]

If iteration 0 was all characters with equal ranking then yes.

 

[Olikus]

let me put it this way then:

from looking at pidge's results on the last page, whats your problem with that particular tier list? you cant throw random numbers out like "70%." if you want just ask koro not to count the votes of people who havent been playing long (and i dont think he will anyway) but youd be getting into another big issue and would have to come up with a criteria of whose vote is allowed which is also subjective in the first place.

waste of time

do what you want. just dont make it official. i dont mind to se a sedda bananabolt tier list. its fun to see peoples opinions anyway. im all about that.but as it stands now waaaay to few good players has contribituted to even think about having an official new tier lsit which shall represant the whole community. but a random oeples list or something like that is no biggie. As I said it would be fun to see.

 

[Sedda]

Yeaaaah don't call it a "Sedda bananabolts" tier list because it's not that. MANY good players have given their opinion, so I don't know what you're talking about and why it should matter that you have a problem with the tier list.

The only thing that I'm hearing is "I don't like this tier list so don't make it official." I would go back and count the number of amazing players who have contributed to the list, but you've already come up with your weird "70% randoms" figure, so it's clear that you don't care enough about the actual numbers.

 

[Dancin Jesus]

Tier lists don't exist
Player skill does

This.

 

[BananaBolts]

Yeaaaah don't call it a "Sedda bananabolts" tier list because it's not that.

Yo, what are you doing?! Olikus is trying to ship us and you're ruining our moment </3

 

[Sedda]

studevolution is the only ship i endorse

it's pratically confirmed

 

[Olikus]

so you basically mean the majority of the votes are really good players? then I guess there are nothing more to discuss on this topic between us.

good luck with the list kids.

 

[KoRoBeNiKi]

so you basically mean the majority of the votes are really good players? then I guess there are nothing more to discuss on this topic between us.

good luck with the list kids.

I didn't say that, there just happens to be enough strong players that the random **** opinion won't matter.

 

[Sedda]

so you basically mean the majority of the votes are really good players? then I guess there are nothing more to discuss on this topic between us.

good luck with the list kids.

no. im saying that you said that 70% of the players who voted are random and that is not true. i also said that you shouldnt call it the sedda bananabolts list because it's not that. dont put words in my mouth lol wtf are u doing olikus seriously.

but ill invite you to count the number of good players who have provided their list since you still havent done it.

 

[Fireblaster]

Guys, whatever tier list comes out of the mass opinions from koro's thread can just be labeled as "community tier list". There doesn't only have to be one tier list.

 

[THE_MAAFIA]

You guys accepted my list right? Just making sure (it wasn't a joke list btw)

 

[Olikus]

I didn't say that, there just happens to be enough strong players that the random **** opinion won't matter.

yeah I know it wasnt a reply to you. Guess I was unclear.

no. im saying that you said that 70% of the players who voted are random and that is not true. i also said that you shouldnt call it the sedda bananabolts list because it's not that. dont put words in my mouth lol wtf are u doing olikus seriously.

but ill invite you to count the number of good players who have provided their list since you still havent done it.

take it easy man. save the temper. As i mentioned I have no interest in this list anymore. so just do what you need to. Maybe it even turns out great who knows.

 

[The Star King]

take it easy man. save the temper.

nothing in Sedda's post suggested he was angry, don't do this "u mad" **** in a serious discussion it's childish

/petpeeves

 

[asianaussie]

stop bullying the norwegian, you know that internet nuances arent part of their culture