The North Carolina Melee Power Rankings! Updated 8/14/14!

[lord karn]

balksdj;flaksjdfalksdjfasdf

 

[lord karn]

In a nutshell, things the panel has affirmed about the PRs:

1. The PRs are used for seeding
2. the PRs are not a list of the best people that compete in NC tournaments
3. The PRs are not based off tournament placing(who does best at NC tournaments), for the reason of #4 being more important.
4. The PRs are based off who v. who

Karn's gripe's with the my propositions:









Additionally, 5. Karn mainly does the seeding, and Kevin(see quote directly above).

So... shall I starkly delve? It is starkly so!


There are the two tournaments out of four Karn attended, that counted for the most recent PR period.

http://www.smashboards.com/showthread.php?t=311005&highlight=tgm3

Karn gets 7th out of 13th. He beats no one notable, he lost to Chris and Stingers.

http://www.smashboards.com/showthread.php?t=311870&highlight=ncsu

Karn gets 17th out of 25. He beats no one notable, he lost to Mahone and Sushi.

---

Alright, so here comes the latest NCSU, karn attending so I can only assume that he did the seeding, and if not certainly assisted with it.

Link here: http://www.smashboards.com/showthread.php?t=314800 (it shows the bracket and pool data... stingers, you are the MAN btw)



Why was karn placed in the easiest(debatable with 2nd easiest, I guess) pool if seeding is TRULY done by the PR / the process that dictates the PR?




What other conclusion can I make by looking at the data then seeing which pool you were in? What I've said all along, the seeding wasn't done right. You do realize that messing up the seeding in pools can have an even worse affect than just seeding a bracket off CORRECT data, right? It would seem so from your posts, but not demonstrated.

I then notice that you received the 2nd best #1 seed from your performance in the easiest pool, not to mention that even concerning performance in pools Boss and Danny Mexico both had better set count records than you.


So, Karn got a good placing without actually beating anyone(an exact scenario he mentioned), which is unfair to non-PR'd people just like he said....It's literally one of the main situations he described. But wouldn't ya know it, it was from bad seeding(Big Daddy Dorsey's memory is going but it almost seems like I've said this before). I guess the plus side is seeing Boss clean sweep all our semi-decent players that attended, with generic yet effective plumber combos, in tourney opposed to just half of them(half-way sarcastic with that last sentence).


In short:

Worst case scenario with basing our rankings off of tournament placing: We have a list of who does the best at NC tournaments. We should do this.

SEEDING is what should be based off who beats who, and the panel should control this. If you beat someone that IS seeded above you or below you, your seeding value (x) should fluctuate as it does in ANY competitive seeding, poll, or ranking. It's a self-sufficient process, but the basis would undoubtedly(something to initially go by) be started out with some kind of "pre-season rating"(like other competitive events), in which objective data is plugged into to CONSTANTLY until it turns into an ever-changing, beautiful process. Take "pre-season" college basketball rankings for an example: The polls already had a good, subjective idea of who the top teams are, and they use this subjective seeding data to measure the impact certain wins/losses will have on a team's rank. Amongst the 30+ games played by each team in a season though, the list is fine-tuned over-time with objective data which in essence, IS who v. who(except there's no sport seasons.. competitive smash is continuous). Amongst the panel's discretion with who v. who, they should be at least comparing seeding value (x) to seeding value (y) to better measure future seeding. I mean wtf... How can you realistically seed a handful of above avg players in our state properly? With an actual seeding point system you could have 5 players within a few tenths of a point, but at least you have the data to show WHY they got that certain seed.

This is what we should not do:

Seed players SLIGHTLY different because they are from the same region.
Seed players SLIGHTLY different to avoid sheik ditto MUs, IC v. peach, etc.
Seed players SLIGHTLY different because of any preconceived notions.
etc.

If the pools are seeded properly(unlike at the recent NCSU), then great. If there isn't pools, and say 18 people come to a tournament, you just list them 1-18 solely dependent on who beats who. #17 and #18 play a pre-lim. match to see who plays the #5 seed, seeds 1-4 all get BYES. And don't make any "SLIGHT" changes.....

Karn, sorry to call you out. In light of realizing I was being mocked for my ideas on this for months on the brocator thread by my name, I decided to share it.

Everyone is welcome to tear my post apart like ravenous wolves. Bring it on.

Just saw this. I didn't do the seeding at NCSU. I went there directly from work and barely made it in time. Someone else made the pools. I'm not sure who did it. As I said in my previous post, Josh probably ended up in Boss's pool because of the last minute switch. That should not have been the case. It should have been either me or someone ranked lower than me perhaps. But once we have already done pools, we can't not seed by them, so it kind of messed things up.

And you're right, technically we should not change things even slightly because of region/matchup/whatever. However, it is really lame for people from the same region play each other. I personally don't care if we slightly mess things up to prevent this from happening. If people feel differently, then we can stop doing it. I've said it before, but I really don't care about PRs. I only do it to help out because I think I understand it. I'd rather mess up the PRs than screw over a couple player's tournament experience (and in fact, creating accurate PRs helps to not screw up their experience, by making seeding more accurate). However, at bigger tournaments (where we get multiple instances of individual seeds), we can fix by region without even changing the players seeds. And I never change anything because of matchups either. Just ask Gofg, who somehow always ends up playing peaches.

And yeah, I wasn't even on the most recent PRs, lol.

 

[-ACE-]

Pretty sure Dorsey is against people from oos being in the same pool lmao. We need top cherish our visitors and treat them fairly, as they serve as valuable experience for our players' tournament performance. The main point he was making is why you had it unfairly easy that tournament when your recent data had no notable wins, I believe.

 

[Juno McGrath]

Pretty sure Dorsey is against people from oos being in the same pool lmao. We need top cherish our visitors and treat them fairly, as they serve as valuable experience for our players' tournament performance. The main point he was making is why you had it unfairly easy that tournament when your recent data had no notable wins, I believe.

 

[bossa nova ♪]

lmfao



karn gets more **** than any panelist

 

[-ACE-]

He didn't say pr's should be used for seeding...

PR's should be based off tournament performance. Placing shows tournament performance. If you look at the PR's and say "wow, that doesn't look like a list of our top 10 players" that's good, because a PR is not a list of the 10 best players in the state; it is a list of players who perform the best in tournament. If you see the list and say, "it's sort of obvious that #8 is more skilled than #5" this sort of problem will be solve itself, unless the player ranked 8th continues to place worse than the player ranked 5th... and if that is the case, is that player really more skilled? The beauty of the system: the more accurate the seeding (panelists who are knowledgeable and fair determine this), the closer the PR (determined by placings/attendance) will be to a top 10 list of our most skilled players (which will eventually make seeding insanely easy). And as time goes on, there will be so much data that factoring in attendance becomes less complicated (not that it is to begin with), and rankings will become more solidified (in the sense that there won't be any ties and thus the data shows a sharper contrast between players ranked next to each other, not in the sense that it would be harder for newcomers to attain ranking... all they have to do is place well).

Would anyone care to give me their thoughts on this matter? Take your time.

 

[Juno McGrath]

What are you, Karn?

 

[lord karn]

Would anyone care to give me their thoughts on this matter? Take your time.

It all depends on two things. Whether PRs should be based off of tournament performance, and whether or not placing shows tournament performance.

If you define tournament performance as simply who placed the highest (or who made the most money), then placement is the best thing to go off of. If this is how you define tournament performance, then I just don't think it is what PRs should be based off of. I don't think it should be based off of this pretty much for the reasons in that big post of mine Dorsey quoted. It creates a really uphill battle to become power-ranked, which is ultimately unfair to everyone in the tournament.

 

[-ACE-]

The former of those 2 stipulations is actually in the mission statement of the NC PR's. The latter involves opinion. I believe consistently placing higher than someone in tournament means you perform better than that person in tournament, lol. Is virtually impossible to cruise through an easy bracket to a high placing EVERY TOURNAMENT. That is literally the only way it could go wrong. And if that were to happen, the problem lies within the seeding aka the panelists. I also don't see how this system automatically makes it hard for new people to be ranked. All they have to do is place well.

 

[Bl@ckChris]

losers brackets are funny. they end up determining placements, even though seeding doesn't really affect them as much directly. i think thats where part of the issue lies. but idk lol.

 

[-ACE-]

No, you're right Chris. That's the major flaw of standard brackets. Round robin, although too much of a pain to carry out, would provide a virtually unabridged data set.

 

[Bl@ckChris]

or we could just seed a single elimination bracket, and then seeding -> placement -> skill would be a slightly more direct correlation.

i think.

but i don't think that's what we want.

 

[-ACE-]

Single elimination takes away too much from the "skill gets rewarded" aspect of tournament play. Similar to comparing a bo5 set vs a bo1.

 

[Bl@ckChris]

yeah, but i still think part of the issue with placement is that when i think about it, only the winners bracket is really seeded. i do think it's very possible to have some standout wins and still place badly. and i think it's possible to get a really easy losers bracket sometimes. because what matters for placement is the losers bracket, but what matters for seeding is the winners bracket.

i guess some seeding in the winners bracket get taken to the losers side, but idk, sometimes people get slapped to really weird places from one side to the next lol.

 

[lord karn]

The former of those 2 stipulations is actually in the mission statement of the NC PR's. The latter involves opinion. I believe consistently placing higher than someone in tournament means you perform better than that person in tournament, lol. Is virtually impossible to cruise through an easy bracket to a high placing EVERY TOURNAMENT. That is literally the only way it could go wrong. And if that were to happen, the problem lies within the seeding aka the panelists. I also don't see how this system automatically makes it hard for new people to be ranked. All they have to do is place well.

Well, what I'm defining as tournament performance is how well players do vs. each other. I suppose we could be more precise. Honestly, I don't think I've read the first page in a long time.

And it definitely makes it harder to get ranked. You start with a low seed because you are new, you play good people first, you lose, and you get a really low placement. Then you're ranked low again. . . and only get on the rankings after you've beaten good people multiple times (and the people on the rankings before didn't necessarily).

In the current system, as soon as you beat someone good you have a shot at getting PR'd, making the vicious cycle a lot less vicious.

The alternative is to not seed based off of the PRs. If that is the case, then I think we would need a secondary system for how to seed players, or else it will just be completely up to whoever is seeding and will be very biased. And that is what my system is. . .

 

[DJRome]

maybe smash should just be swiss. no hype

 

[-ACE-]

yeah, but i still think part of the issue with placement is that when i think about it, only the winners bracket is really seeded. i do think it's very possible to have some standout wins and still place badly. and i think it's possible to get a really easy losers bracket sometimes. because what matters for placement is the losers bracket, but what matters for seeding is the winners bracket.

i guess some seeding in the winners bracket get taken to the losers side, but idk, sometimes people get slapped to really weird places from one side to the next lol.

Of course only the winners bracket is seeded... Again, that is the major flaw of placing off a standard DE bracket. However, that will not happen often enough to matter. The people who consistently place higher WILL have notable wins. They have achieved their consistent placing through thick and thin, both easy and hard brackets. There would just be to much data for one strange instance to have a significant effect. Assuming no upsets, with good seeding, the people in the losers bracket were expected to lose. So if you get upset early on you shouldn't have to play loZr second round losers. You can still place well. I'd there are a ton of upsets, it might be time to adjust seeding.

Well, what I'm defining as tournament performance is how well players do vs. each other. I suppose we could be more precise. Honestly, I don't think I've read the first page in a long time.

And it definitely makes it harder to get ranked. You start with a low seed because you are new, you play good people first, you lose, and you get a really low placement. Then you're ranked low again. . . and only get on the rankings after you've beaten good people multiple times (and the people on the rankings before didn't necessarily).

In the current system, as soon as you beat someone good you have a shot at getting PR'd, making the vicious cycle a lot less vicious.

The alternative is to not seed based off of the PRs. If that is the case, then I think we would need a secondary system for how to seed players, or else it will just be completely up to whoever is seeding and will be very biased. And that is what my system is. . .

You should have a low seed if you are new. You will play PP first, and start from losers bracket basically. But as I started above, with proper seeding, there should be some relatively beatable players in losers bracket. If you are not good enough to defeat them, EVER, should you be ranked? Keep in mind, YOU determine the seeding, not the PR. You can give someone a higher seed if you feel they've proven themselves somewhat (it will get easier for them). That does not, however, warrant an immediate ranking imo.

 

[-ACE-]

8pm is pretty early for this thread to die lol. Does anyone finally understand how this is actually a great system? I kept my mouth shut the other times Dorsey brought this up since I'm sure many of you would assume I'm only agreeing with him since he's my brother. Hopefully I explained it clearly enough.

 

[Dorsey]

Need to fully read this last page... but...

If loser brackets are done correctly(keep in mind, the WB seeding must be proportionally correct), then it seeds itself out just fine. LB does nothing but help things out, it seeds itself if you actually place peeps to their corresponding loser's match properly lol.

I've seeded tons of brackets... I'll always come up with instances like, ahh.. this new school guy is pretty good but true seeding places him against a very defensive player who will probably win but has far less individual talent(referring to an fps)..... but that's just how it goes, w/e. Smash is even more crazy in the seeding aspect, because a "seeding value" that I mentioned earlier considers match-ups(MUs involved in the data used, anyway) but over-all / generally... So technically, due to the nature of smash and assuming a PERFECTLY seeded bracket, upsets are still quite plausible due to match-ups, as well as the range of MU knowledge v. a lack of MU knowledge, for a particular match-up.

anyway... I meant to post that last night but passed out. Still haven't read this last page, hitting the road, see you later today Dave. PAYCE..

happy holidays to all

 

[Bl@ckChris]

is there any way to factor the strength of a tournament, or does getting 5th every time mean the same thing? and it still seems weird to have one method for seeding and another method for a PR...

i would still say that beating $mike in winners but somehow getting 9th still would be better than someone beating stingers and foxy on the way to 5th or something like that. idk how a bracket would make that happen, but somehow judging PR's based on placement seems to put more weight on the losers bracket than the winners...

i mean, we could do a point system like what floridas doing with their thing, and i think that's what stingers is having anyway with the NCSU's, and they look pretty similar to our PR's anyway...

 

[-ACE-]

I knew that was still an issue (weight of a tournament). I haven't figured it all out yet, but it would have to be some sort of point system where players like yay, loZr, dop, mike, etc being in attendance would make the tournament worth more. Individual players would have some sort of point value themselves which would contribute to this.

PR's are strictly tournament performance. Seeding is panelists filling in the gaps that the proposed system cannot fill, as no system is perfect due to tournament setup, varying difficulty in matchups, and no way to measure "true skill".

so, for example, if I enter a tournament where everyone else just happens to play pichu, Yoshi, Kirby, roy, etc. even if they are better players than me overall in the eyes of the panelists, I should have a higher seed than normal since I'm probably going to **** on nearly all of them with Ganon. Otherwise, I will be taking out the best players first and GF's will be a lot less hype. I would seed yay as #1 in a tournament of all ganons even though I think linguini and kage are more skilled. Just my opinion.

 

[Bl@ckChris]

so if the panelists fill in the gaps for seeding, and the panelists use the PR for seeding, but the problem we seem to have is the panelists filling in the gaps for the PR, then can't we just skip the step of having assigning weird points to people and have things done the way they are?

it seems like a weird way to solve the problem of the OMG SUBJECTIVITY factor if it's being taken into account for seeding anyway...

 

[-ACE-]

No. THE PANELISTS DO THE SEEDING. what I meant before was that they have the option to use the placings as a rough guideline. And like I said before, in time, the list will become more and more like a list of our "most skilled" players, and that its what will give the panelists a head start in seeding.

You should seed, imo, according to the probability of the outcome of that particular tournament. Aka, it HAS to be subjective.

 

[Bl@ckChris]

this still seems like a "which came first, the chicken or the egg" kind of argument...

to my understanding, the whole problem with the current way the PR is done is because it's too subjective. the fact that we don't have hard numbers or whatever makes things ambiguous or something. but if we do seeding by this same idea, and the panelists are the same ones who do this seeding, but the "power rankings" are done by numbers from the results based off the subjective (or biased if you don't like it ) seeding, then i don't really see what changes.

 

[lord karn]

The PRs are not subjective. -_-

Dave is suggesting something where the idea of what a good seed is what the outcome of the tournament is most likely to be. We currently seed tournaments based off of who has done the best vs. others in the previous ranking period. In Dave's system, there is subjectivity, because the seeders will have to make the call of who would win in a match between PP's kirby and his own ganon. This IS a judgment call. My system reserves judgment and merely presents data.

I also happen to think that if we had enough data, our system would also pretty accurately predict tournament results, outside of players getting better between tournaments. But, in my opinion, we should not take a player's improvement into account in the seeding of a tournament until they have actually proved themselves in tournament (which Dave agrees with).

 

[Dark Hart]

8pm is pretty early for this thread to die lol. Does anyone finally understand how this is actually a great system? I kept my mouth shut the other times Dorsey brought this up since I'm sure many of you would assume I'm only agreeing with him since he's my brother. Hopefully I explained it clearly enough.

Dude

It's Christmas lol

 

[-ACE-]

I think you should re read some of my posts, answering that will cause me to repeat concise answers to similar questions.

I feel you karn.

 

[Bl@ckChris]

...nvm, i'll just keep playing this game, placing 5th at small tournaments and 7th at larger ones, and whatever happens from that happens.

 

[-ACE-]

Don't be spiteful, this is just a serious suggestion and I want input. This is nowhere near a chicken/egg argument. Well maybe I'm wrong, because clearly the chicken came first. Lol. How does evolution create an egg? Haha. I can see a chicken evolving and eventually laying eggs after millions of years... Lol /tangent

The subjectivity will be minimized as data is collected (depending on how the panelists want to run things, they will have tons of data at their disposal.

 

[Bl@ckChris]

i'm not being spiteful, i just sincerely don't understand enough about what the new system would entail, and how things would change.

i don't see where the problem lies in using who beats who instead of placement, because placement seems like such a fickle thing between tournaments. while i understand that there should be a direct correlation between placing and skill, and i believe there is for the most part, i still find it hard to imagine the ranking being based off of (5th, 5th, 7th) vs (4th, 9th) vs (2nd, 3rd) vs (3rd, 5th, 9th) etc in a ranking period.

 

[Dorsey]

I'll explain Alllll of this when I have a keyboard. A big lol @ ewb having the "same seed" as Josh. And no, I'm not laughing at a difference in skill level.

But yeah, I'll post on all of this when I have a keyboard(I have answers for everything, look forward to it)

 

[$mike]

karn's the homie

 

[Dorsey]

Alright, primarily concerning my seeding comment(snap's comment about ewb and josh having the same seed)...

No one has the same seed. Unless(example) you are comparing seeds 8 and 9 of a 16 man bracket(or similarly proportioned to the amount of participants.... or comparing seeds 4 and 5 from an 8 man bracket, same difference) then why in the **** would 2 people have the same seed. You have got to be ****ing kidding me........ I mean wtf. It's inconceivable to me. I make brackets all the ****ing time under more watchful eye than the nc melee pr thread lmfao. Even my drunk *** can't even get this ****(I am very drunk... x-mas eve activities). So let me be clear here....... NO ONE HAS THE SAME SEED. NO ONE. What is this seeding 1-4 value **** or whatever bull**** I've seen at tournaments? I've seen that stupid horse **** at bill fests and that's the only way something like the ewb v. josh thing could happen. The seeding is BULL****. I repeat, no one should be considered the "same" in a seeding value. It's a ****ing bracket. You seed the bracket(everyone has a different seed) then it plays out. The same concept applies to pools...... which may I remind you, this same concept was NOT applied to the most recent NCSU tourney pools(you were also given the 2nd best #1 seed, which was deserved by boss for out performing you in pools, and that's not even considering what whack *** pools they were).

And yeah, that 1-4 seeding billfest value **** should stop. I may be wrong about the 1-4 part(it might be 1-3 or 1-5, lmfaooaoaoaooaa), but I've seen it with my own eyes(people having the same "seed"/numeric value going into the tournament) attending these tournaments. I've seen myself and mohr been given a greater seeding value than dark hart and players alike(whether we're good at melee or not, still had no previous tourney experience). Karn, to say that your methods aren't subjective in comparison, let me just say that you just need to lay off the crack pipe buddy. As a panelist, you also need to step up and not let TREMENDOUS seeding discrepancies happen like they did in your favor at NCSU 4. WTF...All I gotta say.

Seeding should change after every tournament. They should NOT be based off the PR's, although the PR is essentially a good seeding system.. but it should fluctuate after every tournament, instead it just stays the same after 4 tournaments go by all the while pp and karn try to make sense of it all when the seeding for all tournaments wasn't even absolute.

And karn man... I can't believe you said your system could come close to accurately predicting tournament results. I mean....... idk what to say. In perfect competition, seeds in smash bros. are based off who does the best v. who OVER-ALL... How on earth could your system, or any system, predict a bracket in regard to melee character match-ups? Gotta be kidding me man... I mean take me for example; I'd get a horrible seed but if you put me up against a ganon main I am going to eat his *** for breakfast(smirking at chris right now).

Ranking = who does the best in tournament

NC PR = a rating/poll(good for seeding purposes, but seeding purposes demand more frequency than every 3-4 months to actually be accurate)

alright... I need to get another beer. I don't even remember what I just typed... but I will read later and add, just you wait NC.

$mike's the homie

--

k, Now I'm adding.

Seeding should be based on who v. who(what the PR is). Players should have a seeding value. A system should compare wins and losses in terms of one seeding value(x) v. another one (y)... the range(difference) between these seeding values would dictate the fluctuation of the future seeding changes. That way when the panel actually analyzes this "who v. who" crap they'll actually have something to go by. A lot of it is easy calls, but the panel is indeed being subjective by not doing what I'm saying, by the book and for everyone to see, whether they claim to have similar motives or not.

The process is self-sufficient, as I've said. As for the "strength" of the tournament(chris's comment), that could be easily factored in by the sum of the individual seeding values I discussed.... this would simultaneously account for the "strength" of the tournament(how much it's weighted in comparison to others) as well as attendance.

 

[RaphaelRobo]

Why don't we modify the ELO system and use it for smash? That way it would be updated after every tournament, and give us a decent idea of who is good against who? You get more points for beating a better player, less for beating a worse player, etc

 

[stingers]

Modified elo is what my ncsu tournaments ranking thing uses...I think

 

[Dorsey]

Yep, ELO(from what I've seen in some FL rankings) seem to parallel what I'm talking about with the seeding portion of my post.

my point that our rankings should equate to who performs the best in tournament, depends on the use of good seeding(assuming ELO is in this category).

And chris... your comment about comparing ties with 7th 5th and 13th placings, or whatever... since the tournaments are weighted with specific values as well as the individuals(in regard to how much their place even matters), there probably wouldn't be any ties man.

edit: Yeah, I noticed your ncsu standings looked like the ELO ones. Very good **** actually publishing that data for people to see Alex....

 

[RaphaelRobo]

If there's support for it, I'm happy to run an ELO list for every tournament that takes place in NC. All I'd need is a few days to write a program (it would take a few days because I'm probably not going to write it today or tomorrow), and then I'd need to look back at some NC tournaments and add the results from those. Stingers' formula would probably be good for it.

EDIT: Thinking about it, I could also write a program that automatically seeds pools based on the ELO rankings, updates the rankings after each game, and automatically seeds brackets based on the pools results and the ELO rankings. If there aren't any pools, then it just seeds based on ELO rankings. I'd also write it in Java, so people aren't restricted to one OS. In that case, it wouldn't be too difficult to create an android version of the program.

 

[bossa nova ♪]

ELO seems like some virgin *** **** imo

 

[Juno McGrath]

ELO seems like some virgin *** **** imo

 

[ZoSo]

Dylan, where do you find all these random gifs

maybe smash should just be swiss. no hype

I think you're joking, but Team Ben did swiss for singles and doubles at Gettin Schooled 2 (2005 yeah I'm old school) and imo it's a great format. The main issue is that it's kinda time consuming.