Elo Ratings in Competitive Melee

[Vts]

I also beat you at a St. Louis tournament in pools, I think it was HD Remix and it was not long ago whatsoever.

past is past settle it at either genesis or mist depending on when it is alright guys.

i wanna see some **** going on from either of you alright! rivals are good thing to have so temporally rivals are sinister and peef alright guys

now let's get back on topic of how to set these ELOs ranking per state or how to do it nationally

 

[Rubyiris]

swiss is a style where everyone plays for X number of rounds and you play people with the same score as you. generally in HUGE tourneys it ends up with like 2 or 3 people going undefeated.
http://en.wikipedia.org/wiki/Swiss-system_tournament

In swiss, if you have more than one person undefeated at the final round, you MUST extend it to another round or two until only ONE person remains undefeated.

Also for those interested, Arizona's tournaments have been ran using swiss elimination to determine seeding for the past two or three months. Swiss has been attracting new players, and most of the existing players are very fond with the system, and our tournaments have been ending several hours earlier than they had been when we did pools.

 

[Zivilyn Bane]

Ruby, while I definitely like the idea of doing swiss to seed, there is no unwritten law that states that if two players are tied the round must be extended. Especially in chess tournaments where games could be 1-6 hours long and time constraints prevent that. I have definitely seen a blitz match playoff if there were multiple people tied with perfect scores (5 minutes each player). But at the same time I've seen tournaments where they just used standard tie breakers and called it a day. In Melee however, it would be a lot more practical seeing as how with enough TVs you could run a round pretty darn quick.

 

[G. Vice]

Ziv, much props for working so hard on this dude. And good matches a few weeks back, keep working on that marth, hope to play you again sometime soon.

Also, Peef vs Sinister MM better get recorded, I look forward to seeing it.

 

[-Sinister-]

I also beat you at a St. Louis tournament in pools, I think it was HD Remix and it was not long ago whatsoever.

edit: you know what **** it. accept the $50 mm if i suck so bad it should be free $.

 

[crush]

peef would win imo

 

[Zivilyn Bane]

peef would win imo

Oh cool, you've played them both? Because if you haven't your opinion is completely useless and unwanted. Notice Peef has disappeared from this topic the moment Sinister asked for a MM. If I had known that would've got him out of here I would have tried to MM him a long time ago.

Anyways, please no more talk about irrelevant matters in this topic. I don't want this closed because it's gone haywire. So let's keep this about elo guys.

 

[crush]

Oh cool, you've played them both? Because if you haven't your opinion is completely useless and unwanted. Notice Peef has disappeared from this topic the moment Sinister asked for a MM. If I had known that would've got him out of here I would have tried to MM him a long time ago.

Anyways, please no more talk about irrelevant matters in this topic. I don't want this closed because it's gone haywire. So let's keep this about elo guys.

wow i am so sorry. i feel ashamed. how can i make it up to you? want me to send you 5 bucks in the mail?

 

[PEEF!]

wow i am so sorry. i feel ashamed. how can i make it up to you? want me to send you 5 bucks in the mail?

=D

edit: you know what **** it. accept the $50 mm if i suck so bad it should be free $.

I don't think you suck bad at all in IC chat I said that you were really good and i said it here too. I also haven't played once this summer in favor of Starcraft.

I'm broke as s*** but I'd love to $5 MM you somewhere. I would never risk 50 bucks against anyone.

I've been gone from this thread for forever ZivBane, prolly 2 posts per week on smashboards this summer.

 

[-Sinister-]

=D



I don't think you suck bad at all in IC chat I said that you were really good and i said it here too. I also haven't played once this summer in favor of Starcraft.

I'm broke as s*** but I'd love to $5 MM you somewhere. I would never risk 50 bucks against anyone.

I've been gone from this thread for forever ZivBane, prolly 2 posts per week on smashboards this summer.

god damn it. now i seem like the ****. ok i'm sorry for getting on your case and i'll take you up on he $50 mm.

are you coming to MIST?

 

[Zantetsu]

Lol @ so much tension in this thread.

 

[Vts]

god damn it. now i seem like the ****. ok i'm sorry for getting on your case and i'll take you up on he $50 mm.

are you coming to MIST?

I'll $5 mm you for giggles

 

[crush]

Using swiss rounds in melee would make the community grow a lot. You'd start to see many more low leveled players at tournaments and the community would grow a lot, but things would need to be much more organized because you cannot have some people playing their 2nd round of swiss while some people play their 3rd round of swiss.

 

[-Sinister-]

I'll $5 mm you for giggles

i'm game 4 dat

 

[Rubyiris]

Ruby, while I definitely like the idea of doing swiss to seed, there is no unwritten law that states that if two players are tied the round must be extended. Especially in chess tournaments where games could be 1-6 hours long and time constraints prevent that. I have definitely seen a blitz match playoff if there were multiple people tied with perfect scores (5 minutes each player). But at the same time I've seen tournaments where they just used standard tie breakers and called it a day. In Melee however, it would be a lot more practical seeing as how with enough TVs you could run a round pretty darn quick.

In competitive TCGs, this IS a rule. I don't play chess, I have played MTG/Yugioh/Pokemon/Etc competitively since 2000 which is where my Swiss experience comes from. The TOs ALWAYS extend the rounds if one of two things happens (in most cases):

1. If there are more than one player with an undefeated record
2. If there are players with only 1 loss that don't make the play offs (this is subject to change due to time constraints, or if there are so many entrants that this is bound to happen.)

Furthermore, rounds only last 40 minutes. There is a 40 minute time limit, and an end of match procedure to dictate matches that haven't ended by then, so a round will take roughly an hour to complete, which is why these rules are viable.

Remember that there are OTHER games that use Swiss Elimination for seeding. It's not just Chess. :\

Also from my experience with melee, it takes about 20 minutes for a room of 20-24 entrants to run through a round with 5 tvs.

 

[S l o X]

how do i find our my elo with a few tournaments i've been to? o: i've read cobalts post but . . . yea.

 

[hectohertz]

hey y'all, so i can't go into too much detail, but we're working in the backroom on creating a really nice elo system that's going to kick ***

what would be really useful now is if someone could go and collect an archive of .tio files from tournaments, so that when we have the system working we have some data to make rankings from

 

[PEEF!]

hey y'all, so i can't go into too much detail, but we're working in the backroom on creating a really nice elo system that's going to kick ***

what would be really useful now is if someone could go and collect an archive of .tio files from tournaments, so that when we have the system working we have some data to make rankings from

Har har i can't get into detail because the information could be dangerous if leaked to the common public har har now go do work for me commoners

Nobody is stopping you from going into detail.

 

[hectohertz]

Har har i can't get into detail because the information could be dangerous if leaked to the common public har har now go do work for me commoners

Nobody is stopping you from going into detail.

alright, we're building a two-stage elo system (PROGRAMMERS WANTED) :
- a tio plugin that supports fetching/submitting ELO data
- an online database to receive tournament results, subject to approval, and then update the ratings-db

but if you don't know how to program, what would be awesome is if you could help us compile some data so when we have this thing up and running, we can back-log it with tournaments


Better? Sorry I was a little short yesterday...

 

[The Star King]

Nobody is stopping you from going into detail.

...you're usually not supposed to leak information from private rooms without permission. That would kind of defeat the purpose of rooms being closed to the public.

Not saying that I support the existence of Backrooms or anything.

 

[hectohertz]

don't pay attention to peef he's just a troll

anyone interested in helping out collect some archives of tourneys would be awesome though

 

[Nintendude]

Mytummyhertz - I'll make a thread for it in Tourney Discussion once some details are discussed.

 

[PEEF!]

alright, we're building a two-stage elo system (PROGRAMMERS WANTED) :
- a tio plugin that supports fetching/submitting ELO data
- an online database to receive tournament results, subject to approval, and then update the ratings-db

but if you don't know how to program, what would be awesome is if you could help us compile some data so when we have this thing up and running, we can back-log it with tournaments


Better? Sorry I was a little short yesterday...

Cool =D

...you're usually not supposed to leak information from private rooms without permission. That would kind of defeat the purpose of rooms being closed to the public.

We are playing super mario smash yoshi masters and we hide stuff from eachother in secret rooms LOOOOOL

BTW not trolling

 

[Battlecow]

Seriously, what's the point of having a discussion that only the BR can see? What conceivable harm could come from non-BR'ers seeing the discussion?

 

[Nintendude]

It just so happens that the BR is officially working on developing the project. There's not much actual discussion going on there besides just programming the system. You really aren't missing out on much.

 

[G. Vice]

Lol Peef is hilarious.

 

[Rockenos]

As one of the many players who goes to tournaments just to get better and doesn't plan on winning money even locally, I love this idea. There are numerous reasons, the largest one likely being that the $200 bucks I spent to go to CEO seems less idiotic lol.
But also, gas money and entrance fees for locals get me EVEN MORE than just a great tournament experience. Can't stress how much I love this.

 

[sirlaser]

+1

Is there an estimated arrival date?

 

[Cactuar]

2020.

Problem?

**

 

[NeutralDamage]

Great idea, let's go for it!

 

[sirlaser]

2020.

Problem?

**

You remind me of this video game character:

Except he's serious but easily interpreted as sarcastic, while the overall vibe your profile + troll face gives off is sarcastic but may be interpreted as serious

 

[ajajayjay]

Sorry if this has been asked before, but don't all matches count equally in an ELO rating system? regardless of WHEN they were played? As in, do Mew2king's losses from 2006 and 2007 get weighed equally with matches today?

If so, is it possible to set a 1 or 2 year range of influence on which matches affect ELO rating?

 

[TheCrimsonBlur]

2020.

Problem?

**

More like 2012. Get scraped Cactus.

 

[Spyro]

I know this tread is old but wouldn't this system be a little inaccurate for mid / low tier mains?

 

[Strong Badam]

Why? Character selection is unfortunately part of what determines skill in this game, as is anything that can contribute to victory. I see no problem with having a lower Elo than I could simply because of my character choice, much like how I don't have one with placing lower due to my character choice. They go hand in hand, in most cases.

 

[Kal]

Yeah, skill should remain objectively measured as a function of winning. You don't get bonus points for playing a certain way. And, while it surely is possible that a low tier main might be "better" in some way than a player of the same skill who uses a good character, it's incredibly difficult to account for.

It also leads to being forced to try and account for other such differences. "That player doesn't wavedash. Surely he's better than the player with the same Elo who does!"

 

[Spyro]

Would you expect someone to win 5/8 times if they are playing someone that has 100 fewer points than them but when it's Yoshi (guy with +100 points) vs Sheik? That is kind of what pop in my head when I first saw the rating system.

*EDIT*: I get what you guys are saying now. When I posted about the whole yoshi player vs sheik player thing, the more I think about it, the more I think that the rating would be accurate if there is now "bonus" points added (I'm taking about the ratio being accurate here). However I still think that it's not a accurate representation of their skill.

 

[Varist]

It still works. Just differently. With the Elo system, we can almost definitively say who has the best Yoshi by scrolling down the top 100 list and finding the earliest yoshi mains.

When you choose to play a low tier, you're voluntarily taking yourself out of the real competitive race for #1 in the world. And you'll see it all the time. "I just want to be the best Link," or "I want to show people Ness is competitive."

Because that's what they're competing for. Best ____ main. Ask them.

 

[Kal]

Yes, the Elo still accounts for it. We assume that two players are simply playing the game sans restrictions; no one is forcing that player to stick to only Yoshi when his opponent is clearly going to go Sheik. What you're asking for, Spyro, is something totally different: you want a ranking system that puts restrictions on its players (specifically according to the player's "mains").

Skill is an abstract notion, but we can estimate it based on the probability of a player winning. If a player is stupid enough to restrict himself to only Yoshi, his Elo will reflect this by being low. Otherwise, either:

1) Yoshi does not suck
2) He (and the rest of the community, arguably), have not played in enough tournaments, c.f. this post.

 

[N64]

Choosing to play yoshi in tournament is a decision just as choosing when to usmash or how to recover, etc. I don't see a need for any elo padding for 'low tier mains' or anything really. Elo doesn't represent skill, it represents probability of winning.