Fighting Games Weekly | Nov 4-10 | I-No what that finger smells like [18+]
- Thread starter enzo_gt
- Start date
poopninjamvc3mk
I sucked six dicks to get this tag.
FGCgaf forget those fruity ass top 50 tier lists. GET in here for TheARTchives:
http://www.twitch.tv/nycfurby
http://www.twitch.tv/nycfurby
Passive Assassin
Member
but triforce is talkin.
MarkMan
loves Arcade Sticks
Put together a short video that shows a more visual representation of Korean levers and their strengths... - http://www.youtube.com/watch?v=5w3ldd5n1Sk
The Last Emperor
Member
Dahbomb, wanna make a serious list on the weekend or something?
Professor Beef
Banned
Who gives a shit
Passive Assassin
Member
460 people do.
ThatCrazyGuy
Member
Battops are best, brother.
poopninjamvc3mk
I sucked six dicks to get this tag.
Dat competitive Japan SFA3 on the ARTchives <3
KOF has 50 players??
Professor Beef
Banned
Well those 460 people are idiots.
Lost Fragment
Obsessed with 4chan
Dear Santahaunts,
Gouken should be able to FADC parry into U2 midscreen.
Thx,
-Me
Gouken should be able to FADC parry into U2 midscreen.
Thx,
-Me
DryEyeRelief
Member
At this point, I believe the FGC has been played by KI "leaks".
Not mad at it or anything, but at some point, you question how consistent something is "leaked" and it isn't unfeasible to think that it's more than poor management.
Please tell me you're kidding.
I can only speak for NYC, but you either know little about public transportation here or choose to ignore it to make that statement (and lump NYC into your generalization). Subways and buses are absolutely vital to this city, and not just for the poor (tons of six-figure salary people take the subway).
That's my routine. I respect that. Feels great when you get that workout in at the start of the day.
The notes are a lie.
Well, EMP
The only way I can begin to see something like this having any validity is by using a strength-of-schedule type formula. Everyone starts with the same base score, and then you start calculating as wins/losses tally up. It can be an ongoing thing.
While SOS sucks for determining actual playoff matchups (tournaments, for example, should not use this to determine who plays in grand finals, look what it did for the BCS) it's great at looking at performance within a group when dealing with a diversity of opponents.
Yep. I thought of running a Twitter poll but then realized that it would be just as unquantifiable as the EH rankings and decided against it (even though it wouldn't be a black box effort like the EH list).
sixteen-bit
Member
reminds me of the ai in harder difficulties on SF4 who will always ultra when you press a button
SousedLouse
Member
I'm getting the impression that if these lists were called popularity contests outright, nothing would change.
Or how the AI in Super Turbo read your inputs. I think that devs could set up A.I. that's even more frustrating than that, but they (thankfully) choose not to, so they can avoid scaring off single-player fans.
Or how the AI in Super Turbo read your inputs. I think that devs could set up A.I. that's even more frustrating than that, but they (thankfully) choose not to, so they can avoid scaring off single-player fans.
I'd want one of those battop sticks.
So, I'm kind of sick of not having a statistical basis for the rankings, and I'd like to do something about it.
That said, the first step would be to start gathering brackets for majors for a game (I don't care which game, as long as it's a fighting game).
I'll start looking at the formats that the brackets are in, and importing them into a common win/loss format, as well as normalizing the names.
Once that's done, I can start applying any ranking system to the data.
You can tweet links to me on twitter (@OneFrameLink) or link them to me here.
That said, the first step would be to start gathering brackets for majors for a game (I don't care which game, as long as it's a fighting game).
I'll start looking at the formats that the brackets are in, and importing them into a common win/loss format, as well as normalizing the names.
Once that's done, I can start applying any ranking system to the data.
You can tweet links to me on twitter (@OneFrameLink) or link them to me here.
DryEyeRelief
Member
Why don't they just rank each other based on Evo seed points?
Not comprehensive enough, as there were only five Road to Evo events this year (which offered seeds, not counting the international tournaments, even then, it's not that much).
That, and you're limiting yourself to a small subset of actual results. Why wouldn't you include say, Youmacon from the past weekend? The idea is that the more data you have, the better you can track how that player does against the field.
Has this translation to Daigo's interview regarding Matchup Theory been posted? Awesome read. As long as I play fighting games, I'll be a Daigo fanboy. Credibility be damned.
http://www.eventhubs.com/news/2013/...-details-how-he-approached-ryu-akuma-matchup/
http://www.eventhubs.com/news/2013/...-details-how-he-approached-ryu-akuma-matchup/
If we're counting Blood On The Sand, then Cent should be #1.
Mortal Kombat 2 style. Uppercut? Insta-duck into uppercut. Press button? I'm going to stay right out of range, then walk forward and throw before you can get anything else. Try to throw me?? LOL no. Just keep throwing projectiles and jump back kicking as I walk into it until you win.
Vigilant Gambit
Member
Fighters Spirit sponsors Flocker. I dunno what extent Triforce plays a part in it, but from what I understand, Mike Tokles from central Florida is the dude who sent Flocker to Evo this year. Probably also Youmacon.
This would also be in my top ten.
ipitydatfu
Member
My top 50 list
1. 30 dollar Italian meal
2.
1. 30 dollar Italian meal
2.
That's part of the joke (of striking out "sponsor"); every time that someone refers to EMP as a sponsor, Triforce says they are a "player development" organization.
I can't recall if he's ever elaborated on what exactly that is though.
DryEyeRelief
Member
How do you qualify results from smaller tournaments to bigger ones? Is getting top 16 at Civil War the same as getting it at Final Round?
"God's Beard!"
Member
...that's not even a joke.
What you should have said is:
My top 30 list
- The dollars I spent on an Italian Meal
From the point of Elo, it doesn't matter. The size/setting of the tournament doesn't matter. It assumes the mean value of a player's performance doesn't change dramatically over time (from Wikipedia):
This is the important part:
Whether it's Elo or not doesn't really matter. What matters is that currently, it's not unreasonable to assume that we can capture and collect the raw underpinnings of any system, wins or losses.
That's really the value-add that I'm proposing here; once I'm able to capture wins/losses between players, we can use any method we want to slice and dice the results (Elo, or some other ranking system, hell, we can use multiple).
And that's where I need people's help. Capturing the bracket data (because all wins and losses matter) to have so that we can start asking the questions across the entire set of data.
Esteban (from HBTB) suggested that I start capturing from NEC 2014 going forward. This is probably much easier on me and can be retrofitted to go back in time if necessary.
Again, the focus is on the data collection and cleanup (not everyone will put people in the bracket the same way, mapping to the same names, etc). It's not a difficult task, I just need the data sources.
irfaanator
Banned
The number of points you get/lose is based on who you played correct? So yea bracket wise it's going to be a challenge w/o help from the T.O.s. Plus do you need to count for uneven # of matches played (winners vs losers bracket) or losing to unknown players (e.g. cloud805) who would probably have the initial 1400 rating but should be at a higher rating?
For Elo, yes. However, for the first round of a tournament that the system keeps track of everyone is going to start off with the same amount of points.
Working on that for NEC 2014. People running paper brackets is going to be hardest. Paper brackets are great for running tournaments, but horrible for anything before or after-the-fact.
However, that's not something that can't be overcome, just requires manual entry. This can be done in part through crowdsourcing (even a few people could tackle the paper brackets of a major fairly quickly).
For Elo specifically, you might, but there are other systems which take into account the number of matches played (performance rating, for example).
But that's not really important. Capturing who plays whom, as well as the outcome, that's the prize.
Everything else is easy pickings (it's just an algorithm run over all the data) and can be added to/modified/replaced with another side-by-side on a whim.
sixteen-bit
Member
[QUOTE="God's Beard!";88726859]...that's not even a joke.
What you should have said is:
My top 30 list
GB, are you an illustrator?
What you should have said is:
My top 30 list
- The dollars I spent on an Italian Meal
GB, are you an illustrator?
irfaanator
Banned
Yea definitely. Getting the data is the hardest part. Analyzing is going to be easy
We put all the completed brackets from UFGT9 online in PDF form if you want somewhere to start.
http://ufgtus.wordpress.com/2013/06/05/all-the-completed-brackets-from-ufgt9-online-for-you-to-see/
Evo also has completed brackets in electronic form online. However, Evo's brackets may contain some inaccuracies, since their online app assumes every volunteer ran their bracket correctly. Also, they don't list the names of emergency registrants. (Example: "Bye" in this bracket is Flux.)
2013: http://evo2013.s3.amazonaws.com/brackets/index.html
2012: http://evo2012.s3.amazonaws.com/brackets/index.html
DryEyeRelief
Member
Marn vs Player A,B,C if he wants
According to Floe and Champ, Marn only has a
DryEyeRelief
Member
NerdJosh officially a socal player.
Lazy Nerdjosh doesn't travel enough to the East Coast.
Lazy Nerdjosh doesn't travel enough to the East Coast.
Parallax Scroll
Banned
sounds cool
sixteen-bit
Member
I missed this. Is there an archive.
This please. I need something to hear while working
MiniBossBattle
Member
I want Paradigm, MiniBoss and DrSnakeTits to elevate their game even higher. They are some of the most exciting players on TRB.
Parallax Scroll
Banned
Nah man people would get too salty.
CasperOne has the right idea in that if you want to make a list based on results then you have to do it systematically. I like the system he is trying to come up with where you are ranked based on who you have played and beaten in the past. This allows players like Flocker to score well who even though don't travel as much perform well when they do get to travel to tournaments.
However the stats needed to accurately score players are enormous and we don't really have that information anymore. You would have to have match records for about a thousand people to have an accurate ranking in this system.
Another way to do it would be to simply use what that other site used where they rank based on how they performed at particular events and each event gave different scores based on number of entrants. So Flocker winning EVO would be worth more points than ChrisG winning Yomoucon for example. This is probably the easiest way to determine rank based on tournament performance but it heavily favors players who go to tournaments a lot (like OmgitzAndre).
Snuggle fabric softener