What's new

CrimsonShadow Introduces MKi - New MK9 Player Rankings

CrimsonShadow

Administrator and Community Engineer
Administrator
CrimsonShadow introduces new ranking system solely for Mortal Kombat and soon to be Injustice. Full detail on how the system works can be found below. Check it out and give Crimson your feedback! ~ STORMS

"Ok, so here's the rundown (the About page explains some of it in detail):

As of now, the system ranks people from a bunch of major tournaments, based only on their wins/losses to other players. That allows it to be more accurate, especially with the top players, than previous ranking systems, including NFG. Also, it allows you to see the path someone took to archive the rating -- you can click on a player, and see his rating history based on W's/L's.

The next step, after cleaning up/prettifying the names, and filling out the rest of the pages (Head to Head, Contact etc.) is to build a simple public seeding system on top of this. So you'll be able to input a list of players, and a number of pools, click a button, and have a mock touney seeded for you.

Introducing Mki - And with it a new set of player ratings, based


100% on wins/losses vs. other players in tournament.

Data was crunched from the following tournaments:
EVO, MLG (All 4 from 2012), and NEC XII (SCR will be added this weekend)

http://mkirankings.com

Here are the top 25:

1. REO
2. Maxter
3. CD Jr.
4. Perfect Legend
5. Pig of the Hut
6. 16 Bit
7. Dizzy
8. MIT
9. Detroit
10. Crazy Dominican

11. Tyrant
12. m2Dave
13. Denzell
14. Showtime
15. Riu 48

16. PBoard
17. Forever King
18. Curbolicious
19. Sonic Fox
20. Wafflez

21. Tom Brady
22. Xblades
23. Krayzie
24. Sup
25. Winter Warz

What is this?
It's not a skill ranking, nor a "top8's" ranking; it's a rating of wins/losses vs. other players in tournament.

Which Tournaments Does This Include?
So far it includes EVO, the 4 MLG events from 2012 (Columbus, Anaheim, Raleigh and Dallas), and NEC.

Why These Tournaments?
For two reasons: 1) It was possible to get complete bracket data or Tio files for these tournaments, giving an official record of every win and loss, and 2) These are large tournaments with a good representation of top players from more than one region. As more large tournaments happen this year, they'll be added as long as the full bracket is available.

How Are Players Ranked?
Players are ranked solely by their wins and losses against other players. The formula used to calculate the totals is decades old, and is used for everything from chess to professional ball sports to Starcraft and LoL.

Why is This Important?
Other methods of ranking have significant flaws. Assigning points based on the size of the tournament, the amount of money awarded (really!), randomly assigning much higher point levels to certain tournaments, etc. Also things like only taking a player's top X performances (meaning that if a player has 3 terrible tournaments in a row, it might not affect their rating at all), while superstars who have more than X great performances in a 6-month period are penalized because of the limit.

Also other systems fail to rate players outside the top 3-4 or top 8 for many tournaments, meaning that players who have played consistently well, yet placed just outside the top 4/8 receive no points at all.

By calculating ratings based on player vs. player performance, none of these extra arbitrary "tweaks" are needed, personal opinion is removed, and the results are based 100% on record and math.

It's easy, unriggable, and consistent across the board. And it allows us to rank more than 10-20 players."

Why is This Needed?
Nearly every legitimate sport has a fact/performance-based system for seeding tournaments. Video game-based sports are quickly staking a claim to being legitimate sports as well -- and as dollar amounts and viewership rise, we need to be more careful, less biased, and more factual about how we rate/seed our players.

Also, because not everyone can remember everything that's happened in the past; so it's crucial to have a place where TOs, commentators, and fans can find accurate data about past performance.

Again, this is not an absolute measure of player skill or likeability; it's simply a numeric representation of relative tournament performance.

What Does an Event Need to be Ranked?
In order to be fair, an effort has been made to not include tournaments which feature one region of players almost exclusively. This means no locals, team tournaments etc. If an event meets the qualifications to be ranked, the Tio bracket files (or a complete Challonge bracket) are all that is needed to update the database.

This means that updates are quick, painless, and incredibly fast.

What Does a Player Need to be Ranked?
Players are ranked after they've completed more than 5 matches in ranked tournaments. This makes sure that enough data is available for the rating to mean something.

How Do I Improve My Rating?
Continue to attend events, and do better consistently against better players.

I'd like to thank everyone who helped with this process, including Pig Of The Hut, 9.95 Phil, Tolkeen, EGP Wonder_Chef, and everyone who helped post or collect tournament/player data and info. In the coming weeks I'll be completing the other sections of the site, fixing any errors, and building/posting a seeding tool (open to the public, after we test it) and continuing to double-check all of the info.

Here's to a better, stronger, and brighter 2013!
 

Attachments

Tolkeen

/wrists
This is amazing Crimson, anyone who disagrees needs to go out and get the bracket data from the missing tournaments before just flaming *cough* me *cough*. You're always doing amazing things for the community, thank you!
 

ecilA

Noob
How Are Players Ranked?
Players are ranked solely by their wins and losses against other players. The formula used to calculate the totals is decades old, and is used for everything from chess to professional ball sports to Starcraft and LOL.
What do you mean solely by their wins and losses against other players?
 

Tolkeen

/wrists
What do you mean solely by their wins and losses against other players? Shouldn't where you place at BIG majors be the measuring stick?
Yes and No, big majors are counted, but if I went to ufgt9 and bodied 5 people who have never played before to make top 16, is it more impressive than gga dizzy beating pl, 16bit, and pig of the hut to get to top 16?

*this is all theoretical btw*
 

CrimsonShadow

Administrator and Community Engineer
Administrator
What do you mean solely by their wins and losses against other players?
So basically, you rating goes up/down depending on whether you won or lost. And the amount it changes, depends on your opponent's rating.
 

ecilA

Noob
Yes and No, big majors are counted, but if I went to ufgt9 and bodied 5 people who have never played before to make top 16, is it more impressive than gga dizzy beating pl, 16bit, and pig of the hut to get to top 16?

*this is all theoretical btw*
I'm talking about BIG majors like MLG, NEC, etc.. Placing top 8 at a major with over 100 people at it usually means you did not just body 6 no name scrubs who never played MK before to place top 8. At smaller tournaments I can understand using the who you beat method but not at the super big majors. Also, making a ranking based on 2012 but only count the events you can verify who beat who at is wrong because its an incomplete ranking list and not really accurate for 2012.

I JUST finished listening to the KTP award podcast, a few of the nominees and even the winner for tournament match of the year were at events that are not even counted in this ranking list, yet they all include big name vs big name. Look at Big D, he beat CDjr at Final Round 2012 and beat Maxter at CEO 2012 but those wins are not counted. 2 BIG wins vs players ranked that high and its not counted because the full "who beat who" results cannot be verified. Unfair system for the current ranks based off of 2012 if you ask me.
 

Tolkeen

/wrists
I'm talking about BIG majors like MLG, NEC, etc.. Placing top 8 at a major with over 100 people at it usually means you did not just body 6 no name scrubs who never played MK before to place top 8. At smaller tournaments I can understand using the who you beat method but not at the super big majors. Also, making a ranking based on 2012 but only count the events you can verify who beat who at is wrong because its an incomplete ranking list and not really accurate for 2012.

I JUST finished listening to the KTP award podcast, a few of the nominees and even the winner for tournament match of the year were at events that are not even counted in this ranking list. yet they all include big name vs big name. Look at Big D, he beat CDjr at Final Round 2012 and beat Maxter at CEO 2012 but those wins are not counted. 2 BIG wins vs players ranked that high and its not counted because the full "who beat who" results cannot be verified. Unfair system for the current ranks based off of 2012 if you ask me.
As opposed to the NFG ranking which is an unfair ranking system based off of 2012 if you ask me, the difference is this is a work in progress, as it is stated on EVERY page of the website. Read, then reply! It's based on person vs person, because beating a higher ranked player is a better show of skill than beating a lower ranked player. Not everyone at big tournaments is a threat to a decent player, and as such, you shouldn't get a million points for beating me, and you on a luck draw, when you'll only get a hundred points for beating gga soonk (who is a much better player). It's not a very difficult concept, and its meant to eventually replace a fundamentally broken system, and like my first post said. If you want a tournament included, you have to go track down the tournament bracket, so feel free to do that before you raise any more complaints about a generally great idea, meant to simplify things for the community.
 

9.95

Champion
My major input into this was that pure math won't allow for the randomness that is generated in brackets after the seeding.

If nothing else, math is predictable... it has to be. Math always works the same way which is why it's considered a universal language.

Because of that type of predictability, I brought this up to CrimsonShadow. My concern was that the VERY REASON we separate top seeds from each other, as well as players from the same regions and teams is because none of these players travel all that distance to play the exact same players they play all the time.

The math can dictate a very predictable bracket that would force us to see and use the exact same bracket at EVERY TOURNAMENT without some randomness factored in. It would predict that the players traveling all that distance would have to face the same players(regardless of seed, region or team) because the math puts them in the bracket that way.

What this list is good for, as with NFG, is simply giving us easy access to a way to seed the top 4, 8, or 16 players at a tournament. Once that part is done, then you separate by team/region and place them in the pools and brackets.
 

CrimsonShadow

Administrator and Community Engineer
Administrator
I'm talking about BIG majors like MLG, NEC, etc.. Placing top 8 at a major with over 100 people at it usually means you did not just body 6 no name scrubs who never played MK before to place top 8. At smaller tournaments I can understand using the who you beat method but not at the super big majors. Also, making a ranking based on 2012 but only count the events you can verify who beat who at is wrong because its an incomplete ranking list and not really accurate for 2012.

I JUST finished listening to the KTP award podcast, a few of the nominees and even the winner for tournament match of the year were at events that are not even counted in this ranking list, yet they all include big name vs big name. Look at Big D, he beat CDjr at Final Round 2012 and beat Maxter at CEO 2012 but those wins are not counted. 2 BIG wins vs players ranked that high and its not counted because the full "who beat who" results cannot be verified. Unfair system for the current ranks based off of 2012 if you ask me.
It's only fair to rank tourments we have complete results for. Otherwise I'd have to ensure that every single win/loss was collected, 100% accurately, from whatever other tournament. Also, in an effort to stay current, the consensus was that we should start from EVO 2012 and move forward. So Columbus is only there to 'seed' everything so that not all players would be starting from 0 at EVO.

Moving forward, we can import tournaments of the appropriate size/player distribution with only the Tio files or Challonge bracket.. So we can use data from more tournaments as they happen. This is only the beginning :)
 

ecilA

Noob
So basically, you rating goes up/down depending on whether you won or lost. And the amount it changes, depends on your opponent's rating.

Big D is not ranked in the top 25? Big D placed top 5 at Final Round 2012 with over 120 players, at this tournament he beat CDjr amongst other top players. At CEO 2012 Big D placed top 3 where he beat Maxter, CEO had about 50 players maybe 60.

You do not count those tournaments because you cant get the full "who beat/lost to who" list, so you just ignore it. In fact, you don't count over half of the major tournaments held in 2012 because you cannot get the full "who beat who" list. This ranking is VERY incomplete and is inaccurate for that reason.
 

CrimsonShadow

Administrator and Community Engineer
Administrator
The math can dictate a very predictable bracket that would force us to see and use the exact same bracket at EVERY TOURNAMENT without some randomness factored in. It would predict that the players traveling all that distance would have to face the same players(regardless of seed, region or team) because the math puts them in the bracket that way.

What this list is good for, as with NFG, is simply giving us easy access to a way to seed the top 4, 8, or 16 players at a tournament. Once that part is done, then you separate by team/region and place them in the pools and brackets.

I will give this a try at Winter Brawl.
Yup, and again, as I'm building the seeding system from scratch, and will run through it with you guys repeatedly and put it through the ringer before it's public, we can make sure that appropriate randomness is taken into account and pools won't be exactly the same every time.


Big D is not ranked in the top 25? Big D placed top 5 at Final Round 2012 with over 120 players, at this tournament he beat CDjr amongst other top players. At CEO 2012 Big D placed top 3 where he beat Maxter, CEO had about 50 players maybe 60.

You do not count those tournaments because you cant get the full "who beat/lost to who" list, so you just ignore it. In fact, you don't count over half of the major tournaments held in 2012 because you cannot get the full "who beat who" list. This ranking is VERY incomplete and is inaccurate for that reason.
I don't think you read my post to you. Final Round 2012 was almost a solid year ago. Columbus is only there because was the best choice to kickstart the system thanks to having the entire bracket available. See the rest of what I wrote for why more tournaments will be included coming up.

Now easy there... Take a deep beath :)
 

BillStickers

Do not touch me again.
Big D is not ranked in the top 25? Big D placed top 5 at Final Round 2012 with over 120 players, at this tournament he beat CDjr amongst other top players. At CEO 2012 Big D placed top 3 where he beat Maxter, CEO had about 50 players maybe 60.

You do not count those tournaments because you cant get the full "who beat/lost to who" list, so you just ignore it. In fact, you don't count over half of the major tournaments held in 2012 because you cannot get the full "who beat who" list. This ranking is VERY incomplete and is inaccurate for that reason.
Hi, new person who is surely not an alt yet has an opinion about bracket completeness:

Just because this system is 'young' does not mean that it's not viable and inaccurate. The system will get more accurate over time as more tournaments are added, and any tournaments without complete info can easily be added later as more information is discovered. Complete results are a definite requirement for this system to work.
 

ecilA

Noob
I think the "who beat who" system is fine, but I just don't think it can work mainly because there is no way that every T.O. is going to log every round of who beat who at every major.
 

kronspik

Mortal
You don't need to log every round, right? It's just about who won each set.

If nothing else, being able to see an individual player's win/loss history is pretty awesome.
You should consider expanding this to more than just NRS games. People from other communities might be interested in a system like this.