Why we need point-based ranking system in MMA

zec1234

Brown Belt
@Brown
Joined
Feb 22, 2008
Messages
2,573
Reaction score
0
What I would like to do in this thread is to initiate a meaningful discussion on what kind of ranking system MMA needs.
I will start with a short dissertation and then ask you a few related questions.

Every reputable sport has rankings. Not just sport, but many human activities (like education, employment, finances, ...) have rankings for people and institutions.
So there should be no question if MMA needs rankings.

Question is, do we need the point-based rankings for MMA fighters?
This implies existence of formula and rules that are used to calculate points for each fighter.

There are number of different methods that can be used to create rankings.
Most popular is having a voting panel of experts. Some websites have a panel with one expert, some consist of a few internal experts.
UFC has rankings generated by external voting panel made up of media members. They are asked to vote for who they feel are the top fighters in the UFC. The only dictated placing is champion being automatically at the top position.
Most of the times those experts don't give much explanation on how they created their rankings. Just a simple list from 1 to 10.
Some rankings use polls of non-experts and some use betting markets.
Sometimes we see the power ratings that have extensive textual explanation, but no formula or methodology.

Then there are some systems that combine multiple ranking sources with different ranking approaches. They mix them together and generate some kind of amalgamated score. If that score relies on other voting or poll based rankings, we cannot consider them as real point-based systems. Even if they have some kind of formula that generates score. That is basically just averaging of multiple expert poling / voting rankings.
 
Some self-proclaimed experts and nihilists will instantly say "rankings don't matter".
Most of the time, they are first to respond to the ranking threads. Even though this is not thread for them, they don't like to be ignored. So ignore them.
For that reason I will reserve several posts on the first page t add more material later.

Also, there are some people that are devoted followers of certain fighters and are willing to disrupt any meaningful conversation just to prove their bias point.
This is not for them as well.
 
Why do we need rankings?
It makes sense for a magazines or websites to have their own ranking system. It generates more views and activities on the website.
It is a good conversation topic, generates more interest among MMA community.
Some people use them to show their MMA expertise level and how knowledgeable they are about all aspects of MMA.

That is all nice and dandy, but what is in there for fighters?
Some will say that MMA fighters are glorified prizefighters and they do it just for money. That might be a case for some, but I think that majority of MMA fighters are there because they love the sport and this ultimate way of competing with other individuals.
Some fighters say that they don't care for rankings, but I am pretty sure all of them check different rankings from time to time.
Rankings are very motivating for a fighter. They all like to say "I believe I am in the top 15", or "I deserve to fight somebody from top 10".

For fighters to have trust in rankings, that ranking must be created by a completely objective and unbiased system.
Also we should have one ranking system that replaces countless subjective rankings that are skewed by individual perception, by fighter popularity or by interest of different organizations.

So, we need a formula designed to make the ratings as fair as possible. That formula should use quantitative measures (reliable indicators of fighters performance) proposed by experts. That formula can be used to calculated points on paper, but it is preferred to use computer system.
Computer ranking system affirms objectivity, without giving advantage to a specific player, team, organization, region, or style. Additional advantage of computer systems is that they can accurately evaluate and rank thousands of individuals.
Computer generated scores are verifiable, repeatable, comprehensive, neutral and based on strictly defined criteria.
 
Let's see some examples of rankings outside of MMA world.

Tennis has ATP Ranking that is based on calculating, for each player, his total points from tournaments in the last 52 weeks. It is used to determine qualification for entry and seeding in all tournaments.
Before 1972, tennis tournament entry criteria were controlled by national federations and tournament directors.
That is when the Association of Tennis Professionals (ATP) was created as the trade union. Top tennis players had to boycott the 1973 Wimbledon Championships in order to achieve their goals. One of those goals was introduction of ranking system intended to objectify tournament entry criteria.
First ATP #1, Ilie Nastase, said that now "everyone has a number hanging over them".

That establishment of a computer ranking system provided fair analysis of a tennis player's performance and an objective means to determine entries into tournaments. Original formula changed several times, but today ATP Rankings represents the official ranking system in professional tennis.

Formula and rules used in calculating current ATP score for each player are precisely defined.
Occasionally there are complains about some anomalies and weird rankings, but everybody accept it as one and only way to represent where each player currently stands in comparison to all other players.
That strict implementation of 52 weeks rule created curious rankings like following:
1.
Last year (July 2015) Refael Nadal dropped to #10 in ATP Rankings. At that time David Ferrer and Nishikori were ranked ahead of him. Where they better tennis players at that time. Probably not, but Nadal was injured and lost a lot of points because of inactivity. When he returned, he wasn't in the top form, but soon he collected enough points to go back to #5. Nishikori is now #6 with 4,845 points and he is just 95 points behind Nadal.
2.
In 2011 Serena Williams dropped out of top 10. This was mostly due to long term injury. She missed first half of year and withdrew from few tournaments. It took her more than a year to get back to #1. Did we in 2011 really had 10 female tennis player better than Serena? Probably not, but that is how ATP ranking works.

So, ATP rankings are based on the calculated value called ATP points. Points are also used to show how dominant some players are in a certain point in time.

Novak Djokovic has 6,000 points more than #2 Andy Murray. That is a very dominant #1 and ensures that he will stay #1 for a longer period of time.
Murray has 4,000 more points than #3 Federer. Stan Wawrinka just lost semifinals in Toronto, but that was enough to pass Nadal on #4.
Those points change every week, even when certain fighters are not active.

Average professional tennis players has more than 15 tournaments and 60 matches per year.
This gives a good statistical data sample to calculate scores for each tennis player.
Compare that to average fighter having just 2 or 3 fights per year. Champions often stay inactive for more than 1 year.
That is just one of the reasons why we cannot use ATP system for MMA rankings.
 
respecting the current rankings would be a start. but I'd agree with a point system as well. shouldn't be hard to make one.
 
Another example of points-based ranking system is FIFA World Ranking. This ranking has high importance because it is used to determine groups in the big tournaments.
It is oversimplified and therefore has a lot of flaws.
Formula for points gained in each game is:
Points = M (points for match result) * I (importance of match) * T (strength of opponent) * C (strength of confederation)
The ranking takes the last four years of games into account, with more recent games weighted more heavily: Current year: 100% weight, 2 years ago: 50%, 3 years ago: 30% , 4 years ago: 20% .
That is the whole formula.
Since international football (soccer) games are so infrequent, you have a small sample of data to use in the first place.
International teams sometimes only play a couple of meaningful games per year (and sometimes none at all). So determining a ranking of every team in the world with such a limited amount of information is not an easy task.

This system is very simple and in the past created some very weird results.
Belgium was ranked #1 in the world (in November 2015) even though they only played in one major tournament in the past 13 years.
To give more current look to FIFA rankings, in 2006 they changed using results from 8 year period to current 4 year period. That was one of the factors in pushing Brazil to #22 in 2014.
 
There are ranking systems for fantasy football, for gaming leagues, and any type of individual or team competition.
But there are also rankings for things outside of sport.

Many websites publicize annual lists of Collages. It generates lots of interest and many people use them to make important decisions.
College rankings are often based on opinion and not actual data.
U.S. News uses peer assessment survey. They measured the quality of teaching at one school by asking people who work at other schools how good the teaching is. It would be like basing the Fortune 500 on just the opinions of other CEOs instead of things like revenue and profit. There are number of useful variables, like looking at the acceptance rate or looking at the high school GPAs and SAT scores of students.
Some collage rankings are based on factors that have no demonstrated impact on academic results. Some rankings are created by education experts that are not always impartial. Interesting fact is that only 3 CEOs of the top 20 companies attended elite colleges, and 12 of the top 20 attended public colleges.
 
So, what lessens we can learn from analyzing other existing pint-based rating systems?
From ATP points we can conclude that having a dynamic number value calculated frequently for each player, it improves competitiveness and acceptance by players.
From FIFA score we can conclude that having simple formula can create very strange results.
From Collage rankings we don't learn much because most of them have preconceived notion of what is important and lot of them are trying to push certain agenda.

What are the challenges for those that are trying to construct quality MMA point-based system.

  • Small data sample for consistent statistical analysis.
Fighters are have on average 2.5 fights per year.
Even having 30 fights in the whole career is very limited data sample. That is how many matches tennis player have in 6 months.
Average fight per year count for top fighters is constantly declining.
In 2000, fighters from top100P4P were having 4.2 fights per year. In 2005 that dropped to 3 per year, in 2010 2.5 , and in 2015 is 2.2 fights per year.

  • Irregular fighting frequency.
Some fighters have 1 fight per year and some 10 fights in the same year.
It is common that one fighter have 3 fights in 6 months and then not having another fight the next 12 months.
Also fighting too many times in one period (much more than average) can create problems in score calculation.

  • Long inactivity periods.
Being inactive for 1 or 2 years is not uncommon. This is often happening with champions. They can enjoying their belt and wait for a good financial opportunity.
Of course, lot of long term inactivity are caused by injuries or unavailability of the opponents. Not fighter's fault, but still inactivity.

  • Having fighters split into multiple weight categories.
This is typical for fighting sports (like Boxing or Judo) but is not happening in other sports like Tennis.

  • Having fighters split between multiple organizations with little interaction.
Since UFC is very dominant organization, this is not a big issue right now. But it used to be an issue in the past and might happen again in the future.



Those are some obvious MMA specific problems for those that attempt to create point-based ranking system.
You can probably add more here.
 
Fighters have too few fights for rankings to be meaningful. That would make any point system a mathematical sham. Your proposed rankings would be no better than the rankings we have now: guesses and opinions.
 
all rankings not categorized by statistics are subjective
 
Fighters have too few fights for rankings to be meaningful. That would make any point system a mathematical sham. Your proposed rankings would be no better than the rankings we have now: guesses and opinions.

I appreciate your opinion, but maybe you should first check ScoreCardMMA ranking system.
Let's not look at how and what systems can and cannot be created.

I would like people here to concentrate on defining some basic, general rules that should be used for MMA point-based ranking.
 
I just visited ScoreCardMMA and read everything they say about their scoring methods. I not impressed. They apply a statistical system to small sets of individual samples of wide variance. Some of their factors are just subjective evaluations , for example Fight Marks, that have no quantifiable predictive value. I don't see their system as any better at comparing fighters or predicting outcomes than any longstanding fan's opinion. If you want to gimmick up your own system for your own amusement, fine, but I doubt it provides anything new or valuable to the sport or its fan base.
 
I just visited ScoreCardMMA and read everything they say about their scoring methods. I not impressed. They apply a statistical system to small sets of individual samples of wide variance. Some of their factors are just subjective evaluations , for example Fight Marks, that have no quantifiable predictive value. I don't see their system as any better at comparing fighters or predicting outcomes than any longstanding fan's opinion. If you want to gimmick up your own system for your own amusement, fine, but I doubt it provides anything new or valuable to the sport or its fan base.

{<redford}
 
Most people would probably agree that we should have a point-based system in MMA, maybe something similar to ATP points in tennis.
Let's use term Current Score for calculated MMA points.
This score represents fighter's MMA achievement at this point in time and it is a numeric value that defines where fighter is ranked on the current ranking list.

What we need to do is define the rules for Current Score calculation. Those rules should be acceptable for majority of fighters.
As I said before, this rankings should be a motivating factor for fighters.
Fighters should now what they need to do to improve their score.

Let's start with following two questions:
  1. How should long inactivity affect fighter's score?
  2. What should happen to score when fighter moves from one weight category to another?
 
Fighters have too few fights for rankings to be meaningful. That would make any point system a mathematical sham. Your proposed rankings would be no better than the rankings we have now: guesses and opinions.

This is a very fair point. Also, stylistic variance in terms of opponents faced is much larger than say, in tennis. I'm not against a more objective ranking system, but I think a large amount of subjectivity will be needed to do it right. Not all wins are the same, and not all losses are the same either.
 
Back
Top