MMA By the Numbers

By Adrian Meader

Intro to Statistics

I have been an avid Mixed Martial Arts fan since its inception in 1993 when the Ultimate Fighting Championship held its first tournament to test and see which martial art would reign supreme. Mixed Martial Art or MMA as it has become to be known is a full contact combat sport that allows the use of both striking and grappling techniques, both standing and on the ground, including boxing, wrestling, Brazilian jiu-jitsu, muay Thai, and judo. MMA has grown and evolved by leaps and bounds over the past 20 years. MMA often times is compared to a game of rock, paper, scissors where certain stylistic martial arts fare better against some but not so well against others. In the early 1990s, practitioners of grappling based styles such as Brazilian jiu-jitsu and wrestling dominated competition in the United States. Practitioners of striking based arts such as boxing, and kickboxing, who were unfamiliar with submission grappling proved to be unprepared to deal with its submission techniques. Shoot wrestling practitioners offered a balance of amateur wrestling ability and catch wrestling-based submissions, resulting in a better-rounded skill-set. The shoot wrestlers were especially successful in Japan. As competitions became more and more common, those with a base in striking arts became more competitive as they cross trained in arts based around takedowns and submission holds, leading to notable upsets against the then dominant grapplers. Likewise, those from the varying grappling styles added striking techniques to their arsenal. This increase of cross-training resulted in fighters becoming increasingly multi-dimensional and well-rounded in their skills. In certain circles there is a running joke about what is known as MMA math, typically personified in that person x beat person y who beat person z so person x logically should be able to beat person z. It is also a currently held belief that if a fighter comes from a wrestling background he supposedly has the best training foundation to produce effective fighters in modern MMA competitions, and that basically wrestling wins fights because it allows the fighter to determine where the fight occurs (either standing or on the ground.). My desire was to apply statistical based math to determine if this was in fact true or if there was only minor correlation to this prevalent theory within the MMA community, or if there was a factor more prevalent. The data sample consists of 169 current fighters on the Ultimate Fighting Championships (UFC) roster. The UFC is currently the largest MMA promotion and is considered to have the most elite fighters in the sport. Fighters were selected from www.Fightmetrics.coms data base. I established a ground rule that a fighter needed to have a minimum of three fights within the organization, which I felt was sufficient to demonstrate their comparison to other fighters at a similar level. Fighters from all of the

different classes were examined so as not to skew the results towards a particular weight division. Career Statistics looked at within the striking category included: SLPM-significant strikes landed per minute, Str. Acc.-significant striking accuracy, SAPM-significant strikes absorbed per minute, Str. Def.significant strike defense (The percent of opponents strikes that did not land.). For the grappling aspect of the sport the following data was reviewed as well. TD. Avg.-average takedowns achieved per 15 minutes, TD Acc.-takedown accuracy, TD. Def.-takedown defense (the percentage of opponents takedowns that did not succeed.), and Sub. Avg.-average submissions attempted per 15 minutes. Also looked at were fighter win/loss records, in order for ease of use I converted these to a % based number. I.E. a fighter with 5 wins 5 losses has a 50% win record. I felt this also allowed accurate weight of a fighter with only 3 fights vs. a fighter with 30. These win% numbers were based solely on the data available through fightmetric, as some fighters full records were not listed with accompanying statistics. I still feel though that there was enough raw data to get a good cross section of the various fighters styles. First we will look at strikes Landed per minute in correlation to win percentage. See fig 1. There is only a .1639 correlation demonstrating that in MMA it is not necessarily the amount of punches you land. Upon reviewing the data the outliers to the right of the graph were well known boxing style fighters that specialize in a style known as punches in bunches. I also found it interesting that along the top line (100% win mark). The range of volume striking was well represented. Fig 1

Next we take a look at striking accuracy to determine that if volume is not the key what about being able to land effective strikes accurately. It stands to reason that this will take into account the power punchers that have the ability to know an opponent out with one good strike. See Fig #2

This correlation came in at .139 again surprisingly low. Again you can notice the range across the top for undefeated fighters in this category.

Next we will look at strikes absorbed. This is a measure of how many strikes a fighter takes in the course of his bout. As you can see the regression line trends downward with a correlation of -.3888, a much higher correlation than seen before. But still surprisingly low. You would believe that common sense would dictate that the more you get hit the less likely you are to win the fight.

The next graph shows the ability of a fighter to defend against and opponents striking game. This data shows the percent of opponents strikes that did not land. The correlation for this regression line is .1917. Again demonstrating very little correlation but falling within the same realm as the first 2 graphs. You will also notice that very few fighters represented had less than a 40% ability to defend against strikes. I believe this is indicative of the bare striking fundamentals one must have to compete at this level.

Now we move onto the grappling aspect of the sport. The following graph represents takedowns achieved per 15 minutes of fighting. Its important to note that these are successful attempts not failed attempts. The correlation here is .3409 Still low correlation but coming in higher than some of the striking aspects.

Next we look at takedown accuracy; this is a measure of how successful a fighter is in his attempts to take his opponent down. The correlation came in at .2721, again lower than expected based upon the prevailing theory that wrestling wins fights.

Finally we look at Takedown defense, or the ability for a fighter to make his opponent fail at his take down attempt. This also lends to the wrestling background theory as it is believed the good wrestling not only allows you to take an opponent to the ground but it also allows you to keep the fight standing when you so desire. The correlation arrived at was .3227.

The submission data was discarded as the correlation came in at -.0159 showing no correlation what so ever to attempted submissions to winning percentage. Data was not available for successful submission accuracy which I feel is a better gauge than just attempts.

Conclusion: The expected results based upon the prevailing theory were that the grappling aspect of MMA was to have a strong correlation to winning fights. However correlation scores never exceeded .3409, indicating that with the sample provided there was not a strong enough correlation to prove this statement. However it is important to note that the grappling correlations did come in at higher scores than the striking aspect of the sport. MMA contains innumerable variables that are difficult to measure. Things like a fighters cardio, speed, strength advantage over his opponent, injuries, ones heart and, the will to win are difficult things to boil down to numbers. This was never more present and visible than the famous UFC 117 fight in 2010 pitting Chael Sonnen against the Champion Anderson Silva. Chael Sonnen dominated the scheduled 25 minute fight for 23 minutes. He out struck Anderson Silva 89 to 29 strikes and took the champion down 3 times. With 2 minutes left in the fight and Anderson Silva on his way to being defeated. Sonnen had Silva on the ground and was rocking him with punches, Silva threw a desperate submission attempt and forced Sonnen to tap out and concede the match. It is considered one of the all-time greatest comebacks in the fight industry. In viewing the fights statistics it does not paint an accurate picture of what took place. Such is the case when utilizing MMA math.

Bibliography "FightMetric LLC :.: The World's First Comprehensive MMA Statistics Provider." FightMetric LLC :.: The World's First Comprehensive MMA Statistics Provider. N.p., n.d. Web. 12 June 2012. <http://www.fightmetric.com>.

Snowden, Jonathan , and Kendall Shields. The MMA Encyclopedia. Toronto: ECW Press, 2010. Print.