You are on page 1of 12

12/26/2015 CageMatch|PopularScience

How science is transforming the sport of MMA ghting

BY MATTHEW SHAER Posted AUGUST 28, 2012

Greg Jackson, the single most successful trainer in the multi-billion-


dollar sport of professional mixed martial arts ghting, works out of
a musty old gym in Albuquerque, New Mexico, not far from the base
of the Sandia Mountains. On a recent morning, the 38-year-old
Jackson, who has the cauli owered ears and bulbous nose of a
career ghter, watched two of his students square off inside the
chain-link walls of a blood-splattered ring called the Octagon.

One of them was Jon Jones, the light heavyweight champion of the
Ultimate Fighting Championship (UFC), the premier MMA league.
In four weeks, Jones would be defending his title against Rashad
Evans, an expert ghter and his former training partner. To prepare
him, Jackson had set up a sparring session with Shawn "The Savage"
Jordan, a heavyset ghter from Baton Rouge.

Jones and Jordan met in the middle of the ring. Jordan threw rst.
Jones backpedaled and protected his face with his forearms.

"Look for that space, Jones!" Jackson hollered. "You. Do. Not let him
close those angles on you." Jordan threw a urry of blows. To me,

http://www.popsci.com/science/article/201207/cagematch 1/12
12/26/2015 CageMatch|PopularScience

the exchange appeared disorganized, nonsensicala blur of esh,


sinew and the red ash of Jordan's mouth guard.

To Jackson, it was a logical sequence, one with only one possible


effective response. "Jones," he said, "move inside." The ghter
seemed to hesitate. If he moved within range of Jordan's sts, he
risked catching a glove square in the face.

"Go on," Jackson said.

Jones ducked under one st and whipped his right leg out in a short
arc. The kick missed. Jordan threw again. This time Jones dropped
down, icked his head to the side, and, leaping off one foot,
launched a ying jab followed by a knee to Jordan's midsection,
which landed with a wet . Jordan groaned and crumpled
onto the mat.

"Goddamn, Jones!" Jackson yelled. "Exactly correct."

Producing a notepad from his back pocket, Jackson sketched a


spiderweb of circles and lines. It was a game tree, he explaineda
graph game theorists use to analyze a sequence of decisions. In a
traditional game tree, each circle, or node, represents the point at
which a decision can be made. Each line, or edge, represents the
decision itself. Game trees eventually end in a terminal nodeeither
a tie or a win for one of the players. This game tree, Jackson told me,
showed the exchange between Jones and Jordan from Jones's
perspective.

At the start, the two men stood a few feet apart. Jackson drew a
circle. The node had three edges, or moves that Jackson was
training Jones to use. He could execute a leg kick, or a punch, or he
could shoot for a takedown (attempt to grab Jordan by the backs of
his legs and drive him into the ground). But the initial node was not
http://www.popsci.com/science/article/201207/cagematch 2/12
12/26/2015 CageMatch|PopularScience

"optimal," he said, because it allowed Jordan to swing freely with


both sts. Although it seemed counterintuitive, the fast track to
what Jackson calls the "damage" node (in this case, Jones's
advantageous position following his hard knee) was to move in
close, where Jordan would not be able to fully wind up. Another
circle, representing Jones's inside position, and a series of edges,
representing his potential decisions from there, appeared on the
notepad.

"From inside," Jackson said, "he can do a knee, he can do an


uppercut, he can do elbows. He could have done anything there, and
done it effectively."

Since 1992, when he opened his rst gym, Jackson has been using
math to inform his training techniques. Unlike other MMA coaches,
he continually collects data while watching live bouts, logs old ght
videos to determine which moves work and when, and lls
notebooks with game trees to determine the optimal nodes for
various situations in a match. "I've always seen the ring like a lab," he
says. "I've tried to think rigorously, logically."

"I've always seen the ring like a lab," Jackson says. "I try to think
logically."Jackson's attempts to impose some measure of order on
the primal, violent world of MMA mirror a larger movement within
the sport. Science may not be civilizing cage ghting, but it is
re ning it. Specialty rms compile detailed statistics on matches.
MMA pros appear on ESPN rigged head to toe with sensors and
monitors that measure their striking power and speed. Academics
are writing peer-reviewed articles on subjects such as the
physiology of top ghters and the role that fear plays in the
Octagon. And now ghters, most of them trained by Jackson, are
beginning to use this data and analysis to become ever more

http://www.popsci.com/science/article/201207/cagematch 3/12
12/26/2015 CageMatch|PopularScience

brutally effective in the ring.

The very rst UFC event took place before a crowd of about 7,800
in a Denver auditorium in 1993. It was an odd spectacle. Karate
masters clashed with boxers. Kickboxers dueled with sumo
wrestlers. There were few real rules.

Over the next decade, in an effort to placate critics and state


athletic commissions, the UFC introduced a comprehensive set of
regulations that outlawed especially dangerous moves such as low
blows and hair pulling. The campaign was largely successful, and by
the mid-2000s, dozens of states had agreed to sanction MMA
events.

TV networks, meanwhile, noticed the UFC's large following and


began to broadcast highlights from the big bouts. A popular reality
show called debuted, and a mixed martial
artist appeared for the rst time on the cover of .
Ticket prices kept increasing. So did the size of the sport's fan base.

Among the many die-hard UFC fans was Rami Genauer, a journalist
based in Washington, D.C. Genauer had read , Michael
Lewis's best seller about Oakland Athletics general manager Billy
Beane and his statistics-driven approach to player evaluation. He
dreamed of analyzing mixed martial arts in the same way.

"There were no numbers," Genauer says. "You'd try to write


something, and you'd come to the place where you'd put in the
numbers to back up your assertions, and there was absolutely
nothing."

In 2007 Genauer obtained a video of a recent UFC event, and using


the slow-motion function on his TiVo, he broke each ght down by
the number of strikes attempted, the volume of strikes landed, the
http://www.popsci.com/science/article/201207/cagematch 4/12
12/26/2015 CageMatch|PopularScience

type of strike (power leg versus leg jab, for instance) and the
nishing move (rear naked choke versus guillotine, and so on). The
process took hours, but the end result was something completely
new to the sport: a comprehensive data set.

Genauer titled his data-collection project FightMetric and created a


website to house the information. Some UFC fans registered their
disapproval on Web forums. "'We don't need math with our
ghting,' people would say. I disagreed," Genauer says.

In 2008 he managed to persuade the UFC to use FightMetric data


from past matches to support a televised event in Minneapolis. "The
idea was that this would be good for the producers, who could use
the numbers to illustrate the story," he says. "It'd also be good for
the broadcasterthey'd have ammunition, something to rely on just
like they do in other sports."

Of cials liked having Genauer's ght data, and when the UFC began
spif ng up its broadcasts with more graphics and statisticspart of
an effort to make MMA seem like a real sport instead of a series of
cage brawlsit hired FightMetric as its statistics provider. Genauer
quit his job and opened an of ce in D.C.

Today FightMetric has ve full-time staffers and a rotating cast of


15 specialists who collect a large data set for each ght using a
video feed, proprietary software and a video-game controller with
which they can record every type of strike. Among the statistics
they track: each ghter's number and type of strikes, number of
signi cant strikes (de ned as all strikes landed from a distance, as
well as power strikes landed from close range) and the accuracy and
location of kicks and punches.

The FightMetric team collects the strike and location statistics in

http://www.popsci.com/science/article/201207/cagematch 5/12
12/26/2015 CageMatch|PopularScience

real time. The UFC uses some of the data for graphics during
broadcasts and on its website. FightMetric goes into even greater
detail on its own website, presenting statistics over outlines of a
human body. Colored lines indicate the accuracy of each type of
strike, and boxes show which ground move, whether arm bar,
kimura lock or triangle choke, each ghter used to try to induce a
submission. The analysis is strangely disconnected from the
violence of the Octagona savage ght broken down into simple,
neat gures.

As the available body of data from FightMetric (and its main


competitor, CompuStrike) grows, Genauer and others are
attempting to analyze it in new ways. Already Genauer and his
colleagues have identi ed some clear trends in MMA matches. For
instance, the number of ghts that end in decisions, especially at the
lower weight classes, has risen from a third in 2007 to half today.
That's a signi cant change from the wilder early days of the UFC,
when ghters swung crazily and the vast majority of bouts ended in
knockouts. It points to increasing skill levels among UFC ghters
(knockouts usually happen when one ghter is obviously superior to
the other), a factor that could affect ghters' styles and training
methods. A lighter-weight ghter, expecting now to go the distance
in his next ght, might accordingly develop his aerobic threshold (so
he can wear out bigger opponents) rather than his ability to throw
rst-round knockout blows.

Earlier this year, John Ruggiero and Trevor Collier, economists at


the University of Dayton, and Andrew L. Johnson, an engineering
professor at Texas A&M, released a study called "Aggression in
Mixed Martial Arts: An Analysis of the Likelihood of Winning a
Decision." With data from FightMetric, the researchers estimated
the probability of winning based on ghter characteristics like

http://www.popsci.com/science/article/201207/cagematch 6/12
12/26/2015 CageMatch|PopularScience

height and age. From a sample of 946 matches, they measured


dozens of variables, including blows attempted versus blows
landed, stand-ups, knockdowns and slams. Next they ran that data
through a binary response model (a kind of algorithm) to determine
which characteristics or approaches most affected a ghter's
chances.

Some of the study's conclusions were surprising. For example, in


ghts that end in decisions, the number of strikes thrown appears to
be more important than the number of strikes landed. This may
have something to do with the vantage point of the judges, who
can't always see the ghters clearly, and so occasionally in error
mark a thrown strike as a landed one. Or it may be that a high
number of thrown punches simply contributes to the appearance of
dominance. Either way, the study is something a ghter can use: The
more punches you throw, the more ghts you'll win.

Researchers studied matches to determine which variables most


affected a ghter's chances of winning.Genauer says he is
constantly working to improve both the hardware and software
used to collect ght data. As collection methods improve, the data
will become richer, analysis will become more granular and the
results more useful. That's been the case in other sports such as
baseball, which have changed as statistical analysis of in-game
strategies has become more sophisticated (as rst
highlighted). Stats have suggested, for example, that sacri ce
bunting is not as useful as previously thought, leading many teams
to attempt it less frequently. In MMA, trainers might nd
demonstrable proof that certain moves, like sidekicks or ying
punches, are less effective than others, like knees or arm triangles.
They might see the consistent success of a shoulder lock or the
repeated triumph of the arm bar. They might rely on that data to

http://www.popsci.com/science/article/201207/cagematch 7/12
12/26/2015 CageMatch|PopularScience

engineer a better approach to MMA ghtingone, as FightMetric's


website advertises, "rooted in data and demonstrated effectiveness
rather than in gut feelings and bandwagon jumping."

"Data and demonstrated effectiveness" is something that Greg


Jackson has stressed for years. Unlike other MMA coaches, Jackson
holds no belt in any martial art and has no allegiance to any guru. In
fact, he had hardly any formal training at all. He opened his rst gym
at the age of 17. In the absence of a particular ghting style, he
experimented with practically all of them: aikido, karate, Jiu-Jitsu,
Muay Thai, kickboxing, straight-up boxing. "All I was doing was
looking for empirical evidence," he says. "I'd form a hypothesis and
I'd try it out in a ght. If it didn't work I'd get rid of it, and if it did I
kept it. It was science at its purest. It was driven by need."

Jackson would have two evenly matched ghters spar 10, 15, even
20 times in a row. Waiting nearby, notepad in hand, he would
assiduously track which moves worked in the greatest number of
situations. Unlike most trainers, he held no sentimental attachment
to any speci c moves. If he found that a ying sidekick didn't
consistently do enough damage, he'd stop teaching it.

By the early '90s Jackson had incorporated his results into his own
homegrown martial art, which he dubbed Gaidojutsu"way of the
street," roughly, in Japanese. Gaidojutsu combined rudimentary
striking with grappling and wrestling. At the time, it was rare to
blend ghting stylesmost ghters trained in a single discipline. But
Jackson's students relished the chance to play mix-and-match, and
his stable of trainees grew. A few of them persuaded him to let them
compete in bare-knuckle tournaments, where they dominated their
undisciplined opponents. By the time the UFC came around,
http://www.popsci.com/science/article/201207/cagematch 8/12
12/26/2015 CageMatch|PopularScience

Jackson says, he was completely addicted to winning competitions.

But he knew the UFC would be a far cry from the bare-knuckle
bouts. He'd need to further re ne his methods. One person he relied
on for help was Jim Dudley, a close friend and mentor who also
happened to be a mathematics lecturer at the University of New
Mexico. Dudley gave him private math lessons in the desert, giving
him assignments from books on subjects such as discrete
mathematics and discussing how he might apply math in an MMA
match.

"My rst memory is Greg asking me about fractals," Dudley says.


"Then it was game theory. I had no idea at rst that all of this
pertained to ghting. When he nally told me, I thought, 'OK, that's
odd.' But then again, I knew [math] could be applied to very
surprising topics. It made sense that Greg would be nding these
interesting patterns in ghting."

The patterns that Jackson found were sequences of moves and


positions that most consistently led to success in the Octagon. "I
saw these certain positions over and over again: the side-mount, for
instance, or the full-mount," he says. "And I started thinking of them
in terms of edges. Judging from the data, which positions offered
the most opportunities? Which left the ghter in trouble? And
which allowed him the quickest path to victory?"

What Jackson was developing was a new way of thinking about


ghting, one informed by mathematical and logical frameworks
rather than gut instinct. Crucial to that was constant data
collection. Where other coaches might drift in and out of the gym,
catching snippets of training rounds here or there, Jackson almost
never leaves the apron of the Octagon. He is responsible for
approximately 60 professional ghters, some champions and some
http://www.popsci.com/science/article/201207/cagematch 9/12
12/26/2015 CageMatch|PopularScience

up-and-comers, and every day he watches almost all of them spar


for hours on end. When he is not watching training bouts or
traveling with his team, he is clicking through clips of older matches
on his iPhone, on the TV, on one of the scarred laptops that sit in his
cluttered of ce alongside a photograph of Albert Einstein and one
of his personal heroes, the famous logician Kurt Gdel. His desk
spills over with handwritten logs of successful ghts, hastily
scrawled game trees of sparring sessions, points about form and
function and technique.

All of these notes contain usable data. Analyzing his game trees
shows him the best moves to make at different points in a match,
while logs of his ghters' and their opponents' past matches help
him predict how long an upcoming ght is likely to last, when in each
round the opponent will strike and what moves he'll make. It's an
advantage no other trainer yet has.

In early April Jon Jones defended the light heavyweight belt against
Rashad Evans. The ghters were once friends who trained together
under Jackson, but they'd had a falling out. In the weeks before the
bout, they spent plenty of time trash-talking each other in the
media. The ght was a true grudge match, as the UFC billed it, and
by the time Jones and Evans climbed into the Octagon at Philips
Arena in Atlanta, anticipation (and the noise level) was at a peak.

The ght opened slow. The ghters danced around each other
warily. Evans, shorter and stockier than Jones, snapped away with
his jab. Jones slipped around him, throwing a mix of "superman"
punches (a punch executed while leaping forward) and ying knees.

Near the end of the rst round, Evans caught Jones with his foot,

http://www.popsci.com/science/article/201207/cagematch 10/12
12/26/2015 CageMatch|PopularScience

sending him off balance. The bell rang. Jackson was waiting for
Jones in the corner, a red cap pulled over his shaved head. His gaze
was intent. He knew Evans had a superb defense and fast hands,
limiting Jones's options. He began constructing a game tree in his
mind. In the rst node, the two men were squared off against each
other. Jones could punch away, but Evans would block most of the
blows. He needed to move to another node, one with more edges.

One node appeared optimal: If Jones could manage to get in


position to effectively neutralize both of Evans's hands, he might be
able to land at least one big shot. Jackson shouted in Jones's ear. His
student nodded.

Toward the end of the next round, Jones, heeding Jackson's advice,
squared up against Evans and extended both hands, open-gloved.
Evans matched him, and for a moment it looked as if the two men
were about to play patty-cake. This was the node that Jackson was
looking for. Evans was momentarily exposed. In dazzlingly quick
succession, Jones threw a right elbow, then a left, then another
right. Evans wobbled, and Jones surged forward with a knee and a
left hook.

By the third round, Jones had his opponent on the defensive. Evans
turned one way, and Jones was there. Turned another, and there he
was again. In the fourth, Jones buried his knee in Evans's stomach,
and the crowd, more than 15,000 strong, roared its approval.

At the end of the night, Jones was awarded a unanimous decision.


He would keep his belt. But it was the work of the FightMetric data
collectors, not the judges' decision, that revealed how truly
dominant Jones had been. Their report showed that he'd landed
116 strikes, 105 of which were deemed signi cant. Evans, by
comparison, landed only 49 strikes, 45 of them signi cant. Jones not
http://www.popsci.com/science/article/201207/cagematch 11/12
12/26/2015 CageMatch|PopularScience

only ran Evans ragged around the ring, but he also doubled his
output, continually nding the node where he could throw the most
blows.

A few days after the ght, I spoke to Jackson by phone. Already he


was dissecting what had happened, picking out the things that
Jones had done right to further hone his ght strategies. But he
realizes that a time will come when other trainers, eager to gain any
advantage they can, will begin to emulate his methods. Eventually
more and more mixed martial artists will base their training and
match plans on statistical probabilities instead of instinct and
tradition, raising the quality of competition.

That means Jackson will have to work harder than ever to stay on
top of the sport. But when I asked him how important winning is to
him, he got quiet. "Never put a node for victory," he said nally.
"That doesn't mean we don't want to win. I want my guys to be
thinking about trying to get to the strongest position they can, with
the most edges, over and over. Like any science, it's more about the
process than it is the outcome."

Among Righteous Men.

Copyright 2015 Popular Science. A Bonnier Corporation Company.


All rights reserved. Reproduction in whole or in part without permission
is prohibited.

http://www.popsci.com/science/article/201207/cagematch 12/12

You might also like