Roger Gracie won his most-cited IBJJF World title in 2009 by submitting all six of his opponents, across two brackets, from the same place — the mount. This is the high-mount attack system that finished them: flatten, climb to the chest, and run a small closed web of the cross-collar choke, the ezekiel, and the armbar that no amount of foreknowledge could stop.
At the 2009 IBJJF World Championship in Long Beach, six men lost to Roger Gracie in exactly the same way, and every one of them saw it coming. He entered two brackets that weekend — the super-heavyweight division and the open-weight absolute — and won both. What made the run the most-cited campaign of his career was not the double gold. It was the method. Each of the six opponents was passed, then flattened, then mounted, and then finished from the top. Not out-pointed. Not ridden to an advantage. Finished, on the ground, from the mount.
The finishes barely varied. Four of the six tapped to the cross-collar choke — one hand deep in the far collar, the second crossing under it, elbows drawing down until the carotids closed. The other two tapped to the ezekiel choke, a near-relative of the same grip thrown from the same seat on the chest. Bruno Bastos, Rafael Lovato Jr., Claudio Calasans, Romulo Barral, Ricardo Abreu, Bernardo Faria — a field of world champions and future champions — were all beaten by a top-position attack a white belt is shown in his first month, run from a position every practitioner learns to hold before he learns anything else.
This is the paradox the mount system keeps insisting on. In a decade when the sport's vocabulary was multiplying — berimbolos, leg-lock entanglements, lapel guards nobody yet had clean names for — the most dominant gi competitor of the era won by climbing onto his opponent's chest and doing three things extremely well. He did not out-invent the field. He out-executed it, from a position so old and so plain that a room full of black belts could name the attack in advance and still be unable to escape it.
This is not a story about the mount. The mount belongs to everybody. This is a story about what one competitor built on top of it.
Roger Gracie's mount attack system is the clearest argument in the sport's record that the mount is not a control position with some finishes attached, but a finishing position in its own right — a small, closed web of the collar, the ezekiel, and the arm, drilled so deep that knowing what is coming stops being a defense.
Roger Gracie Gomes was born in Rio de Janeiro on 26 September 1981, into two of the most consequential bloodlines in the sport at once. His mother, Reila Gracie, is a daughter of Carlos Gracie Sr., the family patriarch. His father, Mauricio Gomes, is a black belt under Rolls Gracie — long regarded within the family as the finest technician of his generation — who moved to London to teach, splitting Roger's upbringing between England and Brazil. The competition game, though, was built in Rio, at , under ("Carlinhos"), whom Roger has consistently credited in interviews as the architect of the game that would go undefeated in World finals. The formal lineage reads Carlos Gracie Sr. → → Carlos Gracie Jr. → Roger Gracie.
That lineage matters to the mount system, because the mount is where the family's oldest idea comes to rest. The Gracie curriculum has always treated the top of the mount not as a place to grind out points but as the position from which a fight is ended. Hélio's self-defense syllabus, Rickson's competition game, and Roger's IBJJF run all share the same conviction: get to the chest, kill the escape, and finish. Roger inherited that premise and did something specific with it. He stripped the mount's menu of attacks down to the smallest number of high-percentage finishes that could still beat elite defenders, and he drilled them past the point where an opponent's foreknowledge could matter.
The competitive record is the overture. Roger won IBJJF World titles as a blue belt (2000), a purple belt (2001), and a brown belt (2002), then, at black belt, won ten World Championships between 2004 and 2010 — the super-heavyweight division seven times and the open-weight absolute three times, in 2007, 2009, and 2010. He added double gold at the 2005 ADCC in the no-gi arena, where there is no collar to grip and the same top-pressure game simply migrated to the arm-triangle and the rear-naked choke. Across all ten of his black-belt World finals, not one went to a points decision. He finished. The mount is where most of those finishes happened.
The codification — the part that turns a competitor's habits into a nameable system others can learn — came later, on video. Roger's mount game is set out in The Roger Gracie Mount System (BJJ Fanatics), a three-volume instructional in which he teaches the position as a discrete method rather than a survey: how to flatten an opponent and control the mount, how to climb from flat mount to a seated and then high mount, how to break defensive grips and prevent guard replacement, and how to run the cross-collar choke, the ezekiel, the armlocks (including from the Americana grip), and the wrist lock as connected attacks. It is the paid-video equivalent of the closed-door Gracie curriculum, and it is what lets this article treat "Roger's mount" as a system with an author and a source, not a vibe. It is a distinct product from his closed-guard system — the bottom game that, in competition, was frequently just the delivery vehicle that carried him to the top position this article is about.
A section on the generic mount has to catalog the whole menu: every control, every escape to defend against, the full spread of collar chokes, arm attacks, and transitions available from the top. Roger's system is the opposite exercise. It is defined by how much of that menu he refused to touch, and by the order in which he stacked the little he kept.
The spine is four beats, in order: flatten, climb, break posture on the arms, attack the collar. Every good thing downstream is a branch off that spine, and the spine barely changes from match to match. What makes it a system rather than a favorite move is that each beat feeds the next, and each branch loops back to the spine when it fails.
Beat one — flatten the opponent and kill the bridge. Roger's mount was famously heavy and famously low. Chest to chin, hips settled, weight carried forward over the opponent's head, he took away the bridge-and-shrimp escape before any attack began. This is the least glamorous part of the system and the one that makes the rest legal: an opponent who cannot bridge cannot buck the mount off, so his defense narrows from "escape the position" to "fight the hands" — a much smaller problem, and one Roger had already solved.
Beat two — climb from flat mount to high mount. Roger did not attack from a low, straddling mount. He walked his knees up the ribs into the armpits — the high or technical mount — so that his hips sat on the opponent's chest and the defending arms were pinned above, robbed of the leverage to frame or to strip a grip. The higher the mount, the shorter the opponent's arms become relative to their own neck. Climbing is not a flourish; it is the move that makes the choke possible against a defender who knows the choke is coming.
Beat three — break the arms and the posture of the defense. With the mount high, Roger's job was to clear the defending hands off the neck and off his own gripping arm. His instructional dwells on exactly this: the grip breaks and the small pressure sequences that peel a frame away and prevent it from resetting. One of the most reliable tools here is the arm-trap — the control the wider community calls the gift wrap, in which the opponent's own arm is dragged across his face and pinned, so that one of the two arms that could defend the neck is removed from the equation entirely. A defender fighting a choke with one arm is a defender who is already losing.
Beat four — the cross-collar choke is the threat that organizes everything else. The first hand feeds deep into the far collar, palm up, four fingers in, as deep as the grip will travel; the second hand crosses under it to the near collar; the elbows drive down and the forearms rotate like scissor blades so the pressure lands on the carotid arteries at the sides of the neck, not the windpipe. It is a blood choke, which is why grit and pain tolerance do not save the defender — once the arteries close, consciousness is on a short clock that toughness cannot reset.
Here is the part that makes it a system rather than a technique. The cross-collar choke is a real finish and a piece of bait at the same time. Almost nobody at black belt is choked cold from mount by an attack they saw coming; what actually happens is that the defense to the choke opens the next beat, and Roger already knew which one.
What a section on the generic mount cannot say is this: Roger's system is not the menu of options above, it is the deliberate refusal of the rest of the menu. He is not hunting a dozen different collar variations, a spread of leg-trap chokes, or the flashier arm attacks the position can produce. He runs a flatten, a climb, an arm-trap, and a choke he can throw two ways from one seat — and he runs them at a depth of detail (which knuckle, which angle of the elbow, how the hips shift a few degrees to clear the choking arm) that turns "the opponent knows what is coming" from a defense into an irrelevance. The system's genius is subtraction. Breadth is what he gave up to buy that depth.
Roger stopped competing in jiu-jitsu at the peak of the gi run — his last IBJJF World title came in 2010 — and turned to MMA before returning to grappling only intermittently; his competitive career effectively closed in 2017. So the mount system's "modern era" is not a story of Roger defending a title in 2026. It is the story of the idea outliving the athlete.
The most telling endorsement comes from the stylistic opposite corner of the sport. John Danaher and the no-gi schools that produced Gordon Ryan and the modern submission-only elite are about as far from a gi mount-and-collar artist as the sport gets. Yet they have repeatedly cited Roger's 2004–2010 run as proof of the principle they teach: that a small number of positions, understood to their absolute depth and connected into an inescapable web, beat a wide, shallow game. Roger did with the mount and the collar what the New Wave did with the back and the leg. The vocabulary differs; the thesis — depth over variety — is identical.
The instructional market tells the same story from the commercial side. The Roger Gracie Mount System sells on a deliberately unfashionable pitch: this is the boring stuff that actually wins. In a catalog crowded with novelty, "here is exactly how Roger Gracie flattens people and chokes them from the top" remains a durable offer because the material never went out of date. The mount itself is not a contested position in 2026 the way the guard is; nobody claims the mount has been "solved." What Roger's system adds to the plain position is the argument that the mount is a finishing position and not merely a scoring one — that the correct thing to do from the top is not to sit for points but to climb, trap, and end the match.
The one real evolution is contextual. In Roger's competitive prime, the terminal choke often landed from a stationary high mount he could sit in for as long as the finish took. The modern no-gi elite reach the same top positions inside faster, more chaotic scrambles, and finish before the position ever fully settles. Roger's system was never networked into that kind of flow; it did not need to be, because the flatten-and-climb sequence bought him all the time the finish required. What survives, undiminished, is the underlying claim: that a competitor who has arrived at total command of the mount has not fallen behind the innovation curve. He has stepped off it.
The strength of this system is that, unlike the closed guard, its decisive moments are cleanly documented — the 2009 campaign is one of the best-recorded submission runs in the sport, opponent by opponent.
White belt learns the cross-collar choke from mount in the first month and concludes it does not work, because every training partner defends it. This is the correct experience. The white belt's lesson is that the choke is real; keeping the mount is the hard part.
Blue belt learns that the choke is a mount problem before it is a hand problem. If the mount is low and light, the opponent bridges and the choke never arrives. The blue belt's whole job is to make the mount heavy — chest to chin, hips low — so the hands have time to work.
Purple belt learns to climb. The realization that the attack lives in high mount, knees in the armpits, with the defender's arms pinned above the fight — not in the low straddle a beginner sits in — is the shift into a real top game. Purple belt is also where the first hand starts going in deep and going in first, before the second grip is even considered.
Brown belt learns the web. The cross-collar is not a standalone finish; it is the center of a small system — ezekiel when the collar is defended, armbar when they frame, the back when they turn. The brown belt closes the exits, so that every defense to the choke opens a different finish and the man underneath is choosing between three losses without realizing he is choosing.
Black belt rebuilds the whole thing from the opponent's point of view. He knows every escape and pre-empts it, arriving in a high mount the opponent cannot leave and a choke the opponent cannot see coming even though he knows it is coming. This is the Roger Gracie level: the system is public, written on an index card, and still cannot be stopped.
1. Attacking from a low mount. The single most common error, and the one that quietly defeats everything downstream. A low, straddling mount leaves the opponent room to bridge and the arms free to frame. Climb first — knees to the armpits, hips onto the chest — and then attack the neck. The choke fails because the position was never finished.
2. A light mount that gives up the bridge. If the hips are high and the weight is carried back, the opponent bridges and shrimps out before the grips finish. Chest low, weight forward over the head, kill the bridge, then work the hands. The finish is downstream of the pin, not a substitute for it.
3. Grabbing the second collar first. The instinct is to set both hands at once, which produces two shallow grips and no choke. The deep first hand goes in alone, all the way to the back of the neck if it will travel, before the second hand crosses. A shallow first grip is a wasted attack.
4. Pulling instead of spreading. Beginners yank the collars toward themselves, which chokes the windpipe (uncomfortable, survivable) and postures the opponent's defense back into the fight. The finish is a spread — elbows drive down, forearms rotate, pressure on the sides of the neck. If they can talk, you are on the trachea; move to the arteries.
5. Treating the choke as a single technique instead of a web. The practitioner who commits everything to the cross-collar and quits when it is defended has thrown away the system. The whole point is that the defended choke is the entry — to the ezekiel, the armbar, or the back. If a defended choke ends your attack, you were running a move, not Roger's system.
At open mat this week, split the system into its spine and drill one beat at a time, because the whole point is depth, not variety.
Round one — flatten and climb. Start every round in mount, on top, against a same-rank partner, with one job: keep the mount heavy, kill the bridge, and climb from flat mount to high mount with the knees in the armpits. No submissions. Your partner's only job is to bridge, frame, and try to replace guard. Fifteen minutes. You will fail, then fail more slowly, then hold the high mount. That is the position's whole curriculum in miniature.
Round two — the choke as bait. Add the cross-collar choke and one branch: whenever your partner defends the collar by burying his chin or framing, switch to the ezekiel or trap the framing arm and hunt the armbar. Do not force the cross-collar; reward the defense with the branch. This trains the through-line — one seat on the chest, three finishes — that carried the 2009 run.
Round three — close the web. Let your partner defend fully and reward their choice: when they turn belly-down, take the back and switch to the collar from behind; when they frame and press, take the arm; when they bury the chin, take the ezekiel. Now you are drilling the system, not a technique — which is the actual thing Roger Gracie did six times in one weekend of World finals.
Repeat weekly for six weeks. The point is not to learn more; it is to make less inescapable.
Roger's mount attack is a specific, reduced dialect of the sport's most dominant top position, and it links directly to the pages that hold the rest of the map.
The through-line is that none of these are advanced positions. Roger's system is the first positions in the sport — the mount, the collar, the arm — understood last.
Publication date will be set at Sanity ingest. This log will be appended to over the life of the article. Any factual corrections will be dated and specific.
Publication draft prepared 2026-08-07. Awaiting hero image (needs permissioned photo — placeholder pre-launch). Source-check pass required before status: scheduled in Sanity.
Consistency note (with articles 01 closed guard, 36 Roger Gracie, and the Roger Gracie closed-guard system). This article uses Roger's IBJJF World Absolute title years as 2007, 2009, 2010 (seven super-heavyweight titles 2004–2010, three absolutes), matching the corrected article 01 and the athlete profile 36. It does not assert the unverified "25 May 2013 Grand Slam de Los Angeles vs. Buchecha" match. It treats the mount attack system as distinct from both the generic /positions/mount page and Roger's closed-guard system: the closed guard was, in competition, largely the delivery vehicle that carried Roger to the mount finishes this piece describes. The 2009 all-from-mount campaign is verified here against the IBJJF's own retrospective, which resolves the per-match method split (4 cross-collar, 2 ezekiel) that articles 01/36 had left hedged; those pages can now be updated to cite the same source.
Known open items:
</content> </invoke>