A small, flexible teenager at Carlson Gracie's Copacabana academy built a guard to survive men who outweighed him, then used it to beat an unbeaten Gracie. Four decades later, the position he improvised carries his name into every elite gi bracket in the world — and became the on-ramp to the berimbolo, the highest-scoring sequence in modern sport jiu-jitsu.
The position is named after a man, and the man was small.
In January 1980, a teenager named Ricardo de la Riva Goded walked into Carlson Gracie's academy in Copacabana, in the south zone of Rio de Janeiro. Carlson's competition room in that era was not a gentle place. It was stocked with the heavy, aggressive passers who would later be known collectively as the Arrebentação — the "demolition" or "wrecking" crew — men who trained guard passing as a form of controlled violence and who outweighed the new kid by twenty or thirty kilos.
De la Riva was flexible, quick, and light, and none of that was enough. Flat on his back under those passers, he was going to get flattened. So he started doing something unusual: he threaded a foot around the outside of the passer's lead leg, hooked it deep, and used that single point of contact to spin the bigger man off his base before the pass could land. Opponents playing against it described a wobble, a loss of balance that arrived from nowhere. In the room they started calling it the Guarda Pudim — the "pudding guard" — because the man on top went soft and unsteady, like he was trying to stand on custard.
It stayed a gym curiosity, a specialist's trick, until the 1985 Copa Cantão. There, de la Riva used the hook against Royler Gracie — Hélio's son, a fixture of the family's competition line, and to that point undefeated at the event. De la Riva beat him with the guard he had been building in the demolition room. Brazilian grappling media covering the tournament needed a name for the strange leg entanglement that had just upset a Gracie, and they reached for the obvious one. They called it the De La Riva.
This is not, in the end, a story about a single tournament win. It is a story about what happens when a position invented to survive is handed to a generation that used it to attack — and turned it into the highest-scoring guard in the history of the gi.
The closed guard was built to keep a bigger man from passing. The De La Riva guard was built to spin him off his feet — and four decades later it is still the position from which the modern gi game launches its most valuable sequence, the berimbolo to the back.
The lineage matters here, because it is often told wrong.
De la Riva is a Carlson Gracie black belt — promoted by Carlson himself in 1986, six years after he first walked into the Copacabana gym. That places the guard's inventor one clean step down a specific branch of the family tree: Mitsuyo Maeda taught Carlos Gracie; Carlos taught his younger brothers and his own sons; his eldest son Carlson Gracie ran the toughest competition academy in Rio through the 1970s and 80s; and de la Riva came up inside it. The De La Riva guard is a Carlson-academy invention, not a Gracie-family-office one — a distinction that shows up in the guard's personality. It is a competitor's tool, built under live passing pressure, not a self-defense concept refined in a demonstration.
In 2014, Keenan Cornelius pulled his own lapel out of his belt, threaded it under an opponent's leg, and won a match with a position that had no name. The worm guard he built from it is the clearest modern proof that a guard can be engineered on purpose — a lapel-control system whole enough that its inventor codified it into a flowchart.
The leg drag existed before Rafael Mendes. What he and his brother Guilherme did, out of the Art of Jiu Jitsu room in Costa Mesa, was turn it into a complete passing framework — one that reorganized guard passing around controlling the legline and the hip rather than crushing weight, and that punished the very defense their own berimbolo forced.
The guard's early identity was defensive. In its first decade it was understood as a way for a smaller player to survive the top game of a larger one — an anti-passing device, a way to keep spinning and off-balancing until something better opened up. De la Riva himself used it that way against much heavier opponents throughout the 1980s and 90s, and for years it remained a niche technique associated mostly with him and with a handful of Rio competitors who had felt it firsthand.
The transformation from survival tool to scoring machine came later, and from a different generation. The bridge was the berimbolo — the inverting, spinning back-take that begins from the De La Riva hook. The berimbolo's origin is usually credited to Marcel Ferreira, another Carlson Gracie product, who as a purple belt in the early 1990s began inverting underneath a training partner who kept stepping around his De La Riva hook. The movement was first shown on a major competition stage by Samuel Braga at the 2005 World Championship, and it was André Galvão who is generally credited with attaching the playful name — berimbolo — that stuck.
But the athletes who made it the center of the sport arrived at the end of the 2000s. Rafael and Guilherme Mendes, training out of the Atos system, weaponized the De La Riva-to-berimbolo sequence and won the black-belt lightweight and featherweight divisions with it relentlessly from roughly 2010 onward, adding the leg-drag finish that made the back-take cleaner. Close behind them came João and Paulo Miyao, whose inversion-heavy, near-obsessive De La Riva and reverse-De La Riva game defined the look of elite gi grappling in the 2011–2013 window and after.
By the time that wave crested, the guard de la Riva had built to avoid being crushed had become the single most productive scoring position in gi competition — because the sequence it enables, sweep to back-take, is worth more points than any other chain the rulebook allows.
The De La Riva guard is defined by one signature point of contact and a supporting cast of grips. Get the hook wrong and nothing downstream works; get it right and the whole system of sweeps, back-takes, and submissions opens from a single frame.
1. The De La Riva hook. This is the position. From the bottom, you thread your outside leg around the outside of the opponent's lead (forward) leg, wrapping so that your foot and instep come to the inside of their thigh, behind the knee, hooking their leg from the outside-in. The foot is flexed hard — a pointed toe is a hook that slips. Depth is everything: a hook that reaches only the ankle or shin gives almost no steering, and the opponent simply circles out and passes. A hook driven deep, up behind the knee with the calf loaded against their leg, turns your leg into a lever on their entire base. The single most common way the position dies is a shallow hook that slides out of place.
2. The far-side foot on the hip (or the biceps). Your other foot posts on the opponent's near hip, or in some variations on the same-side biceps or shoulder. This is the strut that manages distance. Bend it to let them close in when you want to load them; extend it to keep them off you and preserve the frame. Without this second point of contact you have a hook and nothing to balance it against, and the opponent folds you toward the hooked leg.
3. Two grips, never one. The hallmark error at every level is a single grip. The De La Riva player wants two: typically the same-side sleeve or cuff, plus a second grip on the collar, the pant leg, or the hooked-leg's ankle. Two grips form what practitioners call a steering wheel — the ability to rotate the opponent's upper and lower body independently. A player who holds only a sleeve, or only the ankle, has given the opponent one free limb to post, base, and walk out of the hook.
4. Stay on your side; never go flat. The De La Riva guard is played on the hip, with the bottom shoulder off the mat. Flatten out and the position collapses into a bad, passable open guard. The tension of the hook and the on-the-side posture are what let a lighter player torque a heavier one. This is the through-line back to de la Riva himself in the demolition room: the guard only produces the famous off-balancing wobble when the player underneath stays angled and springy, never square and flat.
A correctly framed De La Riva hook opens onto four broad families of attack. Every named technique from the position is a member of one of them.
Family A · The De La Riva sweep. The foundational scoring move. Using the deep hook plus the sleeve-and-collar grips, you load the opponent's weight over the hooked leg and extend the hook to tip them past their own base. In the gi this scores two points and delivers you to top position. It is the guard's original purpose — off-balance the heavier man — expressed as a point.
Family B · The berimbolo to the back. The reason the position sits at the top of the modern meta. Keeping the hook, you invert underneath the opponent — rolling onto your shoulders and upper back — and spin through toward their hips, arriving behind them to take the back. In IBJJF gi scoring the sweep component and the back-take stack, making this the highest-value single sequence available under the rules. The kiss-of-the-dragon is a well-known berimbolo variation in which the inverting player's face ends up pressed against the opponent's thigh on the way through to the back.
Family C · The triangle and omoplata. When the opponent postures up or reaches across to fight the grips, the De La Riva player can release the hook and attack the upper body — kicking a leg over the shoulder into a triangle, or swinging it across a trapped arm into an omoplata. These are the "he defended the sweep, so I take the finish" branches.
Family D · Transitions to the leg-lock game. Releasing the hip foot and swinging underneath the opponent flows naturally into single-leg-X and X-guard, which in the modern no-gi world are the launch points for ashi-garami and the heel-hook game. This is how the position, born in the gi, has found a second life in no-gi: not as a scoring guard, but as a corridor into leg entanglements.
Every other De La Riva attack — the balance-disruption sweeps, the back-take variations, the crossovers into reverse De La Riva — is a variation inside one of these four families, not a fifth.
In 2026, the De La Riva guard occupies two very different places depending on which ruleset you are watching.
In the gi, it is not a niche technique or a specialist's trick — it is close to mandatory equipment. Any competitor playing bottom in an IBJJF gi bracket from rooster weight up through the middle divisions is expected to have a De La Riva game, because the sweep-to-back sequence it enables is simply the most efficient way to accumulate points against a competent passer. The reason is arithmetic. A guard that produces a sweep and then a back-take chains two of the rulebook's scoring events into one motion, and the back is the highest-value control position in the sport. No other guard converts a single defensive frame into that much of the scoreboard.
The position also spawned its own most important counter, which is now standard gear alongside it: reverse De La Riva, where the hook goes to the inside of the opponent's leg and spins them the opposite direction. Reverse De La Riva emerged largely as an answer to opponents defending the classic hook with knee-cut and toreando passes, and elite gi players now flow between the two based on how the top player is trying to pass. The two positions function as a matched set.
In no-gi, the story is different. Stripped of the sleeve and collar grips that make the classic De La Riva sweep and berimbolo so controlling, the position is less dominant as a scoring guard. But the hook itself survives, and it has become an entry rather than a destination — a way to off-balance the opponent and thread underneath into single-leg-X and the ashi-garami leg-entanglement game that defines elite no-gi. The same leg position that scored gi points for the Mendes brothers now feeds heel-hook setups for a different generation of grapplers.
What has not changed is the identity of the hook. Whether it is spinning a passer off his base for a two-point sweep, launching an inversion to the back, or feeding a leg entanglement, the De La Riva hook does the one thing its inventor built it to do: it takes a bigger, heavier top game and makes it unstable.
Three moments that mark the guard's arc from gym trick to competition centerpiece. The dated modern match here is verified against public records; see Known open items for the specific caveat.
The naming event. De la Riva submitted the previously-unbeaten Royler Gracie at the Copa Cantão using the leg hook he had developed against Carlson's heavier passers, and Brazilian grappling media attached his name to the position afterward. This is the moment the Guarda Pudim became the De La Riva. The exact date within 1985 and the precise finish should be confirmed against a primary Brazilian source before publication (see Known open items).
The rivalry that dragged the De La Riva-and-berimbolo game to the center of the sport. Their 2010 Worlds final in Long Beach went the distance and was decided 4–4 on a 5–3 advantage count in Mendes's favor — his first black-belt world title, won with a game built around the inverting De La Riva attacks. Mendes and Cobrinha met in the black-belt featherweight Worlds final four times (2010, 2012, 2014, and 2015); Mendes won all four. Their broader head-to-head across all events ran heavily in Mendes's favor.
João and Paulo Miyao's inversion-heavy De La Riva and reverse-De La Riva game, campaigned hardest in the roughly 2011–2013 window and after, became the visual template for what elite gi guard play looked like for a generation of competitors — hooks constantly threading, bodies constantly spinning under the pass. Specific title results are left uncited here pending record verification (see Known open items).
White belt learns the hook and learns to hate how often it slides out. The lesson at this rank is depth and flex: the foot goes deep behind the knee, the toes pull back hard, and the hook gets checked and re-set constantly. Most white belts also learn the position the hard way — by getting toreando-passed the instant the hook goes shallow.
Blue belt learns the second grip. The realization that the guard is a steering wheel — sleeve plus collar, or sleeve plus ankle, never a single grip — is the blue-belt shift. This is also where the first real sweep starts to land against resisting partners.
Purple belt starts inverting. The berimbolo entry — rolling under, keeping the hook, spinning toward the hips — is a purple-belt project, and it is humbling. The player learns to invert after the hook and grips are locked, not before, because inverting into empty space against a set opponent is how you get stapled and passed.
Brown belt learns the switch. Classic De La Riva and reverse De La Riva become a matched pair, and the brown belt flows between them based on which pass the opponent commits to. The guard stops being one position and becomes a small system.
Black belt hides the entry. At the top level the hook is threaded before the opponent realizes the guard is being set, and the sweep, the back-take, and the leg-entanglement exit are all live from the same frame, chosen based on the opponent's reaction rather than a pre-set plan. The black belt does not "play De La Riva"; the black belt steers exchanges into it.
1. A shallow hook. The single most common error at every rank. A hook that only reaches the ankle or shin gives no leverage on the opponent's base, and they circle out and pass — often with a toreando in under a few seconds. Fix: drive the hook deep, up behind the knee, and flex the foot hard so the instep bites. Check it every few seconds and re-set without thinking.
2. A single grip. Holding only the sleeve, or only the ankle, leaves the opponent one free limb to post, strip, and walk out. Fix: two grips, always — a steering wheel, not a handle. Sleeve-and-collar or sleeve-and-ankle at minimum.
3. Going flat. The De La Riva guard played from a flat back is a passable open guard. The off-balancing power lives in the on-the-side, hip-up posture. Fix: keep the bottom shoulder off the mat and stay angled; the moment you square up, you have lost the position's leverage.
4. Inverting before the position is locked. Attempting the berimbolo before the hook and grips are secured sends the player spinning under a set opponent, who simply follows the hips and smashes. Fix: hook first, grips second, invert third. The inversion is the last step in the sequence, never the first.
5. Forgetting the far-side frame. Players fixated on the hook often let the hip-post foot drift, which lets the opponent close distance and fold them over the hooked leg. Fix: the second foot on the hip or biceps is not optional — it is the strut the whole structure balances against.
At open mat this week, isolate the hook. Fifteen-minute round against a same-rank partner with one rule: you are only trying to establish and keep a deep De La Riva hook plus two grips while your partner is only trying to strip the grips, clear the hook, and pass. No sweeps, no inversions, no submissions. You will lose the hook constantly at first, then lose it more slowly, then hold it. That progression is the whole point.
Add a second round in which you are allowed a single attack from the frame — the classic De La Riva sweep, and only that sweep. No berimbolo, no triangle. This forces the mechanics of loading the opponent over the hooked leg without the temptation to bail into a flashier finish.
Leave the berimbolo for its own dedicated day, drilled slowly with an emphasis on inverting after the hook and grips are set. Repeat the hook-retention round weekly for six weeks and the guard will change.
The De La Riva guard is a child of the open guard and a parent of the modern back-take game.
Every one of these positions was reached, historically, by someone who started with a foot hooked around a bigger man's leg and asked what else it could do.
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 2026-08-04. Awaiting hero image (needs permissioned photo — placeholder pre-launch). Two-pass edit not yet complete. Source-check pass required before status: scheduled in Sanity.
Known open items:
Keenan Cornelius did not invent one guard from a stray lapel — he invented a way of reading the gi that produced a whole family of them. The worm guard was only the first. Behind it sits a mapped category — squid, reverse de la worm, gubber, ringworm — codified in his Lapel Encyclopedia as the closest thing the sport has to a formal specification of lapel play.