Intro
This piece began as a response to the following comment on reddit by the user monoplatajones. I have no beef with this person and wish them only the best. I simply disagreed with them and got to typing a response that grew far too much to fit into the reply chain. The original thread was a discussion in which a relatively new grappler asked if De La Riva and other “gi guards” were necessary, or if “no gi” guards like butterfly and half guard were enough. The comment was as follows:
“People really like to point to Marcelo or Wardzinski but there’s a really good reason almost no one plays inside position based guards as their primary attacking position in 2026, people are too good at using pant grips to attack from the top position and you’re giving them easy access by putting the feet in front.
Marcelo was winning in the gi 15 years ago, things have come a LONG way at the top.”
Firstly, I have to point out that this specific assertion, worded the way it was is objectively incorrect. While supine without any grips, the guard player’s feet are much more “in front” than they would be from seated just by simple anatomical fact. Becaue of this, almost every athlete, regardless of level, plays the engagement phase of guard sitting up while they fight for an initial grip (assuming a disconnected state to begin with). The actual point here is that most gi guard players do drop back to supine position once they have established a grip, and I believe that the point the user was trying to make was that the reason for this is because of the threat of outside passing the pants. I do not necessarily disagree with this, but I do question handwaving the best counterexamples. Even if you dismiss Marcelo as too far in the past to be relevant (I wouldn’t; people stood up to pass, death gripped the pants, and blasted both gear and torreando passes since long before he retired), you have to contend with the fact that Wardzinski was the #1 ranked gi athlete in the world a short time ago, and seemed like he could have kept it up for a while longer if he so chose, and ignoring that level of excellence would be an enormous mistake. It also can’t be explained away with him being a physical outlier, as I have seen people attempt to do in the past. He’s a big, strong guy, but he’s also a professional athlete who weighs over 200lbs. Everyone at that size is strong. If anything, he looks and performs like way less of a physical freak than many of the others in his weightclass. I would submit that this indicates that his game is effective in the gi at the highest levels, and doesn’t rely on some unique set of physical attributes. For a further example, you could also look to Bruno Malfacine, whose career took place largely in between those of Marcelo and Adam, and also favored inside position guards.
Lesser Reasons For The Disappearance
So, we have established that there have been dominant champions throughout the history of Jiu Jitsu who play butterfly, and that they do not all share some unique athletic/physical profile. Why, then, is it still a relatively rare style? That is a question that is impossible to answer in absolute, objective terms, but I can at least offer a few hypotheses:
The ideal style for seated guard is not culturally favored in most BJJ gyms. When you watch someone like Marcelo Garcia play guard, it looks very little like the mental model of a laid-back, technically superior Jiu Jitsu master setting traps and waiting for a reaction. He is bolt upright, aggressively seeking his preferred grips and attacking sequences, instantly stripping any unwanted connections his opponent makes. This is unquestionably still good Jiu Jitsu; his success in the absolute division against titans is proof enough of that. But it is also the sort of thing an instructor is likely to tell a white belt to chill out with. That isn’t a knock against coaches: I myself spend a great deal of time telling beginners to calm down, but at the formative stages of development, it can be difficult to distinguish between “seize the initiative” and “club my training partner with all my might for no reason”.
Seated position doesn’t allow the bottom player to abuse flexibility for retention. I expect this could be a little more controversial of a point, but I will defend it. Being able to walk into the gym and put both your legs behind your head without warming up doesn’t mean you have “good guard retention” any more than being able to squat 700lbs and flinging people off you means you have “good escapes”. It is not good Jiu Jitsu, and I think it is silly that we pretend otherwise. Nevertheless, it is a common part of the game, especially at lower weight classes, and a lack of access to their normal self-pretzeling abilities to bail them out of bad spots almost certainly contributes to people choosing not to play the seated style.
People follow established metagames. While I did point out that there are examples of dominant butterfly guard players across the history of Jiu Jitsu, there are even more examples of people dominating in the gi with supine guard positions, and success begets success. Those athletes of course spread their style to their own students and fans, and further generations of supine guard players are born.
I would submit that the relative rarity of inside position has more to do with Brazilian convention than most people realize. This is far more relevant in the gi than without it, because if you haven’t taken a look at IBJJF rankings recently, just know that Brazilians continue to dominate. And if you talk to many Brazilians, you will notice that they consider things like double sleeve or collar/sleeve just…the way you play open guard. That’s their starting point. The idea that you’d do something different is almost alien.
The Big Problem
All of those prior explanations may or may not hold water based on the reader’s evaluation of my arguments, but most of them are cultural or historical in nature, not technical. This begs the question: Are there actual, technical reasons that seated guards may be less favored in the gi? I believe the answer is yes. I think that it is fundamentally true and demonstrable that seated/inside position is a “live by the sword, die by the sword” choice, particularly in the gi. Losing the ability to use leg pummeling against pant grips is a major issue, even when the guard player is not unduly relying on flexibility. In fact, I would call it the central problem of the position. One thing Wardzinski and Marcelo Garcia have in common is that sometimes people would manage to blow by their legs at the very beginning of the match, using athleticism to quickly secure a momentary “pass”. I put “pass” in quotes there because, while the opponents did often manage to get around their legs, securing position and scoring was far less common. This is due to the fact that, in addition to butterfly, they are also both masters of late defense (Wardzinski more with turtle, Marcelo more with late seated retention and early side control escapes), and were excellent at parlaying near-disastrous positions into a favorable connection vs someone who preferred to remain disconnected on the outside. That’s praise for their skills, certainly, but it’s also an undeniable downside of the system. In a perfect world, our guard protects us and gets us to good position without incurring risk. However, there is another side to that coin: these people also never really get stymied or stalled out. It is very common for supine guard players to go down by sweep/takedown points (or even just an advantage) and get stalled out for minutes at a time, unable to create meaningful action. By contrast, guys like Wardzinski (Marcelo was basically never down on points later in his career in the gi, but the same would apply in theory) can consistently mount comebacks when swept or taken down because their system allows them to push the action in a ways more a passive one simply cannot.
Solution Models
That of course leads us to: if standing pant grips have the potential to be such kryptonite for seated guard positions, how have there been dominant champions playing those positions throughout the history of Jiu Jitsu? Are they magical, special outliers whose success is impossible to replicate? Of course not. Fortunately, this is an arena in which we do not have to engage in cultural speculation, as the athletes in question have left us a beautiful, concrete legacy in the form of their recorded matches and technique.
The first of these to go over will be Adam Wardzinski, who is the most recent competitor and also the most “pure” guard player I will cover. By that I mean that while Adam is certainly capable of takedowns, and is fantastic on top, you rarely see him truly jockey for the takedown from standing position. Instead, he opts to pull virtually every time if that battle is contested. I will keep this section relatively brief, because other people have already done a fantastic job of analyzing his game. Perhaps best known in this space are BJJScout and RixtyGarami’s excellent analysis videos of his game. However, most relevant to us here is Corey Walker’s excellent article for Flo Grappling, wherein he explains how Adam funnels standing passers into his game almost verbatim how I would. To quickly recap his content, as well as summarize my own understanding of his game:
If Adam can obtain a right-handed grip on the opponent’s left sleeve, he is, as the kids say, “big chillin’”. He is able to use that grip to block passes, attack dummy sweeps, and enter shin to shin.
If Adam cannot obtain that grip, or finds himself defending against circling/knee cuts, he will use a half guard/reverse De La Riva hook and a shin to shin on the same leg to stretch and ground the opponent. From there, he threatens a double attack with his knee lever sweep and the half butterfly with an over the back belt grip.
The next athlete I will touch on with this is Marcelo Garcia, who is probably the most historically discussed in this space. Unlike Adam, Marcelo was more willing to battle for the takedown in his matches, but still pulled frequently. As for how Marcelo navigated this situation, I believe it can be boiled down to two things: phenomenal handfighting to obtain his preferred two on one grip, and fierce, nuanced late defense. The actual process of his hand fighting could merit a novella length piece of writing on its own, but suffice it to say that I think Marcelo Garcia might be the best handfighter the sport of BJJ has ever seen. The consequences of this are obvious: he was usually able to find his two on one grip, and when you have a tight two on one grip, outside passing suddenly goes from deeply threatening to a near nonissue. Some great examples of his handfighting with the two on one in action can be seen here and here. In both matches, the story is similar: Marcelo pulls guard, either already with the two on one, or fights to it immediately afterward. He then leverages the grip to weather the initial passing storm and eventually find his preferred bottom position with inside hooks. On the flip side, Marcelo did not really seem to have anything like Adam’s shin to shin+half guard hook to ground disconnected standing passers. He either got the two on one grip, or he leaned on that aforementioned late defense. By “late defense” I mean the family of techniques that he pioneered/made popular that blur the lines between wrestling, guard retention, and side control escapes. Marcelo would combine the conventional frames, shrimping, and leg pummeling with less common but still powerful tools like short arm frames, the “James Brown” split, and the elbow push escape to make sure that even when opponents “passed” his legs, they never really secured the position as a real pin that scored points. Because all of these defenses are “late”, they also allowed him to reenter his offensive cycle, because the opponent would inevitably be close enough for Marcelo to engage his butterfly hooks (or stand up).
Lastly, I will point to an athlete that I have as of yet barely mentioned: Bruno Malfacine. I am not quite as familiar with him as with the other two, but his success and longevity are undeniable, and he did indeed play a significant amount of butterfly guard in his career. Those with longer memories may recall that Malfacine was known more for his top game than his guard, which is true. However, I think this overlooks the extreme efficiency he had in the guard position, using his mastery of inside position to sweep and wrestle up against some of the best competitors to ever do it. Indeed, while writing this I rewatched a number of his old matches and struggled to find a single instance of his guard really failing him; virtually all of his losses came about as the result of his opponents’ incredible guard work, not as a result of his failures in bottom position. How then, did he solve the problem I described with the engagement phase? Largely, he didn’t. The double guard pull is incredibly common at rooster weight, and was even more so before the rules changed to limit it. That in mind, Malfacine’s competition strategy seemed to be to use a flying “sit” guard pull to get forward momentum and pull at the same time as his opponent. He would then use his inside position, superior wrestling, and compact frame (not to mention unusual physicality) to come up on top, starting the match up an advantage. That is all fine and good, but what about if the opponent successfully came up on top, or if the opponent declined to pull altogether? In the first case, the double pull itself obviates solving the engagement problem. By definition, someone coming up from a double pull is already engaged and connected, which lets the bottom player immediately begin to use their hooks. As for when the opponent did not pull (rarer than you might think at rooster, if you’re unfamiliar), the flying nature of his pull would often win him connection without a particularly technical or involved process. The most interesting case, however, is what happened when it didn’t. As mentioned, Malfacine enjoyed superiority on the feet to most of his opponents, and so in this case his strategy preyed on what I believe is the primary weakness of outside passing with pant grips: before strong grips are established, it does little to keep the bottom player down. Malfacine in these positions would not tolerate a cagey guard passer leaping around his guard. He would instead go to his knees, and seek grips and/or takedown opportunities from there. This is perhaps best illustrated in his match with Gui Mendes at Pans 2012, where he used this strategy to score and eventually win vs a legendary competitor while giving up size (although yes, he absolutely stalled while on top). I would also note that this strategy is not without value even if you cannot assume better wrestling. Hand fighting from the knees/one knee is immensely easier than from seated position (more extendable length, better mobility), and building height also gives access to “flying” guard pulls, or at least sitting with forward momentum once grips are obtained.
Synthesis
So what is the consequence of all this? I feel that I have made a reasonably strong case for the viability of butterfly guard in the gi, regardless of era or weightclass. That is all fine and good, but it would be better if we could extend this beyond the realm of academic argument. More to the point, how can this impact you, the reader, and potentially help your Jiu Jitsu? If you choose to play seated guards, we can distill our takeaways into a handful of points:
Learn Some ****ing Takedowns.
Remember how I called Wardzinski closer to being a pure guard player? Here he is hitting a single leg in the finals of the world championships. Pulling guard should be a tactical choice, not an excuse to ignore standing position. The increased access to takedowns from bottom position is one of the key advantages of seated guard systems, and it would be wildly remiss to leave them on the table.
Controlling A Sleeve is Key
All of the athletes we have covered share this same characteristic: their choice of weapon against the pant grip is to control a single sleeve (double sleeves provide comparatively little value for a seated player). Marcelo uses the two on one. Adam uses the cuff grip one on one. The results are similar: relative safety and a powerful offensive platform. Consider handfighting from the feet or knees a la Malfacine if a sleeve isn’t immediately available; every moment you spend on your butt disconnected from an athletic enough guard passer is a moment you could be passed in a flash.
The Collar May Be A Trap
None of the examined athletes seek a collar grip as their primary platform, and that is worth noting. Playing seated with a deep cross collar grip is one of the most classic attacking platforms in the gi, but you’ll see it sought out seemingly not at all by the sport’s best butterfly players. The answer for why this is the case might be as simple as the fact that collars are much further away than sleeves (for the guard player). I am not about to claim that the cross collar grip is weak, but it is much more difficult to find than sleeve grips vs a standing opponent, and it is very plausible that the extra time and space required to get it will allow an athletic, savvy opponent too much time to take their own grips and begin passing.
Seated Retention Is Not the Same as Supine Retention
An enormous deal has been (rightly) made of supine guard retention in recent years, and the average Jiu Jitsu practitioner is better for it. The same cannot be said of seated retention. Ideas like the short arm frame, the wrestler’s switch, and heisting out of pass attempts have yet to really catch on with the rank and file grappler. Many of the best exemplars of these techniques in the modern era are no gi athletes, but some of the best butterfly players of the past used them extensively (Marcelo), and they remain transferable to the gi to this day.
What Remains To Be Done
The above summarizes (on a very surface level) some of the options and principles that contributed to the success of butterfly players in the past, but what of the future? After all, one of the only constants in Jiu Jitsu is change, and it would be implausible (not to mention disappointing) to think that the seated guard of 2036 would look identical to the seated guard of today. Here, I lay out some prospective tools that I think will be used and problems I believe may need to be solved moving forward.
The Belt Grip
Belt grips (with an underhook, an over back grip, or a cross arm and belt) are nothing new for seated guard. However, they have primarily been used against kneeling opponents, and usually reach the back of the belt for elevation or butterfly sweeps. The front of the belt is also a powerful grip, and with a stiff arm is incredibly difficult to pass. Furthermore, it is frustratingly difficult for the passer to break, often requiring a commitment to big, opportunity-creating motions to accomplish. Saulo Ribeiro is an advocate of this grip for butterfly, and Terere used it some, so there is also historical precedent. It has fallen off for the moment, but I think the right athlete could have great success with it in the future.
A Bridge From The Outside
Sometimes, no matter how good a player’s seated guard is, the smart retention option will be to go to their back and use their legs to ensnare the opponent. This can be done with either wide “outside position” legs blocking the torso/outsides of the legs, or with a half guard/reverse De La Riva hook. Adam Wardzinski has provided us with a superb model for how to bridge the latter with butterfly guard techniques, but no real, battle-tested equivalent exists for outside position. I do not know what this might look like. I’ve experimented with it, but found nothing satisfactory. What I do know is that a powerful, consistent transition from this spot would be a great boon to inside position players everywhere.
Gamesmanship
I do not like this aspect of competitive Jiu Jitsu, but it is important: A major reason we do not see people using heisting and handfighting from the knees/more expanded positions like Malfacine is that under many rulesets, were they to be knocked down (two inches into the position they were in moments ago), it would score for the opponent. In my ideal ruleset, this would be irrelevant (people would be rewarded for positions rather than actions), but I do not make the rules. If seated guard players could figure out a way to insulate themselves from this possibility, it would no doubt make playing the position more appealing for IBJJF competition.