Breakfalls in Seven Systems — the Ground Decides, Not the Style

System	The ground	The back means	Thrower keeps hold?	Basic solution
Judo	tatami, 40–50 mm	ippon — the match is lost	yes, to the end	slap
Aikidō	mat	nothing — no competition	often yes, at the wrist	roll
Sambo	mat, shoes worn	total victory for the thrower	no — he must stay upright	slap
Wrestling	thick wrestling mat	a fall — the match is lost	yes, he holds you down	avoid the fall
BJJ	mat	guard — an attacking position	usually yes, he follows you down	roll
Capoeira	floor of the roda, often hard	part of the playing surface	no — no grip at all	roll
Parkour (control case)	concrete	nothing — no opponent	nobody	roll

Overview

The breakfall is the only technique in the entire field that nobody wants to win with. It scores no point, impresses no spectator, and in most systems it is taught in the first lesson and never seriously discussed again. That is exactly what makes it revealing: there is no pride distorting the answer. What gets taught here is what the conditions force.

And the answers diverge widely. Judo has you slap the mat with a flat hand. Aikidō rolls. Wrestling has no breakfall curriculum at all. Capoeira treats ground contact as a movement option rather than an incident. And parkour, which works with the hardest ground of them all, explicitly forbids what judo requires.

This article places seven systems side by side: judo, aikidō, sambo, wrestling, BJJ, capoeira — and parkour as a control case, because there is neither an opponent nor a rule book nor a mat.

The thesis is not “which breakfall is best”. It is:

Breakfall training does not answer the question “how do I fall”, but three questions of rules: How hard is the ground? What does the back mean? And does the thrower let go?

Answer those three for a system and you can predict its breakfalls without ever having seen them.

The seven systems side by side

System	Ground	The back means	Thrower keeps hold	Basic solution
Judo	tatami, 40–50 mm	ippon — the match is lost	yes, to the end	slap (ukemi)
Sambo	mat, shoes worn	total victory for the thrower	no — he must stay upright	slap
Aikidō	mat	nothing, no competition	often yes, at the wrist	roll, otherwise high fall
Wrestling	thick wrestling mat	a fall — the match is lost	yes, he holds you down	do not fall at all — bridge
BJJ	mat	guard, an attacking position	usually yes, he follows you down	falling is a move
Capoeira	floor of the roda, often hard	part of the playing surface	no — no grip at all	transition, acrobatics
Parkour	concrete	nothing, no opponent	nobody	roll, never slap

The first question of rules: how hard is the ground?

Judo competition takes place on tatami for which the world federation prescribes measurable values: 40 to 50 millimetres thick, at least 55 per cent impact absorption, 4 to 8 millimetres vertical deformation. This is not a marginal detail but the precondition for everything practised on it. A mat that swallows more than half of the impact energy turns a fall from two metres into something the body can take repeatedly — a hundred times per session, for decades.

Only this mat makes slapping possible. Striking the ground with hand and forearm sacrifices a small body part to relieve a large one: the arm absorbs part of the impulse before shoulder, trunk and head arrive, and it spreads that impulse over a larger area and a longer time. On a mat that yields, the arm is equal to the load.

On concrete it is not. Which is why parkour sits at the other end of the scale — and there the rule is exactly reversed. The traceur rolls diagonally across one shoulder to the opposite hip, forms the hands into a triangle that keeps the head off the ground, and avoids contact with anything close to bone. Nothing is slapped. Strike concrete with your forearm and you break it.

Both systems solve the same physical problem — a mass meeting a surface at speed — and arrive at opposite techniques. The difference does not lie in the martial art. It lies in the surface.

The second question of rules: what does the back mean?

Here the systems separate most sharply, and it is purely a question of rules.

Judo defines ippon as a throw that lands the opponent largely on the back — with force, with speed, and controlled by the thrower until the end of the landing. The back is defeat. That creates a peculiar double objective: the person thrown must survive the impact and must simultaneously avoid arriving flat on the back. This is precisely where the sideways ukemi with an angled body comes from — it is not only protection, it is also an attempt to deny the score.

Sambo uses the same notion of the back but attaches a second condition: total victory is awarded only if the thrower stays on his feet throughout the throw. That single clause changes the breakfall noticeably. Whoever is not dragged along falls alone — and a thrower who wants to remain standing does not follow the thrown player down. Sambo therefore produces more free, unaccompanied falls than judo.

Wrestling is the extreme case. Both shoulders on the mat end the bout immediately. A wrestler who developed a breakfall for the supine position would be practising a technique for the moment of his own defeat. So he does the opposite: he builds an entire repertoire for not reaching his back — the bridge, in which only feet and head touch the mat, is the best-known form. Wrestling has no breakfall because its rule book does not know falling as an event to be survived, but as a state that must not be reached. On top of that, wrestling takedowns are on average of lower amplitude than judo throws, so the impact is milder to begin with.

BJJ inverts the question entirely. There the supine position is the guard — an attacking position with a vast technical repertoire of its own. The rules draw the consequence: a takedown scores two points only if the thrower also stabilises the position; whoever lands inside the opponent’s guard gets nothing, however large the throw. And pulling guard is explicitly permitted — you may lie down voluntarily. A system in which the ground is the actual playing field needs no emergency technique for getting there.

Capoeira goes one step further: there the ground is not even a destination but a third level alongside standing and jumping. The queda de rins, literally “fall on the kidneys”, is a controlled descent onto one elbow from which one evades, attacks or moves into acrobatics. It is called a fall but is not one — it is a position with exits.

The third question of rules: does the thrower let go?

The least noticed of the three, and in practice the most consequential.

A roll requires forward momentum and a free arm. Both exist only if the thrower releases or if the throw runs forward. Hold the arm and the roll becomes physically unavailable — and then only the hard landing remains.

Aikidō is the proof. It rolls wherever it can: the forward roll across shoulder and opposite hip is the basic form there, not the exception. But precisely in the techniques where tori holds wrist, forearm or elbow — kote-gaeshi, sumi-otoshi and their relatives — there exists a category of its own, the high or jumping breakfall (tobi ukemi, in some schools hiyaku ukemi). It is not an advanced flourish but the only solution for the case where the roll is blocked.

Judo has this problem as a permanent condition. Under the rules the thrower must control, therefore hold — control until the end of the landing is a scoring criterion. Where the grip stays, the roll is not an option. That is why judo’s breakfall is a school of slapping while aikidō’s is a school of rolling. Not because Kanō and Ueshiba disagreed, but because their systems leave the thrown player different arms free.

Capoeira is the opposite pole: no grip, no throw in the narrow sense, only leg sweeps such as the rasteira. Someone with no opponent attached to them can roll, turn or escape into a handstand at any moment — and does.

What has been measured — and how clearly

Unlike stick fighting, where no comparative data exist at all, the evidence on breakfalls is unusually clear. It covers only one system, however: judo.

In 2020 Murayama and colleagues had one fifth dan throw another with osoto-gari, four times, with a six-degrees-of-freedom sensor on the thrown player’s forehead. With clean ukemi the head registered 10.3 ± 1.6 g translational and 679.4 ± 173.6 rad/s² rotational acceleration.

On their own these figures say little. They become interesting next to the thresholds:

For translational acceleration, 98 g is the value at which the risk of mild traumatic brain injury reaches 75 per cent. What was measured is a tenth of that.
For rotational acceleration, 4500 rad/s² is taken as the concussion limit. What was measured is barely 680.
The comparison value for an unchecked impact on the back of the head, recorded with an instrumented dummy head and collated in the systematic review by Lockhart and colleagues (2022, 16 studies, 158 judoka), is 5081.3 ± 691.8 rad/s².

The injury threshold therefore lies exactly between the two cases. Above it without ukemi, at roughly a seventh of it with. It is rare for a martial arts technique to be documented this cleanly.

Why osoto-gari of all throws. It lays the thrown player backwards without giving their head any way to avoid the mat — the angular acceleration at the head comes out higher than with ippon-seoi-nage or tai-otoshi. Which is why this one throw sits so far up the Japanese accident statistics.

And the statistics are the real justification for the whole discipline. In 2013 Kamitani and colleagues analysed the damage reports of the All Japan Judo Federation for 2003 to 2010: 72 catastrophic injuries, of which 30 to the head and 19 to the cervical spine — and 27 of the 30 head injuries involved players under twenty. Over the longer term, the Japan Judo Accident Victims Association counted 118 deaths in school judo between 1983 and 2011, disproportionately often first-year students, that is, beginners.

The common denominator is not the harshness of the sport but the absence of breakfall skill. The injured are those who cannot yet do ukemi — and who train with stronger partners who can already throw.

The 2022 systematic review also shows where novices and experienced judoka measurably differ, and the result is counterintuitive: not in neck flexion. The difference lies further down. Experienced judoka extend the knee faster and further and keep the trunk straighter; novices flex hip, knee and trunk more, and it is precisely this greater flexion that correlates with the higher readings at the head. Tucking the chin is therefore not what makes the difference — it is the line of the body below it.

Slap or roll — the real dividing line

This settles the most common dispute on the subject.

Slapping the hand is usually justified by saying that it dissipates energy and enlarges the contact area. Both are true, and both are a mat argument. The arm can only dissipate energy if it survives it, and it survives only if the ground does its share.

Parkour supplies the control: harder ground, greater drop heights, no rule book forbidding anything — and still nothing is slapped. If slapping were the superior falling technique, it would have to appear exactly where the conditions are hardest. It does the opposite.

From this follows a clear assignment:

Slapping occurs where there is a cushioned mat and the thrower keeps hold and the fall runs backwards or sideways (judo, sambo, parts of jūjutsu).
Rolling occurs where forward momentum and a free arm are available (aikidō, capoeira, parkour, large parts of BJJ).
Avoidance occurs where the back is defeat and the impact is mild anyway (wrestling).

It is a remarkably complete explanation: for none of the seven systems does one have to fall back on tradition, origin or philosophy. The rules and the surface suffice.

What is the same everywhere

Three things are found in all seven systems, and are therefore probably forced by the matter itself.

The chin comes to the chest. The back of the head is the only spot whose impact is immediately life-threatening — the Japanese fatalities are almost uniformly acute subdural haematomas. Even systems without any breakfall curriculum teach this.

Area is enlarged and time extended. Whether through the slapping arm, the rounded back rolling through, or yielding in the leg: all seven spread the same impulse over more area and more time. That is not a technique, it is the only physics available.

Nobody lands flat. No system lands on both shoulders, the pelvis and the heels simultaneously. There is always a sequence — bottom to top, from one edge to the next, along a diagonal. An impact that happens in sequence is several small ones instead of one large one.

What this means for your own training

Change system and you must change breakfall. A judoka who slaps in aikidō at kote-gaeshi instead of taking the high fall lands on a shoulder that is still attached to a held wrist. An aikidōka who tries to roll out of osoto-gari in judo does not have the arm free that this would require. Both errors share one cause: transferred technique without the transferred condition.

Breakfalls belong in the warm-up, not in the beginners’ syllabus. The accident figures come from beginners training with advanced partners — that is, from the window between the first throw one receives and the first breakfall that holds up. That window is closed by practising falling for years, not by learning it once.

On hard ground you roll. For everyday self-defence that is the only transferable part. The pavement has 0 per cent impact absorption. What spares the arm on the mat breaks it there.

The line of the body counts for more than the chin. That is the practical yield of the measurements: extended knee, straight trunk. Curling up as you fall measurably worsens your readings — even with a perfectly tucked chin.

Common fallacies

“Slapping is the correct way to fall.” It is the correct way to fall onto a judo mat while someone holds your arm. Outside those two conditions it is the wrong one.

“Wrestlers have no breakfall because wrestling is incomplete.” Wrestling has none because its rule book does not know falling as an event. What exists there instead — the bridge, the turns out of the supine position, the whole repertoire of not being put down — answers the question the rule book actually asks.

“Ukemi protects against head injury, so judo is safe.” The measurements document the effect of the technique, not the safety of the sport. The Japanese figures show where the risk sits: with those who cannot yet perform the technique and are thrown anyway.

“Aikidō falls are spectacular, therefore unrealistic.” The high fall is the consequence of a held hand — not showmanship but the only solution when the arm is not free to roll. The reproach hits the demonstration, not the technique.

“Chin to the chest and nothing happens.” Necessary but not sufficient, and precisely the point at which novices and experienced judoka do not differ significantly. The difference sits at hip, knee and trunk.

On the strength of this comparison

The measurements apply to judo and to a single throw. They come from laboratory trials with very small samples — in Murayama’s case one thrower and one thrown player, four throws — and from trials with dummies whose neck mechanics only approximate a human’s. The 2022 review summarises 16 studies with 158 judoka, of whom nine were women; for female judoka the evidence is correspondingly thin.

For the other six systems no comparable measurement exists. What is said about them here is derived from their rule books and their documented training practice: the world federation’s tatami standard, the ippon criteria, the sambo clause on the thrower staying upright, the fall by shoulders in wrestling, the BJJ point rules on takedown and guard.

The assignment “slap / roll / avoid” in the panel is an assessment, not a measurement. It is testable, but it has not been collected. A comparison measuring the breakfalls of several systems under identical conditions would be the obvious study — and it does not exist.