GJJ Pedagogy

How and why we teach the way we do at Golden Jiu Jitsu.

The Short Version

Every class is built around games. We use a constraints-led approach where students develop skills through guided play, not repetitive drilling of techniques at zero resistance. Brief drilling introduces movement pathways that students connect to games immediately through the drill-to-game pattern. We teach concepts first, then create environments where students have to use them. Our conceptual framework organizes grappling knowledge into three layers, and we teach from the top down: big concepts that transfer everywhere before position-specific details.

This page explains the learning science behind these choices and how to apply them as a coach.

Games as the Core (and Where Drilling Fits)

Most grappling gyms prioritize fine-detail instruction on random techniques, often with little connection between classes. A coach demonstrates a move, students drill it at zero resistance, then everyone rolls live. The assumption is that “the game teaches the game.”

Live rolling is better than zero-resistance drilling. But unstructured live rounds without specific goals leave a lot of learning on the table. And the disconnect between instruction and practice means the only people who benefit from the technique demonstration are those who happen to land in that exact position during sparring.

Our approach is different. We design games with specific constraints that force students to solve problems using the concepts we’ve introduced. Games are the core of every class because that’s where the real learning happens: perceiving, deciding, and acting against a resisting opponent in real time.

Where drilling fits

Drilling isn’t useless. It has a specific, limited role: introducing movement pathways that students need before they can play a game productively.

Think of it this way. If you invented a game called backflip tag where you have to backflip and then tag someone within three seconds for the tag to count, nobody is going to learn a backflip by playing the game. They need blocked practice first. But as soon as they can reliably execute a not-too-sloppy backflip, there are diminishing returns on drilling more backflips. Playing the game is what develops the skill from that point forward.

Grappling works the same way. If the game starts from a single leg and a student has never shot a single leg, they won’t learn much by getting stuffed every round. A few minutes of drilling the entrance gives them a movement pathway they can connect to the game. Once they have that pathway, more drilling doesn’t help. Playing does.

This is the drill-to-game pattern: brief isolated practice of an entrance or movement, then immediately into a game where that movement is the starting point. The drill gives students a stable waypoint. The game develops everything else.

Progressing from stable to unstable

Drilling also serves another purpose: controlling how much chaos students face. You can start a game from a later, more stable position (the single leg is already secured, now finish the takedown) and then progressively move the starting point earlier and less stable as students improve (now you have to shoot the single leg first, now you have to get a grip before you shoot, now you start disconnected).

This progression from stable to unstable is how we move students from blocked practice to serial practice to randomized practice within a single class or across a multi-week topic. With relatively athletic students, you can often skip the drilling and drop them straight into a game. With less athletic students or complex movements, the drill-to-game bridge matters more. Read the room and adjust.

Entropy always increases

Within any segment of class, chaos should only go up, never down. The progression is always: drill → constrained game → less constrained game → open positional sparring. Each step adds unpredictability, resistance, and decision-making complexity. You never reverse direction.

This is a hard rule because the reverse (game → drill) actively damages the learning environment. When students are playing a game at real intensity and you pull them back to drill a movement, they don’t maintain their game intensity and add the movement on top of it. They downshift. They lower their intensity across the board to accommodate the new, harder thing you’ve introduced. The game degrades into something that looks more like drilling at half speed. You’ve lost the energy, the perception-action coupling, and the engagement you built.

Rob Gray’s concept of satisficing explains the mechanism: when a task conflicts with a performer’s current motivation (in this case, playing at intensity), they’ll find the easiest shortcut that’s “just good enough” rather than doing the harder thing you intended. Introducing a drill mid-game gives students an easy shortcut: just go slower at everything.

Going drill-to-game avoids this because students start at low intensity, build the movement pathway, and then the game gives them a reason to ramp up. They’re adding intensity to something they can already do, not trying to bolt a new skill onto existing intensity.

If you find yourself wanting to stop a game to show a technique, resist the urge. Either address it between rounds with a quick concept (without demonstrating), change the game constraints so the desired movement becomes more likely to emerge, or save the demonstration for the start of the next round as a drill-to-game entry point.

The research

Rob Gray’s work on motor learning shows that skill lives in the relationship between the athlete and their environment, not just in the athlete’s head. You can’t separate perception from action and train them independently. Pulling apart “see the position,” “decide what to do,” and “execute the technique” into separate drills doesn’t transfer to live performance. Students need to perceive and act simultaneously, in real time, against a resisting opponent. This is why games are the core, not drilling.

Doug Lemov’s framework for game-based activities reinforces this. Well-designed games multiply two types of learning events: “mental touches” (reading and adapting to what’s happening) and specific high-value situations (which constraints make occur more frequently but unpredictably). Standing around during extended drilling is counterproductive because students aren’t getting reps AND aren’t reading the game.

The Constraints-Led Approach

The constraints-led approach (CLA) is the engine behind our game design. Instead of prescribing solutions (“do this technique”), we manipulate constraints to guide exploration.

Three Types of Constraints

From Karl Newell’s constraints model:

Individual constraints. What the athlete brings: height, weight, strength, flexibility, experience. Different for every student. This is why we don’t teach one “correct” technique. A 130-pound guard player and a 220-pound former wrestler will solve the same positional problem differently, and both solutions can be correct.

Environmental constraints. Properties of the training environment: mat surface, room size, training partner’s attributes. We can’t always control these, but we can be aware of them.

Task constraints. The rules of the game: starting position, win conditions, allowed techniques, time limits, scoring. This is the lever coaches pull most. By changing the rules, we change what solutions students discover.

How Constraints Shape Learning

They narrow the solution space without prescribing solutions. A game that says “start in guard, win by sweep, no submissions allowed” eliminates some options and focuses exploration on sweeping mechanics. The student discovers how to sweep through play rather than being told the steps.

They amplify errors. Rather than correcting a student directly, you can constrain them so their current approach fails more obviously. If a student disengages every time they’re in a bad position, a game where points are scored for time in contact forces engagement. The constraint makes the error visible and costly.

They create variability. Rob Gray’s research shows that movement variability is not noise to be eliminated. It’s an essential feature for adaptability. Training against different partners, from different starting positions, under different rule sets creates the variability that builds robust, transferable skills. Repetitive drilling of one pattern creates fragile skills that break under pressure.

They maintain the perception-action link. When students drill a technique at zero resistance, the perceptual information they need in live rolling (opponent pressure, timing, weight distribution) is absent. Games preserve this link because the opponent is always responding.

The 70% Rule

Challenge level affects learning. Too easy and there’s not enough perturbation to drive adaptation. Too hard and the student becomes frustrated and stops engaging. The sweet spot is around 70% success rate. This is where FYJJ becomes essential: the more experienced player self-handicaps to calibrate the challenge for both partners.

The 3-Fold View

When we learn a new subject, we build out a mental map over time. A useful way to think about that map is through three different lenses.

Bird’s Eye View

The big shapes of the landscape. In grappling, these are the overarching concepts that describe the sport. They play out in every position.

Mechanical concepts:

Alignment (base, posture, structure)
Core mechanics (frames, levers, wedges, clamps, hooks, posts)
Force vectors
Internal and external rotation
Knee-elbow connection
Momentum and center of gravity

Strategic concepts:

Regulated tension
Pressure and timing
Stability vs. mobility
Control before submission
Alignment before everything
Placeholders (don’t abandon one control until you’ve replaced it)

This is where we want newcomers to start. These concepts give people footing no matter where they find themselves, and they allow beginners to integrate everything that follows much faster. If someone understands frames and levers on day one, they have tools that work from every position they’ll ever encounter.

View on the Ground

The layer of common positions and movement pathways. Names for positions, transitions that connect them, the “map” of how grappling flows.

The most important type of knowledge for practitioners is practical knowledge: being able to do the sweep matters more than knowing what it’s called. But having names for positions and movements still matters. You learn faster when your mental map has labels on it. And shared vocabulary lets training partners communicate and coaches give efficient feedback.

These movement patterns are learned even faster when paired with concepts from the bird’s eye view. Understanding why a movement pathway works helps you modify it on the fly, which matters because no two repetitions are ever exactly the same.

Under the Microscope

Fine details and deep dives. Position-specific systems and algorithms. This is the territory of detailed instructionals: a single topic covered to significant depth.

These details matter, especially for submissions, but we don’t want to miss the forest for the trees.

When Each Lens Matters

In nonlinear situations (guard passing, scrambles, open guard exchanges): concepts and heuristics are the most valuable tools. If we tried to build a guard passing system that accounted for every possible variation, it would be too bloated to use.

In more linear situations (back control with a body triangle, mounted triangle, deep half): ground-level and microscopic knowledge become more useful. These positions are more stable, so systems can be compact enough to retain practical value.

When attacking submissions: fine details are often the difference between a tap and an escape.

As a coach, this should guide your emphasis. Don’t spend 20 minutes on micro-details when teaching guard passing. Don’t skip the details when teaching a choke finish.

Coaching Language

How you communicate directly determines what students learn. Nick Winkelman’s research on coaching language, Rob Gray’s work on attentional focus, and Gray’s ecological coaching framework all converge on the same principles.

Let Constraints Do the Talking

The single most important coaching skill is knowing when to shut up. Gray describes his mentor Coach Phillips as “a master at letting his unique practice activities do most of the talking for him.” Instead of telling a goalkeeper to close down the angle, Phillips removed certain movement options so the goalkeeper discovered the solution through play.

Well-designed constraints teach without words. If you find yourself explaining at length, your game design probably needs work. The constraint should make the desired behavior obvious through experience, not through instruction.

External Cues Over Internal Cues

Over 20 years of motor learning research shows that external focus of attention consistently produces better performance and better learning retention than internal focus.

Internal cue (less effective): “Rotate your hips” or “extend your arm” External cue (more effective): “Drive your knee toward the wall” or “push their head to the mat”

External cues direct attention to the effect of the movement on the environment. Internal cues direct attention to body mechanics, which increases conscious control and actually interferes with performance, especially under pressure.

For jiu jitsu coaching: cue the outcome and the environment, not the body parts. “Don’t let their head touch the mat” is better than “put your hand on their shoulder.” “Drive through them” is better than “extend your hips.”

Analogies Over Technical Prescriptions

Gray’s research shows that analogies produce better transfer than step-by-step instructions because they capture the invariant features of a movement without locking students into specific joint angles.

“Stay connected like a seatbelt” is better than “keep your right arm under their chin and your left hand grabbing your right bicep.” The analogy captures the key concept (continuous connection that adapts to movement) while allowing the student to find their own mechanical solution.

Three types of analogies (from Winkelman):

Scenario-based: Map the movement onto an analogous situation. “Sprint as if climbing a vertical cliff.” These are emotionally provocative and trigger intuitive understanding.

Constraint-based: Establish mental rules. “Front squat as if there are vertical pillars just in front of each plate.” These organize movement patterns without prescribing specific positions.

Object-based: Map movement onto a familiar object’s motion. “Punch through the ceiling” vs. “push toward the ceiling.” Word choice dramatically affects how the body interprets the instruction.

The 3D Cueing Model

Winkelman’s framework for building effective cues:

Distance: Where to focus attention. Close-external (“drive your knee through the gap”) vs. far-external (“end up on the other side”). Beginners often need closer focus; advanced students benefit from outcome focus.

Direction: Which way to move. Use spatial prepositions: “away from,” “toward,” “through,” “out of.” These give the movement a target.

Description: How to move. Action verbs are the heartbeat of effective cues. Specific verbs (“punch,” “drive,” “squeeze”) generate greater motor cortex activation than vague ones (“move,” “do”). Verbs delivered before movement activate the motor system. Verbs delivered during movement interfere with it.

When to Talk and When to Be Quiet

Before a round: Brief concept introduction. One or two ideas. Use permissive language: “You could try…” rather than “You must…” You’re guiding exploration, not prescribing solutions.

During a round: Mostly quiet. Research shows that 64% of coaches give instructions during performance, and it’s consistently less effective than pre/post feedback. If you must cue during play, keep it to one external cue. Avoid technically-focused commentary while students are moving.

During demonstrations: Be quiet. Don’t narrate the mechanics while showing the movement. Show it under different constraints so students pick up invariants (what stays the same across variations) rather than copying one specific execution. Use real-time speed, not slow motion.

After a round: This is where feedback lives. One point per stoppage. Not a sermon.

Fast and Focused Feedback

From Lemov: one point per stoppage. Working memory is limited. If you give students five things to think about, they’ll remember zero.

Aligned feedback is critical. If you stop the game to teach a concept, your live feedback during the next round should reinforce that same concept. If you teach point A then give feedback on points B, C, and D, the message is that point A wasn’t important.

Knowledge of results (did it work?) is essential and can be provided frequently. Knowledge of process (what about the movement led to the outcome?) is useful but should use external focus. Avoid technique feedback that induces internal focus.

External focus with 33% feedback frequency produced similar learning to 100% frequency in research. You don’t need to cue every rep. Resist the urge to over-coach.

Observation and Assessment

How you watch matters as much as what you say.

Look for Invariants, Not Errors

The traditional coaching approach: compare what you see to a mental model of “correct” technique and identify deviations to fix. Gray calls this looking through a “filter” of what things should look like.

The problem: this filter makes you see what you expect rather than what’s actually happening. You look for errors to correct rather than picking up what’s working. You view students as problems to solve rather than individuals discovering solutions.

The ecological approach: look for what all successful solutions have in common (the invariants) rather than comparing to one ideal. When a student passes guard in an unusual way but it works, ask: what essential features did their pass share with a conventional pass? Probably: they controlled the hip, cleared the knee line, and established a connection. The specific mechanics differed, but the invariants were the same.

This matters because there is no single correct technique. Different body types, leverage profiles, and opponent responses demand different movement solutions. A 130-pound guard player and a 220-pound wrestler will have structurally different passes that share the same invariants.

What to Watch For

Focus on essential variables (outcomes and coordination patterns) rather than elemental variables (specific joint angles and positions):

Is the student achieving the positional goal?
Are they maintaining alignment while doing it?
Are they perceiving and responding to their opponent’s movements?
Is the movement adaptive (changing with context) or rigid (same pattern regardless)?

When something isn’t working, resist the urge to prescribe the fix. Instead, design a constraint that makes the current approach fail more obviously and the better approach more discoverable.

Being an Adaptive Coach

Gray distinguishes between adapted coaches (who adjusted to one approach and stay there) and adaptive coaches (who continuously respond to what’s emerging).

Adaptive coaching means:

Being prepared but not rigidly planned. Have a structure for class, but be ready to change direction based on what you observe.
Recognizing that constraints don’t always produce predicted outcomes. Complex systems are non-deterministic. When a game isn’t producing the learning you expected, adjust the constraints rather than adding more verbal instruction.
Developing “knowledge of” coaching through practice, not just “knowledge about” coaching through study. Reading this page gives you knowledge about pedagogy. Running classes and reflecting on what happened gives you knowledge of pedagogy. Both matter, but the second can only come from doing.

Practice Design

Skill Progression

Lemov identifies three types of training activities:

Skill-acquisition activities. Build automaticity in specific actions. Blocked drilling of entries, movements, and transitions. The goal is fluency: reducing working memory load so students can perceive and decide rather than just execute.

Game-based activities. Develop decision-making through perception-rich environments. This is the bulk of our class time. Small-sided, constrained games that multiply decision-making reps and create high-value situations.

Tactical activities. Recreate specific match conditions. Competition prep, specific scenarios from upcoming opponents. Not a regular part of fundamentals training.

The progression is: blocked practice → serial practice (connecting moves into chains) → random practice (full variability). We move through this progression within a single class: a drilled entry feeds into a serial game which feeds into positional sparring.

Creating Variability

Variability is not noise. It’s the raw material for adaptability. Rob Gray’s research shows that there is no single “correct” technique. Every performance context is unique, and skilled performance requires multiple overlapping solutions to the same problem.

Practical implications for class design:

Vary training partners (size, style, skill level)
Vary starting positions within the same topic
Vary rule constraints across rounds
Avoid drilling the same technique with the same partner for extended periods

The “variability-overuse hypothesis” also applies to injury prevention: repetitive stress on the same tissues causes breakdown. Encouraging varied movement patterns distributes stress and reduces injury risk.

Cognitive Load Management

Students have limited working memory. The more they know about a topic, the more new information they can absorb. Beginners have almost no scaffolding and get overwhelmed fast.

For fundamentals: Pair it down to core concepts. Give them one or two ideas, then let them play. Layer information gradually: play a round with no instruction, introduce one concept, play another round.

For advanced: Students have enough scaffolding to absorb more per session. Deep-dive into single topics over multiple weeks. Start with concepts and movement pathways, progressively add detail and systems.

How We Teach Fundamentals

Fundamentals classes prioritize breadth: building a wide, well-labeled mental map so students feel comfortable everywhere even if they’re not yet skilled anywhere.

Every class follows the same four-segment format (see Curriculum Design Report for full details):

Warmup (~2 min): light movement
Evergreen game (~8 min): constant game covering a broad positional category
Retrieval (~10-15 min): revisiting a topic from 2-6 weeks ago
Main topic (~35-40 min): the week’s primary instructional content

The main topic block follows a three-stage progression within each class:

Blocked drilling of the entry to the position
Serial game: drill the transition into the game position, then play
Positional sparring: open play starting from or passing through the position

This progression moves students from isolated technique to game-like application within a single class.

Multiple positional engagement types appear in a single session through the evergreen and retrieval segments. The goal is that before a student touches the main topic, they’ve already done 18-23 minutes of active grappling in other positions.

What We’re Developing in Fundamentals Students

In priority order:

Playfulness and training culture. Can they invest in loss, self-handicap, keep the rally going?
Awareness. Can they read alignment in real time? Can they perceive threats and opportunities?
Problem-solving through play. Can they use the conceptual framework to solve novel problems?
Concrete skills. Do they have a baseline of practical ability across all major areas?

We list concrete skills last because they’re the output of the first three, not the input. A student with strong awareness and good problem-solving habits develops concrete skills faster than a student who memorizes techniques but can’t read a live situation.

How We Teach Advanced

Advanced classes assume students have decent movement quality and a large enough mental map to absorb more per session.

Where fundamentals covers multiple engagements per class, advanced deep-dives into a single positional engagement for several weeks. We’re building deep pathways rather than wide and shallow ones.

The progression across a multi-week block:

Concepts and movement pathways (orient to the engagement)
Key controls and battles (what are both players fighting over?)
Progressively add detail, systems, and fine-tuning
Full positional sparring with increasing complexity

There are differences in how we’d run an advanced series on arm bars (more linear, more detail-friendly) versus guard passing (less linear, more concept-driven). We adjust based on who’s there and how everyone is progressing.

Resources

How We Learn to Move (Rob Gray)

The foundational text on ecological dynamics and motor learning. Covers the constraints-led approach, perception-action coupling, self-organization, and why variability matters. Essential reading for understanding why we teach the way we do.

Learning to Optimize Movement (Rob Gray)

The advanced companion. Covers information sources (tau, optic flow), gaze control, metastability, and how practice design shapes perception. Relevant for understanding how to design progressions and why some drills transfer and others don’t.

The Coach’s Guide to Teaching (Doug Lemov)

Practice design, feedback principles, and the progression from skill-acquisition to game-based to tactical activities. The source for our understanding of cognitive load, aligned feedback, and why “the game teaches the game” is insufficient without intentional design.

Learning to Be an Ecological Coach (Rob Gray)

How to apply ecological dynamics as a coach. Covers the constraints-led approach in practice, observation skills, when to talk vs. stay quiet, adaptive coaching, and the difference between knowledge “of” and knowledge “about” coaching. The most directly applicable of the three Gray books.

The Language of Coaching (Nick Winkelman)

The science of coaching cues. External vs. internal focus, the 3D cueing model, analogies, and feedback frequency. Essential reading for any coach who wants to communicate more effectively.

Ep. 262: Gamification, feat. Rob Biernacki (BJJ Mental Models)

How and why gamification works in jiu jitsu.

Buzzsprout