Why drilling doesn't transfer (and what we do instead)

Every grappler who has trained for more than a few months has felt this. You drill a technique fifty times against a cooperative partner. It feels smooth. You can do it in your sleep. Then a live round starts and the move never appears — or you try it with terrible timing and it fails immediately. The instinct is to drill it more. That almost never fixes it.

What motor learning calls the transfer gap

There's a name for the problem: the transfer gap. It's the distance between what you can do in practice and what actually shows up under live resistance. Decades of research in sport science have looked at this, in everything from baseball to tennis to combat sports, and the conclusion is consistent: the gap isn't caused by too few reps. It's caused by what the reps remove from the situation.

When you drill a guard pass against a flat, compliant partner, you keep the shape of the movement but you delete almost everything else: the resistance, the timing pressure, the shifting weight, the defensive reactions, the grip exchanges. Those things aren't background noise — they're the information your nervous system uses to organise the movement in real time.

Why more reps don't fix it

It's easy to assume practice is a volume problem. Twenty reps not working? Try a hundred. Hundred not working? Try two hundred. But what you're really getting good at is the drill itself — a specific sequence performed against a non-responsive partner. That skill is genuinely real. It just doesn't generalise to a different problem (a real opponent) because the perceptual cues are completely different.

Imagine learning to ride a bicycle with training wheels and never taking them off. You can pedal cleanly, steer cleanly, and stop cleanly. But the moment the wheels come off you fall over, because the training wheels removed the very thing you needed to learn: balance.

What we do instead

The approach we use is called the constraints-led approach (CLA), and it comes from a field called ecological dynamics. The short version: instead of showing you a technique to memorise, we design a small training game with clear rules, a clear win condition, and a real partner who is actively trying to beat you.

A simple example: one person tries to keep their back off the mat for 60 seconds, the other tries to flatten them. That single constraint produces guard retention, framing, hip movement, and grip fighting — without anyone having to memorise a "guard retention technique". You discover the solutions that work for your body, against a real opponent, with all the information present.

Because the game preserves the same perceptual information you'll have in open rolling, the skills transfer directly. There's no second step of "now apply what you drilled in a real situation" — you were already in a real situation the whole time.

Does this mean drilling is useless?

No. Drilling has a place — warming up a shape, exploring a new movement at low speed, or rehearsing something just before live play. The mistake is treating drilling as the training method, and live rounds as a test of what you drilled. In our room, live problem-solving is the training. Drilling, when it happens, is the warm-up.

What this looks like in practice

Every session at Grapplers Collective Lisbon is built around live training games. The games scale: a beginner and an experienced grappler can play the same game and both learn from it, because the constraints create the right challenge for each person. Nobody sits out. Nobody waits in line for a turn at the demo dummy. You're solving problems from the first minute.