Successive Approximations: What the Berimbolo Taught Me About Learning

When you start to learn something, sometimes you get one part right and forget two others. It seems to take a while to come together. Why is this and what can we learn from how we learn?

The other day, I was drilling a grappling sequence called the berimbolo with my friend/teammate/teacher German. I have heard that “berimbolo” means “scramble” in Portuguese, but I have also heard from native Portuguese speakers that it is a nonsense word. Regardless, as you can see in this video, in which world champion Rafael Mendes makes it look effortless, it’s pretty freaking sweet no matter what you call it.

I’m here to tell you that if I’m any indication, there is nothing effortless about the berimbolo. German and I spent maybe an hour and a half drilling the sequence, and as we worked together, I thought about the movements, of course, and about how to get all the details down. As part of that, I also thought at times about tearing my hair out or banging my head against the wall when I wasn’t able to get my body and my brain to synch up. (Sometimes learning sucks.) German has more facility with the movement pattern than I do, and he was kind enough to give me some cues and suggestions for how to think about moving my body.

And while I was contorting, I also thought about the way we were working together and what our learning process was. It occurred to me that over time, as I drilled the move and got feedback from German about how my positioning felt, I became able to execute more and more details of the sequence appropriately.

Back in the day in behavioral research circles, scientists like B. F. Skinner would shape the behavior of pigeons using rewards, such as food pellets. If the desired behavior was that the pigeon would eventually peck at a lever, the scientists would start out by rewarding the pigeon simply for looking in the direction of the lever. Eventually, they would require more specific behaviors, and the pigeon would then receive a reward for walking to the lever, pecking near the lever, and eventually pecking the lever itself. This shaping of behavior resulted in something called “successive approximations,” where, with appropriate motivation, even a pigeon could learn a desired behavior by slowly, over time, engaging in actions that more and more closely resembled that behavior.

When I was drilling the sequence with German, I was tense, whatever the opposite of “crisp” would be in this context, and, as it turns out, only able to nail some of the important elements of the sequence at a time. For instance, as I inverted, I didn’t roll all the way through, which made it more difficult for me to raise German’s hips off the ground so I could get under them. So the next time, I’d make a point of rolling all the way through, but then I’d neglect to keep tension on German’s hip with my knee, something I had actually been doing effectively in previous reps. And then I’d have to try to get the tension back while still rolling through and then adding onto that keeping my head as close to German’s body as possible while inverted. So, slowly but surely, I incorporated more and more of the component movements of the sequence into my execution of the sequence, taking a step or two forward and then a step or two back.

I have experienced this with weight lifting as well. As I back squat, on one rep I may keep my chest up but allow my knees to collapse. And then on the next rep if I am able to keep my chest up and drive my knees out simultaneously, I may not break parallel.

I’m not saying German and I are pigeons, especially because in these cases the motivation is my own intrinsic desire to be able to execute the movements smoothly and effectively. (Though now I keep thinking, “Mmm…pellets.”) For me, the idea of successive approximations provides more fodder for my coaching and learning processes. As an athlete, I can see that the fact it will take me multiple reps (we’re talking thousands) to be able to berimbolo at anything even slightly approaching a Mendes level is part of the process, not some kind of shortcoming on my part. I may never execute nearly as well as Mendes, but over time, with work, I can become a lot better at it than I am now, and I’m already better at it than I was before I worked with German.

As a coach, the idea of successive approximations does a couple things for me. First, it requires that I have and can communicate a clear understanding of what the desired end behavior is. Second, it requires that I have and can communicate a clear understanding of the component parts of that desired end behavior. And finally, it gives me information on individuals and in the aggregate about which elements of that end behavior might be the most difficult to execute and require the most additional support.

So the next time you find yourself taking two steps forward and one step back while learning a new sequence or even trying to refine a familiar one, try to think about this as part of the process rather than a failing. And as a coach, if you see your athlete executing only some of the cues you have given, remember that this is the first step – the first approximation – which will eventually lead to a reasonable execution of the desired behavior.

Photo courtesy of Shutterstock.

Leave a Comment