The science of unlearning a skill

You never forget how to ride a bike. That's a problem.

In 2015, YouTuber Destin Sandlin made a video about riding a bike.

This bike was slightly unusual. It had been welded such that the handlebar moved the wheel in the opposite direction. So, for example, if he wanted the wheel to turn left, he’d have to turn the handlebars to the right. And visa versa. Everything else was normal.

Sandlin was a cyclist the way most people are cyclists. That is, he learned to ride a bike at age 6, and that was it. He was a bike rider from that point on. Not a professional, mind you, but someone who could manage a bike lane. It’s not a particularly impressive skill. Monkeys can ride bikes! The Moscow circus famously employed bears that rode bikes.

What it lacks in impressiveness, though, bike riding sure makes up for in stickiness. So when Sandlin attempted to ride the “reverse” bike, he flailed. At first, he could barely balance himself. His initial few glides were wobbly and violent. He looked, well, like a kid again. Growing frustrated, he took the bike home to practice. He expected that, after a few days of determined effort, he could master this new bike.

It took him eight months.

Destin Sandlin says his experiment with the “reverse” bike “damaged my mind.”

The reason I bring this up is because there was a study published recently in Nature that explores the “reverse bicycle” challenge in more detail. And it offers come constructive insights into how we learn skills to begin with.

The “reverse bicycle” challenge

The first thing to note about riding a bike is that its complexity is underestimated. We take this for granted. We say things like “it’s just like riding a bike,” meant to imply that it’s easy. And for most of us, it is fairly easy, because we learned to ride a bike when we were young, and our brains were fertile and developing. It’s not impossible, but it’s harder to learn to ride a bike as an adult.

What’s required of riding a bike (maintaining balance, subtle manipulations of handlebars, pedaling, etc) is fairly complex, but the instructions to accomplish them are relatively simple — and therefore easier to internalize and retrieve. The more we practice these simple few steps, the more automatic they become.

Think about learning to swing a golf club. You are taught at the outset about the placement of hands and wrists and shoulders, and how to rotate your hips, and where to focus your eyes, and all the little checkmarks of a perfectly fundamental golf swing. And as you’re learning, you have to actively thinking about each one of these steps as you swing. This is dreadfully inefficient for the brain, and prone to errors. But fortunately, there’s a trick: practice. Practice essentially trains the brain to chunk all those tiny steps together as one easily retrievable module. With practice (ideally) you can eventually step up to the ball and hit it.

This is great! Automaticity is a foundational element of skill. Right? When you think of a skilled tennis player, or a skilled skiier, or a skilled pianist, or a skilled anything, they’re not thinking about the initial instructions for how to execute their movements. That’s why they practiced! They put in 10,000 hours (roughly) to get the fundamentals down so that they could use their cognitive resources on other things, like strategizing.

This is great! Until… something changes.

“You  might think that after the millions and millions of balls I’ve hit, I’d have the basic shots of tennis sown up, that reliably hitting a true, smooth, clean shot every time would be a piece of cake. But it isn’t.” - Rafael Nadal

Researchers at USC recently decided to test whether they could confirm what Sandlin had shown during the video. Twenty “novice” participants were asked to practice with a reverse bicycle for 10 minutes for eight consecutive days. By the end, half of them managed to successfully ride the reverse bike for 20 meters without toppling over. Five of them never got the hang of it at all. Cool. Sandlin hypothesis confirmed: learning to ride a reverse bike is pretty dang hard.

The USC researchers took it a bit further. Out of those 10 “fast learners,” they pinpointed three that the authors called “experienced riders.” They were, in fact, three of the papers’ very authors—and they were experienced because they’d been practicing the reverse bicycle “extensively” “as a hobby” in their spare time. All that practice had made them reasonably adept at riding the reverse bicycle, and so they completed the trials along with the novices with relative ease.

The trouble for the “experienced” riders came when they were asked to return to riding a normal bicycle. They were now worse at riding a normal bicycle than novice riders had been when first attempting a reverse bicycle.

Skill interference

The novice participants could return to riding a normal bike pretty easily. Their few days of practice learning on the reverse bike hadn’t yet interfered with their prior knowledge and history of normal bike riding. The difficulty for the “experienced” riders came because they had put so much time into learning the reverse bike. That procedure became engrained, and it crowded out all the prior experience with a normal bicycle. It became automatic.

The problem with automaticity is that it’s inflexible.

The researchers suggest that at some point learning to ride the reverse bicycle requires the unlearning of how to ride a normal bicycle. Until that occurs, the signals for the motor control of both bicycles will get crossed, and you’ll have to be really straining and thinking consciously about your actions, which isn’t ideal for bike riding. But unlearning and relearning is also obviously quite time consuming, frustrating, and humiliating. And it’s a pitfall — a major one — of practicing one thing so much that it becomes automatic.

A common example of this is driving. I grew up in the United States, and thus I learned to drive on the right side of the road. When I visited England — jesus christ, what the hell are they doing over there?! It all got flipped. Driving on the left side totally threw me. I’d have had to unlearn how to drive my normal way in order to learn how to drive in the U.K. (if I hadn’t been too scared shitless to get behind the wheel). After practicing in England for awhile, if I went back to the U.S., it might have taken me some time to readjust to my old “normal” way of driving yet again.

Adaptability is a hallmark of skill

This all brings me to my ultimate point. How much automaticity should we strive for in skill?

This is actually a pretty hotly debated question among movement researchers and coaches. According to the legendary UCLA coach John Wooden, “the importance of repetition until automaticity cannot be overstated!”

That seems to be the wisdom that a lot of coaches have since applied to their teaching. But we know this isn’t correct. Automaticity can’t be the end goal in skill learning. We’ve already seen how one small adjustment can cause an entire lifetime of practice to suddenly collapse and crumble.

Lately, there’s been increasing pushback to the notion of automaticity at all. Proponents of an ecological approach to skill acquisition say the rigidity and inflexibility of automaticity has no bearing in expertise and shouldn’t be conflated with skill whatsoever. Their point is that automaticity (in principle) removes context and the environment out of the equation, and I generally agree with this. Some others argue that expertise actually demands that one counteract the pull of automaticity and seek more control over your performance.

But I’m not sure this is entirely correct either.

The authors of the “reverse” bike study write about three “stages of learning.” The first stage is overcoming the fact that the reverse bike is impossible to ride because your brain keeps automatically retrieving a bad module — the experience riding a normal bicycle. This reflects the bad part of automaticity. A second stage is when riding both the reverse bike and the normal bike become impossible, because your brain has to unlearn its normal-bike procedures. Again, a downside of automaticity.

The third stage is when you get it. You get both of it. Your brain has updated itself to account for riding the reverse bike and you can still select from your prior experience on the normal bicycle.

This is (I think) perhaps an instructive way to consider skill acquisition. Ideally, you want to achieve some degree automaticity (of certain well-trained elements) and use it as a building block. You can go back to it, you can quickly retrieve it, but you’re adaptable enough and sensitive enough to apply changes when the situation requires it. Automaticity, in this regard, is a gift for the expert brain. For one thing, it wards off the psychological demons that bring about choking under pressure (see my previous post!). It also frees your brain for the higher-level control you need to take your game to the next level, such as strategizing or deciding when to disguise your shots (see my other previous post!)

Rafael Nadal wrote in his own autobiography that every shot in tennis — even after millions of shots he has taken — every shot is different and has to be approached in a different way. “No ball arrives the same as another; no shot is identical.”

So it’s good to be wary of the old adage “practice makes perfect.” It also can make you inflexible and vulnerable. Practice enables you to build a reservoir of automatic processes from which you can choose the right one given the circumstance. But elite skill also requires adaptability and intelligence.

The question to ask now is: Are we coaching to achieve that?

Links

Scientists develop mathematical model to optimize sprint performance … Parents, wealth, race drive girls’ chances to play sports … Conquering Pressure! The Effects of Mild-Anxiety Training on Motor Performance Under Pressure During Early Motor Learning … Elite athletes are trained to get a good sleep before competition, but what about after? … It’s the NBA’s ugliest shot. And it keeps going in … Top soccer clubs are using an AI app to scout future stars … Coaches can boost athletes’ mental health by being “authentic leaders” … Professional bowler wins tournament by reading a self-help book throughout

References

Magnard, Justine, et al. "Initial development of skill with a reversed bicycle and a case series of experienced riders." Scientific Reports 14.1 (2024): 4334.