Peak by Anders Ericsson argues that much, if not all, of expertise can be explained by having repertoire of patterns that are easily recognized in the domain at hand. The natural question then is, how can we effectively communicate these patterns through teaching in order to create better expertise?

Pattern recognition: When you are presented with a new problem, what patterns do you see? What ideas or leads come to mind? What do you pay attention to? What do you ignore?

In an expert, these questions are answered autonomously by highly-honed neural pathways. This is why, when presented with a new domain, novices often experience an overwhelming feeling of “Where do I put my attention?”, but experts don’t experience this.

Perhaps many of the problems we face around teaching expertise occur because we don’t have a language for it. Thus I think speaking in the language of pattern recognition can be helpful.

If we want to improve a skill, we should deliberately focus on developing our ability to recognize patterns necessary for the skill.

One way to do this is to ask an expert to articulate their thoughts in real time as they’re presented with a new problem. What pops out at you? What do you focus on? What do you tend to ignore? What specific information primed you to generate each hypothesis that came to your mind?

However, the ability to articulate helpful answers to these questions is a rare skill.

When explanations fail to explain

Some of the problems in my classes require the use of pattern recognition that isn’t explicitly taught. When I ask someone to explain the solution to one of these problems, they’ll explain the solution—but often they won’t explain the pattern recognition necessary to generate the solution anew.

Almost everyone—except the best teachers—fails to helpfully articulate why they thought to do the steps that solved the problem.

Instead, the explanation is usually akin to “well if you do this obscure trick, then it works”. (Gee, thanks.)

But why did you do that specific trick, and nothing else? The space of possible actions was infinite—what led you to choose that specific option? What implicit information did you see in the problem? What specifically primed you to see what you saw?

It’s funny— finding the ‘trick’ of a problem is usually sole point of difficulty, yet we often fail to dwell on the process and pattern recognition necessary for discovering the trick autonomously.

Why does this occur?

Two hypotheses:

  1. We don’t usually have the common language to explicitly discuss the development of expertise.
  2. Much of the problem seems to be due to the curse of knowledge—once X is obvious to you, it’s difficult to simulate your prior mind-state when X wasn’t so obvious. At some point, understanding ‘clicks’, and then we forget what made it click. Because of this, quality, empathetic explanations can be difficult to create.


For these reasons, I deliberately try to ask for the pattern of thought that lead to recognizing the solution, rather than asking merely for an explanation of the solution.

It seems that asking these questions is at least slightly more helpful, but it still usually fails often because the person I’m asking a) isn’t actually an expert; b) had merely memorized the ‘trick’ anyway; or c) is not correcting for the curse of knowledge.

So, for now, my test to determine whether someone is truly an expert and skillful teacher is to ask them to explain their pattern recognition for solving problems.