Picture the best evening of practice you've ever had. Maybe it's piano, or Spanish verbs, or a set of math problems. You did the same kind of thing again and again, and by the fifteenth repetition your hands knew it. You closed the book feeling quietly proud.

Now the uncomfortable part. That glow is real, but it isn't evidence of what you think it is. Often it's evidence of the opposite: you got good at the repetition, and you may have learned less than a messier evening would have taught you.

The two ways to practice

Say you want to learn three things: A, B, and C. There are two obvious ways to schedule your practice.

The first is blocked practice: AAA, BBB, CCC. You finish one thing before touching the next. It's tidy, and it feels like mastery.

The second is interleaved practice: ABC, BCA, CAB. You mix the topics so no two consecutive items are the same kind. It feels scattered. You keep getting pulled out of a groove just as you were settling into it.

A large body of research on learning, much of it from psychologists Robert and Elizabeth Bjork and their colleagues, finds that the scattered version often produces better long-term learning. They call this kind of effect a desirable difficulty: a condition that slows you down and makes practice feel harder, but strengthens what you retain.

Interleaving is related to the spacing effect but not the same thing. Spacing is about when you return to material. Interleaving is about what is next to what. You can do both at once, and they tend to help each other.

What the experiments actually showed

In a well-known 2008 study, Nate Kornell and Robert Bjork had people learn to recognize the painting styles of different artists. Some saw paintings grouped by artist. Others saw the same paintings mixed together. Afterward, everyone was tested on paintings by those artists that they had never seen.

The mixed group did better at classifying the new works. Yet when asked, most participants believed the grouped version had helped them more. Their feelings pointed one way and their performance pointed the other.

A similar pattern turned up in mathematics. In a 2007 study, Doug Rohrer and Kelli Taylor had students practice finding the volumes of different solids, either in blocks or shuffled. During practice, the blocked group looked better. On a test a week later, the shuffled group did dramatically better.

The lesson isn't that blocked practice is worthless. It's that how practice feels is a poor guide to how well it works.

Why shuffling helps

The leading explanation is the discriminative-contrast hypothesis. When examples of the same type come one after another, you never have to decide what kind of problem you're looking at. The last one told you. You just run the procedure.

In the real world, nobody announces the category. A math exam doesn't label each question "volume of a wedge." A conversation doesn't tell you whether someone wants advice or just wants to be heard. A bird in the trees doesn't wear a name tag.

Interleaving forces the step that blocked practice quietly skips: noticing which kind of thing this is, and choosing a strategy. You also see the categories side by side, so the differences between them become vivid. You stop learning "what a finch looks like" and start learning "what makes a finch different from a sparrow." That second kind of knowledge is what transfers to new situations.

A second ingredient is retrieval. When you switch topics, you have to reload what you know about the new one, which is a small act of recall. Effortful recall strengthens memory more than rereading or repeating does.

Why it feels so wrong

Blocked practice produces something psychologists call an illusion of fluency. When you do the same thing repeatedly, performance rises fast and feels smooth. Smoothness gets read as understanding. But the smoothness is borrowed from the short-term context: the last example is still ringing in your head.

Interleaving removes that borrowed help. Mistakes go up and progress feels slower. Most of us read that as a sign something is going wrong, so we go back to the neat, comfortable method.

There is something almost tender in this. The method that makes you feel competent in the moment is not always the one that makes you competent next month. Learning to tolerate the small sting of not-quite-getting-it is part of getting good at anything.

Where the evidence is strong, and where to be careful

Interleaving seems to help most when the categories are similar and easily confused: types of math problems, bird species, painting styles, medical diagnoses, grammar patterns. It forces you to tell near-neighbors apart.

It's less obviously useful when you are brand new to a skill and each piece is completely unfamiliar. A beginner can be overwhelmed if everything is shuffled from minute one. A reasonable approach, and one many learning researchers suggest, is a short stretch of blocked practice to get a first foothold, then mixing things up soon after.

Also, not every study finds a benefit, and the size of the effect varies with the material. Treat interleaving as a strong default to experiment with, not a magic rule.

Your next moves

You can try this today without buying anything or learning a new system.

  • Shuffle your next practice session. If you're working through exercises, flashcards, or problem sets sorted by type, mix them. Shuffle the cards or pick problems from different chapters at random.
  • Name the type before you solve. Before answering a mixed problem, write down one phrase for what kind of problem it is and why. That is the discrimination step doing its work.
  • Compare near-neighbors on purpose. Choose two things you tend to confuse (two grammar tenses, two historical events, two plants on your street) and study them in alternation, noting one difference each time.
  • Do a short blocked warm-up, then mix. If something is new, spend ten minutes getting oriented with one topic, then rotate to the others.
  • Check yourself a week later. Don't judge by how the session felt. Quiz yourself after several days and see which method held up. Your own results beat any article.

Expect to feel clumsy

When you try this, you will probably feel worse for a few sessions. You'll make more errors and you may feel less sure of yourself. That discomfort is not a malfunction. It's roughly the feeling of your brain building the distinctions it will later rely on.

Curiosity works the same way. A mind that only ever digs one hole gets deep and narrow. A mind that keeps hopping between questions starts noticing patterns that no single topic could show it.

That's the idea behind Curiokit. It offers a steady, varied stream of questions rather than a single subject to grind through, so your attention keeps switching, comparing, and reloading what it knows, much like interleaving does. If you'd like a small daily habit that leans in this direction, you can take a look at Curiokit. And if you don't, shuffle your next study session anyway. It's a good experiment.