Your Brain Learns Better When It Feels Like It's Failing

In 2007, Doug Rohrer and Kelli Taylor at the University of South Florida gave students a simple task. Practice calculating the volumes of geometric shapes. Wedges, spheroids, spherical cones, half cones. Basic math with formulas.

One group practiced in blocks. All the wedge problems first, then all the spheroid problems, then the cones. Neat, organized, logical. The other group practiced the same problems mixed together. Wedge, spheroid, cone, wedge, cone, spheroid. No pattern. No rhythm.

During practice, the blocked group looked better. They were faster. More accurate. If you walked into the room, you'd bet money on them.

One week later, on a test with new problems, the interleaved group scored 63%. The blocked group scored 20%.

That's not a small difference. The group that felt worse during practice crushed the group that felt organized and productive. By a factor of three.

The Illusion of Mastery

This is the pattern that keeps showing up across the Practice Paradox. The strategies that feel effective aren't. The strategies that feel like struggle are.

Blocked practice feels good for an obvious reason. When you do twenty wedge problems in a row, you get into a groove. By problem five, you've got the formula memorized. By problem ten, you're on autopilot. Each correct answer reinforces the feeling that you're learning.

You're not learning. You're performing. And performance during practice is a terrible predictor of actual retention.

Richard Schmidt and Robert Bjork identified this distinction back in 1992. They called it the difference between performance and learning. Performance is what you can do right now, in this context, with these cues. Learning is what you can do later, in a different context, without the cues. Blocked practice inflates performance while undermining learning. You look good in the gym and fall apart in the game.

It Works for Everything

The math result alone would be interesting. But Nate Kornell and Robert Bjork at UCLA tested interleaving in a completely different domain in 2008. No formulas. No calculation. Art.

They had participants study paintings by 12 obscure artists. The goal was to learn each artist's style well enough to identify new paintings they'd never seen. One group studied in blocks. Six paintings by Artist A, then six by Artist B, and so on. The other group saw the same paintings interleaved. Artist A, Artist C, Artist B, Artist A, Artist D.

On the test, the interleaved group identified new paintings with 68% accuracy. The blocked group hit 51%.

Here's the part that still gets me. After the experiment, 78% of participants said they believed blocking was the more effective strategy. Even the people who were in the interleaving condition. Even the people who had just performed better because of interleaving. They still felt like blocking would have worked better.

The subjective experience of learning was completely detached from actual learning. Your feelings about what works are wrong. Not slightly wrong. Systematically, reliably wrong.

Why Mixing Works

The mechanism isn't mysterious once you see it.

When you practice in blocks, you don't have to figure out what kind of problem you're looking at. You know it's a wedge problem because you've been doing wedge problems for the last fifteen minutes. The category is given to you. All you have to do is execute the formula.

When problems are interleaved, every single trial starts with a question the blocked group never has to answer: what kind of problem is this?

That question changes everything. You have to look at the problem, compare it to other types you've seen, identify the relevant features, select the right approach, and then execute. Monica Birnbaum, Nate Kornell, Elizabeth Bjork, and Robert Bjork confirmed this in a 2013 study published in the Journal of Experimental Psychology: Learning, Memory, and Cognition. Interleaving's advantage comes specifically from improving your ability to discriminate between categories.

This is exactly what real performance requires. In school, on a test, in your career, problems don't come labeled. Nobody tells you "this is a statistics problem" or "this requires the third approach you learned." You have to figure that out yourself. Blocked practice never trains that skill. Interleaving trains it on every single rep.

The Discrimination Problem

Think about what happens when a medical student studies diseases in blocks. Monday is all heart conditions. Tuesday is all respiratory issues. Wednesday is all gastrointestinal disorders. Within each block, they learn to distinguish between similar conditions in the same category.

But a patient doesn't walk in and say "I have a heart problem." They say "my chest hurts." Now the doctor needs to discriminate across categories. Is this cardiac? Pulmonary? Musculoskeletal? Anxiety? The blocked study approach never practiced that discrimination.

Or think about programming. You could spend a week on arrays, then a week on trees, then a week on graphs. By the end of each week, you'd feel confident with that data structure. Then you sit down for a technical interview and someone gives you a problem. No label. No hint about which structure to use. You freeze. Not because you don't know the structures, but because you never practiced figuring out which one to apply.

I've felt this. I spent years learning things in blocks because it felt productive. Neat categories. Organized study sessions. Clear progress within each topic. Then I'd hit a real problem that mixed concepts and feel like I'd never studied at all. That wasn't a failure of knowledge. It was a failure of discrimination. I had the pieces but no practice assembling them under ambiguity.

The 78% Problem

The Kornell and Bjork finding about subjective experience haunts me. 78% of people believed the less effective strategy was more effective. After experiencing the more effective one.

This connects to something Frank Dempster wrote about in 1988 in the American Psychologist. He called spacing (a close cousin of interleaving) "a case study in the failure to apply the results of psychological research." The evidence was overwhelming by the late 1980s. Nobody changed their behavior. Not students. Not teachers. Not textbook publishers.

Why? Because the strategies that work feel wrong. Interleaving feels chaotic. Spacing feels like you're forgetting. Retrieval practice feels like failure. Every effective learning strategy shares this property: it introduces difficulty that feels like a problem rather than a feature.

Elizabeth and Robert Bjork coined the term "desirable difficulties" to describe exactly this. The difficulty isn't a bug. The difficulty is where the learning happens. When practice feels easy and smooth, it usually means your brain isn't working hard enough to build durable memory.

What This Actually Looks Like

Interleaving isn't random chaos. It's structured mixing.

If you're studying three topics, don't do AAA BBB CCC. Do ABC BCA CAB. If you're practicing guitar, don't play the same scale for twenty minutes. Play three different scales, switching every few minutes. If you're learning a language, don't drill all the verb conjugations for one tense before moving to the next. Mix them.

It will feel worse. Your accuracy during practice will drop. You'll feel less fluent, less confident, less productive. That feeling is the learning happening.

Robert Kerr and Barbara Booth showed this with motor skills back in 1978. Children who practiced throwing beanbags at targets from varied distances performed better on a test than children who practiced from a single distance. Even when tested at the exact distance the blocked group had practiced. The varied practice group outperformed the specialists at their own specialty.

That finding should stop you in your tracks. The people who practiced the test condition exclusively did worse on that condition than the people who barely practiced it but mixed it with others.

The Real Paradox

Every part of this series comes back to the same uncomfortable truth. Your intuitions about learning are backwards. Not occasionally. Systematically.

You feel productive when you reread and highlight. Testing yourself works better. You feel efficient when you cram before the deadline. Spacing over weeks works better. You feel organized when you practice one skill at a time. Mixing skills works better.

The entire educational system, from elementary school worksheets to corporate training programs, is built on the assumption that practice should feel smooth and organized. Textbooks group similar problems together. Training modules cover one topic at a time. Practice schedules repeat the same drill until it's "mastered."

The research from Rohrer, Kornell, Bjork, Birnbaum, Schmidt, and dozens of others says this is almost exactly wrong. The chaos is the point. The struggle is the mechanism. The feeling of "I don't think this is working" is frequently the clearest signal that it is.

Sources

• The Shuffling of Mathematics Problems Improves Learning (Rohrer & Taylor, 2007, Instructional Science)
• Learning Concepts and Categories: Is Spacing the "Enemy of Induction"? (Kornell & Bjork, 2008, Psychological Science)
• Why Interleaving Enhances Inductive Learning (Birnbaum, Kornell, Bjork & Bjork, 2013, Journal of Experimental Psychology: Learning, Memory, and Cognition)
• New Conceptualizations of Practice: Common Principles in Three Paradigms Suggest New Concepts for Training (Schmidt & Bjork, 1992, Psychological Science)
• Making Things Hard on Yourself, But in a Good Way: Creating Desirable Difficulties to Enhance Learning (Bjork & Bjork, 2011, in Psychology and the Real World)
• The Spacing Effect: A Case Study in the Failure to Apply the Results of Psychological Research (Dempster, 1988, American Psychologist)
• Specific and Varied Practice of Motor Skill (Kerr & Booth, 1978, Perceptual and Motor Skills)

---

Part of the Practice Paradox series. Previous: The Worse You Practice, the Better You Learn. Next: 10,000 Hours of What, Exactly?.