The Brain Learns Better When It Doesn’t Know the Answer

The answers that stick with us longest are the ones we didn’t know, even if only for a moment.

That statement runs against our intuitions about learning. It seems obvious that to remember something you first have to see it, read it, and understand it—and to save testing yourself for the very end. Yet it’s exactly that brief moment of helplessness, when you take a guess and miss, that prepares the ground in which new knowledge will sink solid roots.

Most people arrange their studying in a simple sequence: first the material, then review, and a test somewhere at the very end—as a check on what they already know. By this logic, asking a question before we’ve read anything looks like a pure waste of time—after all, if we don’t know the answer, what could come of it besides frustration? So many people put off the harder material until the moment they feel “ready,” convinced that only then does studying make sense. And this is precisely where a mistake hides that you can’t spot with the naked eye. It concerns the order in which we let the brain encounter a new portion of knowledge.

To see this, it’s worth looking at an experiment that, at first glance, looked like a researchers’ joke. Kornell and his colleagues asked participants about facts that had never existed. They were asked, for example, about the treaties ending a war that in reality never took place¹. No one could know the correct answer, so every attempt was doomed to fail from the outset. And yet the people who first had to guess at something, and only afterward learned the real information, remembered it better than those who were shown the question together with the ready-made answer right away. The very effort of an unsuccessful search, even though it ended in error, left a trace in memory that passively reading a “cheat sheet” never created.

You might think this is a curiosity that works only with made-up riddles, but the effect shows up with ordinary study material too. In other studies on pretesting, one group read a text containing unfamiliar concepts only after grappling with questions about them. And it was precisely this group that performed noticeably better afterward—the difference reached 10 to 19% in retention². Tellingly, the advantage held up especially when only those questions answered incorrectly earlier were taken into account. It was the failed attempt itself, rather than a lucky hit, that turned out to be what strengthened later memory.

There is, however, a condition that must not be overlooked. A wrong answer teaches nothing good on its own unless a correction follows it—only feedback turns a slip into lasting knowledge. Research on answer generation and feedback indicates that a corrected error doesn’t harm learning, but an error left without correction can cement itself as a false belief³. What has value, then, is not the guessing itself but the whole cycle: attempt → mistake → immediately seeing the correct answer.

The most surprising thing is what happens when you’re wrong with complete conviction. Counterintuitively, it’s with the answers given with the greatest self-assurance—if they turn out to be wrong—that, after correction, there is the greatest chance you’ll recall the correct version next time⁴. The more firmly you believe you’re right, the more strongly the truth surprises you. That surprise probably then acts like a vivid cue that helps retrieve the correct version from memory. Confidence in error, which we usually feel ashamed of, paradoxically turns out to be one of the best possible situations for learning. But only on the condition that a reliable correction appears right after it.

These findings lead to a surprisingly simple conclusion for everyday learning. When you open new material you don’t yet know, don’t wait for the moment when you feel you already know something. Try to answer right away, even if you’re certain you’re shooting in the dark. The moment you admit you don’t know the answer is not a failure or proof that studying started too soon; it’s the moment when the brain makes room for the information about to arrive. The second thing is being honest with yourself when judging your own answers. Inflating your scores just to avoid admitting ignorance cuts off access to the very mechanism that works in your favor.

This is exactly what Glimsy is built on: the app doesn’t require you to know everything first. Every answer, including admitting you don’t know it, is valuable information for it. After each such signal, the algorithm rebuilds your review schedule anew, so that the harder portions of learning come back to you more often until they settle into memory for good. The button you tap to mark that the material gave you trouble isn’t a reason for shame but data—data that makes the next reviews appear exactly when they’re needed most.

The hardest thing to accept is that good learning can be unpleasant precisely when it’s effective. The effort of failed recall and the small sting of an error aren’t side effects but the very heart of remembering. So take care of this one small change of habit: next time, instead of putting off difficult material for a better moment, try tackling it right away and allow yourself a few mistakes at the start. If you see the correct answer right afterward, you’ll discover that it’s the one that stays in memory the longest.