The Feynman technique: 4 steps, a worked example and why it works

The Feynman technique is a way to find out whether you understand something: pick a concept, explain it in plain words as if to someone who has never heard of it, notice where the explanation breaks down, then go back to the source and fix exactly those parts. It is named after the physicist Richard Feynman, who was known for explaining hard ideas simply. It works because explaining exposes gaps that rereading hides. This guide covers the four steps, a worked example, what the research says, the mistakes that make it feel productive without being so, and how to use it with a board and an AI without handing over the explaining.

What is the Feynman learning technique?

It is a test of understanding that you run on yourself. Most study methods check whether you can remember something. This one checks whether you can explain why it is true, which is a harder bar and the one most exam questions above recall level are set at.

Feynman did not publish a four-step method. The steps below are a later summary of how he described learning, popularised by study writers, and the version taught by university study centres such as York and Colorado is close to it. The name matters less than the move at its centre: write the explanation out, in full sentences, with no jargon to hide behind.

The 4 steps of the Feynman technique

1. Pick one concept and write it at the top of a pageOne concept, not one chapter. "How a vaccine produces immunity" or "why the derivative of sin is cos" is the right size. If you cannot explain it in about a page, split it.
2. Explain it to a beginner, from memoryClose the book. Write the explanation as if to a smart twelve-year-old: plain words, short sentences, one example. Every technical term has to be defined in the sentence that uses it. Draw the diagram if the idea has a shape. Keep going until you reach the end or get stuck.
3. Find the gaps and go back to the sourceMark every place you stalled, waved your hands, or used a term you could not define. Those marks are the point of the exercise. Open the textbook or slides and reread only those parts until you can explain them.
4. Simplify and explain againRewrite the explanation on a fresh page without looking at the first one. Cut the jargon you leaned on and replace it with what it means. Use an analogy only if you can say where it breaks. When the second version runs start to finish without a stall, you understand it well enough to move on.

Some versions add a fifth step: teach it to a real person. If someone is willing to listen and ask "why?", do it. Their questions find gaps you cannot see yourself.

A worked example

Say the concept is how the kidney filters blood, for a physiology exam. Here is what each step tends to produce.

The first version was not wrong. It was a list of facts with the explanation missing, which is what notes made by rereading usually look like, and it is exactly the kind of answer that loses marks on an "explain why" question.

Why it works

Nobody has run a trial of "the Feynman technique" under that name. Its parts have been studied a great deal.

• It catches the illusion of understanding. Rozenblit and Keil asked people to rate how well they understood everyday devices such as a zip or a flush toilet, then to write a step-by-step explanation of how one works. After trying to explain, people rated their own understanding much lower [1]. Writing the explanation is what punctured the feeling. Step two is that experiment, run on your own course.
• Explaining builds understanding. Chi and colleagues asked eighth-graders to explain each line of a text on the circulatory system to themselves as they read. The students who explained most all ended up with a correct model of how the system works; many students who read the text twice without explaining did not [2]. Dunlosky and colleagues' review of study techniques rated self-explanation as moderately useful, ahead of rereading and highlighting [3].
• Teaching helps, and so does preparing to teach. Fiorella and Mayer found that students who studied a lesson expecting to teach it understood it better on an immediate test than students who studied expecting to be tested. Preparing to teach and then actually explaining the lesson produced the best long-term learning [4].

The second explanation is also retrieval practice, since you write it from memory. That is one more reason to close the book before you start.

Mistakes that make it useless

• Explaining with the book open. That is paraphrasing. The test only works if you have to produce the explanation yourself.
• Explaining it in your head. Silent explanations skip the hard parts without you noticing. Write it or say it out loud.
• Hiding behind terms. "The ATP powers it" is a placeholder. If a word is doing the explaining, define it or replace it.
• Doing it for everything. It is slow. Use it on the ideas you need to explain in an exam, and use active recall and spaced flashcards for the facts.
• Stopping at the first draft. Finding the gaps is step three. Most of the learning happens in steps three and four.

The Feynman technique on a board

Paper is enough to start. A board helps with step three, where you need the source next to your explanation, and with keeping each draft so you can see what changed.

In Bord, write your explanation in the notebook on a blank part of the board, in boxes and arrows if the idea has a shape. For the check, add your slides or chapter to the same board as a source (that needs a free account), or ask Ask to lay out the concept in a section beside yours. When Ask works from your file, its claims link to the page they came from, so each gap you mark points at what to reread. Then write version two on a fresh patch next to the first.

You can also have something ask you "why?". Select your explanation and mention it to Ask with a request like "find the gaps in this explanation; don't rewrite it". Or switch to Tutor and have it teach the concept out loud, stopping at checkpoints that ask you to answer in your own words. Keep the order: explain first, then compare. If the AI explains first, you are back to reading a good explanation and feeling that you understand it, which is the illusion the technique exists to catch. There is more on that split in how to study with AI without cheating yourself.

Questions