Explain it to your roommate, who does not care

Reading is silent, and silence lets you skip.

You read a sentence about p-values, some part of your brain files it as understood, and you move on. The skipping is invisible because nothing forces you to produce a complete thought. There's no gap you can see.

Speaking has no such mercy. Sentences have to end.

That's the whole technique. You start explaining, three sentences come out fine, and then you hit a word like "which means" and there's nothing behind it, and you stop.

That stopping point is worth more than another hour of reading, because it's specific. You now know exactly which sentence to fix, instead of vaguely knowing you should review the chapter.

Why out loud specifically

Two things happen when you speak that don't happen when you read or think.

Order is forced. In your head, ideas can sit in a pile — connected in some hazy way, all present at once. Speech is linear. You have to decide what comes first and what depends on what, and that decision reveals whether you have a structure or just a pile.

Connective words have to be earned. "Because," "therefore," "which means," "so." Silently, these slide by. Out loud, saying "therefore" commits you to producing the actual reason, and if you don't have one you'll hear yourself trail off.

Writing does some of this. Speaking does more, because you can't pause and edit.

Who to explain it to

A roommate or friend who doesn't take the course is ideal. They can't fill in your gaps, so the explanation has to be self-contained, and they'll interrupt with "wait, what's a null hypothesis" at exactly the point where you were about to skate past something.

They don't have to care. They mostly won't. They can be doing something else. The point isn't their attention — it's that a person in the room makes you produce complete sentences.

Someone in the course is the second-best option, and there's a good version of it: take turns. You explain the first half, they explain the second, and you're both allowed to say "I don't think that's right."

Nobody works too. Talk to the wall. It feels ridiculous for about thirty seconds. The mechanism doesn't require an audience — it requires speech.

How to run it

Pick a chunk. One lecture, one section, one concept — not "the whole unit."

Close everything. This is not negotiable and it's the part people skip. With notes open you'll glance, and glancing turns the whole thing back into reading.

Explain from the beginning, out loud, as if the person has not taken the course. When you hit a wall — and you will, usually faster than you expect — stop and write down where you stopped. Don't push through it and don't look it up yet.

Keep going past the wall if you can. Get to the end.

Then open the notes and fix only the places you got stuck. That list is your study plan for the next twenty minutes, and it's much shorter than "reread the chapter."

The tell that you're doing it wrong

If your explanation is going smoothly and it sounds like the textbook, you've probably memorized the phrasing rather than the idea.

The test for this: try to explain it with different words. If you can't rephrase it, you don't have it — you have a recording of it. Same thing that makes people freeze when a professor words an exam question differently from the slides.

Another tell: if your explanation contains a term you can't also explain. "It's when the sampling distribution converges" is not an explanation if you can't say what a sampling distribution is. That's a nested gap and speech surfaces them, because you'll hear yourself using a word as if it were a primitive.

When it works best

For anything conceptual. Why does this follow from that. What is this actually doing. Where does the argument come from. This is the method's home ground.

Before office hours. Try to explain it first. You'll walk in with "I get up to here and then I lose it at this specific step" instead of "I don't understand chapter 4," and that's the difference between a five-minute conversation that works and a twenty-minute one that doesn't.

For essay-based courses, unusually well. If you can't say your argument out loud in three sentences, the paper doesn't have an argument yet, and no amount of writing more paragraphs will produce one.

It works less well for procedural skills. You can explain how to do an integration by parts perfectly and still botch the algebra. That's a practice problem, not an explanation problem.

Twenty minutes

Take one lecture from this week. Close everything. Explain it, out loud, to a person or a wall, from the top.

Write down each place you stop. That list is what you didn't know, and until you said it out loud, it looked exactly like the parts you did.