Roshan Singh • 5 January 2026 • 11 min read

Why Cramming Doesn't Work (And What Does)

Your brain deletes 75% of what you learn within a week. Rereading won't fix it. Here's what 140 years of cognitive science says actually works.

You've been lied to.

Every night before an exam, millions of students sit in their rooms, textbooks open, highlighters in hand, believing that if they just read through this chapter one more time, it will finally stick.

It won't.

This isn't opinion. This isn't some motivational speaker nonsense. This is 140 years of cognitive science, hundreds of peer-reviewed papers, thousands of experiments. The research is overwhelming and clear: the way most students study is fundamentally broken.

Nobody told you. Not your school. Not your coaching. Not your parents. So let me.

Your Brain Is Designed to Forget

In 1885, a German psychologist named Hermann Ebbinghaus got obsessed with a weird question: how fast do we forget things?

He spent years memorizing nonsense syllables (things like "WID" and "ZOF" so prior knowledge wouldn't help) and testing himself at different intervals. What he found was brutal.

Within 20 minutes: 42% gone. Within an hour: 56% gone. Within 24 hours: 67% gone. Within a week: 75% gone.

Your brain is not a hard drive. It's a filter. It's constantly asking: "Do I need this? Will this kill me if I forget it? No? Delete."

Now think about what happens in a typical coaching class. You sit through a 2-hour lecture on thermodynamics. You take notes. You go home. You have five other subjects to worry about. You don't look at those notes again for three weeks.

By the time you open that notebook again, your brain has deleted 75% of it. You're not revising. You're relearning. From scratch. And that panicked relearning you do the night before the exam? Your brain will delete most of that too within a week.

This is why you can "study" for hours and still blank on the exam. This is why you've read the same chapter four times and still can't solve problems. Your brain isn't broken. You're just fighting its nature instead of working with it.

So what actually works?

Review within 24 hours. I know, I know, you have other classes, other subjects, you're tired. But even 10 minutes of review the next day makes a massive difference. Then review again after 3 days. Then a week. Then two weeks.

Each review is shorter than the last because you're not relearning, you're reinforcing. The memory gets stronger each time, and the forgetting slows down.

Stop Reading. Start Testing.

Here's the thing that broke my brain when I first learned it: testing yourself is a better way to learn than studying.

Not "also good." Better. Significantly better.

Roediger and Karpicke, 2006. Students read passages. Some reread them four times. Others read once and then took three practice tests (no feedback, just trying to recall what they read).

Five minutes later, the rereaders did slightly better. Makes sense, right? They spent more time with the material.

One week later? The testing group remembered 50% more.

Fifty percent. That's not a marginal improvement. That's the difference between failing and passing.

Why? Because there's a huge difference between recognition and recall. When you reread something, your brain goes "oh yeah, I've seen this, I know this." That's recognition. It feels like learning. It's not.

Recall is when you close the book and force your brain to retrieve the information from memory. That's hard. That's uncomfortable. That's also what actually strengthens the memory trace.

Rereading is like looking at a map. Testing yourself is like navigating without one. Which do you think makes you better at finding your way?

Practical version:

Finish a study session. Close everything. Take a blank sheet of paper. Write down everything you can remember about what you just studied. Don't peek. Just dump your brain onto the page.

Then open your notes and see what you missed.

Those gaps? That's exactly what you need to focus on. You just identified your weak points in 5 minutes instead of discovering them on the exam.

Do this after every study session. It feels terrible at first because you realize how little you actually retained. Good. Now you know. Now you can fix it.

Flashcards work on the same principle. The act of trying to recall the answer before flipping the card is what creates learning. If you just flip through and read both sides, you're wasting your time.

The Spacing Effect: Timing Is Everything

Okay so you need to review things multiple times. But when?

This matters more than you'd think.

Cram everything into one night and you might pass tomorrow's test. But a week later? A month later? On the actual JEE when you need to remember things you studied 8 months ago? Gone.

Space out those same hours of study over several sessions across days or weeks? Retention goes up by 200%. That's not a typo. Two hundred percent better retention for the same total study time, just distributed differently.

There's a whole field of research on optimal spacing intervals. The short version: you want to review something right around when you're about to forget it. Too soon and you're wasting time on stuff you already know. Too late and you've already forgotten, so you're relearning instead of reinforcing.

A simple schedule that works:

Day 1: Learn it
Day 2: Quick review
Day 4: Review again
Day 7: Review again
Day 14: Review again
Day 30: Review again

After that, monthly reviews are usually enough to maintain it until exam time.

There are apps that automate this. Anki is free and popular. It uses an algorithm to figure out exactly when you need to see each flashcard again. The fancier algorithms (like FSRS, which came out of machine learning research) can cut your review load by 20-30% while maintaining the same retention.

I'm not saying you need to use an app. The simple schedule above works fine. But if you're studying thousands of facts for NEET or memorizing reactions for JEE, automation helps.

Why Your Practice Sets Are Sabotaging You

You learn quadratic equations. You do 20 quadratic equation problems. You learn trigonometry. You do 20 trigonometry problems.

This is called blocked practice. It's how every textbook is organized. It's how every coaching module works. It feels efficient.

It's terrible for actual learning.

Rohrer and Taylor, 2007. Students practiced four types of math problems. Half did blocked practice (all problems of one type, then all of another). Half did interleaved practice (problem types mixed randomly).

During practice, the blocked group did better. Felt easier. More confidence. Smooth sailing.

On the test a week later? Interleaved group scored 76%. Blocked group scored 38%.

Double. The "harder" practice produced double the test scores.

Here's why. When you're doing blocked practice on quadratic equations, you already know every problem is a quadratic equation problem. You don't have to identify what kind of problem it is. You don't have to choose a strategy. You just execute the same procedure twenty times.

But what does the exam look like? Problems are mixed. You have to look at a problem and figure out: Is this a quadratic? Should I factor or use the formula? Wait, is this even a quadratic or is it something else?

Blocked practice never trains that skill. You walk into the exam knowing procedures but unable to recognize when to use them.

What to do:

Mix it up. When you practice, include problems from earlier topics. Studying kinematics? Throw in some problems from mechanics you learned last month. Studying organic chemistry? Mix in some physical chemistry.

It will feel harder. You will make more mistakes during practice. Your practice scores will be lower.

Your exam scores will be dramatically higher.

Create your own mixed problem sets. Or just grab problems randomly from different chapters. The discomfort you feel is the learning happening.

When Struggling Is a Waste of Time

Everything so far has been about making practice harder. Testing over rereading. Spacing over cramming. Interleaving over blocking.

But there's a limit.

Sweller and Cooper, 1985. Some students solved algebra problems on their own. Others studied worked examples (problems with step-by-step solutions shown). The worked-example group learned faster, performed better on new problems, and took six times less time.

Wait, what? Isn't struggle supposed to be good?

There's a difference between productive struggle and unproductive struggle.

When you're a complete beginner, you don't have the mental frameworks to approach a problem sensibly. You don't even know what you don't know. So you flail around, try random things, get frustrated, maybe eventually stumble onto an answer. But you learned nothing about the general method. You just got lucky on that specific problem.

Worked examples let you see how an expert thinks. Oh, first they identified this was a quadratic. Then they decided to factor because the coefficients were small. Then they set each factor equal to zero. Now I see the pattern.

Once you see the pattern, then you practice. But struggling before you see the pattern is just wasted time and frustration.

The sequence that works:

Read the concept explanation
Study 3-5 worked examples carefully
For each step, ask: why this step? what principle? what if the numbers were different?
Try similar problems yourself
Check your work, understand mistakes
Move to harder variations

Here's the catch though: worked examples help beginners. Once you understand a concept, they become useless or even harmful. Experts need to practice retrieval, not look at more examples.

So pay attention to where you are. Don't skip examples when you're lost. Don't keep reading examples when you've got it. Adjust.

The Self-Explanation Thing

You can read a worked example and learn nothing. I've done it. You sort of skim through, nod along, tell yourself "yeah that makes sense," move on.

Then you try a similar problem and have no idea what to do.

The missing piece is self-explanation. You have to actively explain each step to yourself, ideally out loud or in writing.

Research shows students who explain each step to themselves learn significantly more than students who just read. It's not even close.

Why does this work? Because explanation forces you to actually process the information instead of just letting it wash over you. When you have to articulate why step 3 follows from step 2, you either understand it or you realize you don't. No more fooling yourself.

Practical version:

Study with the intent to teach. Pretend you're going to explain this to a confused younger student. When you read a step, pause and say out loud why that step makes sense. If you can't explain it, you don't understand it. Go back.

Even better: actually teach someone. Study groups work when people take turns explaining concepts to each other. Teaching is the ultimate test of understanding.

Putting It Together

Here's a simple system. It's not complicated. It just requires consistency.

After each class or study session:

Close everything
Write down what you remember (blank page recall)
Check what you missed
Make flashcards for key facts/formulas

Next day:

10-minute review of yesterday's material
Do a few practice problems (mixed with older topics)

Weekly:

Review flashcards (or use Anki)
Take a practice test on the week's material

Before exams:

Full-length practice tests under timed conditions
Review mistakes, identify patterns in what you get wrong
Focus remaining time on weak areas

That's it. Test yourself instead of rereading. Space out your reviews. Mix up your practice. Study examples before struggling. Explain things to yourself.

None of this is complicated. It's just different from what you've been taught.

Why This Isn't Common Knowledge

If this stuff works so well, why doesn't anyone teach it?

I think about this a lot. Part of it is that the research hasn't trickled down. Teachers are trained to deliver content, not to teach students how to learn. Education schools don't emphasize cognitive science.

Part of it is that these methods feel wrong. Rereading feels productive. Testing yourself feels unpleasant. Interleaved practice feels inefficient. We're wired to prefer what feels easy, not what works.

And part of it is that the coaching industry has no incentive to teach you this. They get paid for lectures delivered, not for knowledge retained. A student who learns efficiently needs less coaching. Bad for business.

Whatever the reason, you know now. What you do with it is up to you.

References

If you want to dig into the research:

Roediger & Karpicke (2006). Test-enhanced learning. Psychological Science.
Rohrer & Taylor (2007). The shuffling of mathematics problems. Instructional Science.
Sweller & Cooper (1985). Worked examples in learning algebra. Cognition and Instruction.
Dunlosky et al. (2013). Improving students' learning with effective techniques. Psychological Science in the Public Interest.
Cepeda et al. (2006). Distributed practice in verbal recall tasks. Psychological Bulletin.

These are the principles I built Eklavya on. The platform applies all of this, retrieval practice, spaced repetition, interleaved problems, worked examples, automatically. It's free at eklavya.io.

But honestly, you don't need an app. You just need to change how you study.

Roshan

Keep exploring