Why watching maths videos feels like revision (and what to do instead)

If you have just watched forty minutes of a paper being worked through and feel like you have revised, you almost certainly have not. Watching someone else do maths builds recognition, not recall — you leave able to follow the method, which is a completely different skill from producing it on a blank page under time pressure. This is an awkward thing for me to write, because I made the maths videos for the AITutor YouTube channel and hundreds of thousands of students have watched them. They work, but only as the first half of something. The second half is the bit almost everyone skips, and it takes about the same amount of time as the video did.

Why the video feels like it worked

The feeling is real and it is not stupidity. A good explanation is built to be followable, so every step arrives exactly when you expected it. That fluency is the problem: your brain reads "this is easy to follow" as "I could do this", and those two things have almost no relationship to each other.

You can test it in ninety seconds. Finish a video, close the laptop, and write the question out from memory. Not the answer — the question. Most students cannot, and the ones who can usually stall at the first line of working. Nothing has gone wrong: you have found the gap between watching and knowing somewhere better than an exam hall.

What the research actually says

This is one of the better-evidenced findings in learning science, and the numbers are worth knowing because they are so lopsided. In Roediger and Karpicke's 2006 experiments, one group of students read a passage four times over; another read it once and spent the remaining time trying to write down everything they could remember, with no peeking. Five minutes later, the re-readers looked like the winners — 83 per cent recall against 71. A week later it had inverted: the re-readers could produce 40 per cent of the material, and the group who had spent most of their time testing themselves produced 61.

So the students who studied the most remembered the least, and the ones who felt best about their revision had learnt the least. A later survey by Karpicke, Butler and Roediger found 57 per cent of students would choose to reread after studying a chapter, against 18 per cent who would test themselves. People pick the method that feels productive.

Maths has a sharper version of this, because a worked solution hides every decision. You see integration by parts chosen; you do not see the two seconds in which someone ruled out substitution — and that decision is the whole question. Watch enough solutions and you accumulate a library of methods you can recognise and cannot select.

The watch-then-do protocol

I give this to every student who tells me they revise by watching things, and it is not complicated:

• Attempt the question cold for five minutes before playing anything. That is enough to find out where you get stuck, which is the only genuinely useful information in the whole exercise — and being stuck first is what makes the explanation stick.
• Watch with the pen down. Copying working as it appears is the most passive thing you can do at a desk: a beautiful transcript, no memory of it.
• Stop the video at the decision points rather than at the algebra. When the method is named, pause and say out loud why that method and not another one. That sentence is what you are actually trying to acquire.
• Close everything and redo the question from scratch, same day. Not "read through it again" — blank page, no notes. If you cannot, you have learnt something true and cheap.
• Do it again two or three days later, with a related question from a paper. The gap is not a formality; it is what makes the method retrievable in June.

Aim for roughly one to one: forty minutes watched means forty minutes of pen-on-paper owed before the session counts as revision. Left alone, most students run at about ten to one, then cannot understand why a mock went badly after hours of revision.

What videos are genuinely good at

I am not about to tell you to stop watching them; that would be daft advice from someone who made several hundred. They are unusually good at three jobs and poor at everything else.

The first is unsticking: when you have tried a question, failed, read the mark scheme and still cannot see why the mark scheme did that, a video is the fastest repair there is — faster than waiting until Tuesday to ask a teacher. The second is watching decision-making on a hard, unfamiliar question, which is why I made the harder series at all. The third is first exposure to a topic missed through illness or a gap in teaching, where someone has to walk you in before practice is possible.

What they cannot do is build speed, accuracy, or the nerve to commit to a method with the clock running. Only your own paper does that. The free video library is best treated as a repair kit you reach for when practice breaks down, not as the practice itself.

Where the hours should actually go

Strip revision back to what earns marks and the core is dull: do questions you cannot yet do, mark them honestly against the mark scheme, and log why each mark went. Everything else — the notes, the videos, the colour-coded timetable — is scaffolding, and scaffolding is only worth building if something is going up behind it.

So the raw material is papers rather than playlists. Work through your own board's past paper archive in the sequence I set out in the right way and the three wrong ways to use past papers, and mark strictly — a video walkthrough is not a substitute for the mark scheme, because it shows one good solution rather than what earns credit. When a question defeats you, that is the moment a walkthrough is worth forty minutes.

Log the reason rather than the topic. "Trigonometric identities" is a topic. "Chose the wrong identity because I did not look at what the answer needed to contain" is a reason, and reasons repeat — which is the whole argument behind the mistakes that cost A-Level students the most marks. Watching a solution rarely surfaces a reason, because on screen the right choice looks inevitable.

The honest version of the fix

Most students who fall into passive revision do not need to buy anything. They need the protocol above, a printed paper, and somebody to hold them to attempting things cold — and a parent who never took A-Level Maths can do the holding perfectly well, because it only requires checking that the page was blank when they started. If your child's mock was disappointing after weeks of apparent revision, this is the first thing I would change, before considering a tutor at all. The four-week mock revision plan is built entirely around it and costs nothing.

Where a tutor earns their money is narrower. It is in the marking — telling the difference between an answer that was fine and an answer that got there by luck — and in choosing the next question, which is a judgement students cannot make about themselves because the questions they most need are the ones they least want. That is also most of what separates an A* from an A. If the watching-instead-of-doing habit has survived every attempt to break it, an external deadline each week is sometimes the only thing that works, and that is a legitimate reason to book lessons. Wanting better explanations is not — those are free, on the channel and elsewhere, and more of them will not fix anything.