“You can follow every step of a mechanism as you read it and still freeze the moment you have to draw the first arrow yourself.”
Chemistry has a tell that gives away who's actually revising and who's just rereading: hand someone a reaction and they can explain what's happening at every step. Hand them a blank page and ask them to draw the mechanism from scratch, and a lot of that confidence disappears. That's not a knowledge problem. It's the gap between following something and producing it, and chemistry tests the second one almost exclusively.
Why chemistry punishes rereading in its own specific way
Most chemistry revision looks like reading through worked examples in a textbook or your notes. You follow the electron-pushing arrows, the logic makes sense, and it feels like understanding. It is understanding, in the sense of comprehension. It is not the same as being able to produce that mechanism yourself under exam conditions, with a slightly different substrate than the one you memorised, and no arrows already drawn for you.
The exam almost never asks for the exact example from your notes. It asks for the same underlying mechanism applied to something you haven't seen before, which means memorising a worked example gets you nowhere if you never practised the actual skill of applying the logic to something new.
What actually works, technique by technique
Mechanisms: draw them from a blank page, with a twist. Don't just redraw the exact example from your notes. Once you can do that from memory, change one thing, a different leaving group, a different nucleophile, a ring instead of a chain, and work through what changes and what doesn't. If you can only reproduce the exact example you memorised, you don't actually have the mechanism, you have one memorised answer.
Problems, not just theory: do them cold, then check. Read a worked example once, then close the book and attempt a similar problem without looking. You will get stuck, and getting stuck at the right point tells you exactly which step you don't actually understand yet. Rereading the worked example again feels productive and skips past exactly that information.
Balancing equations and stoichiometry: never check as you go. Work the whole problem through to an answer first, mistakes included, then check against the correct one. Checking after every line lets you course-correct without ever noticing you'd have gone wrong on your own, which is exactly the situation the exam puts you in.
Naming and functional groups: use them, don't just recognise them. Being able to identify a functional group when it's labelled is one level. Being able to name a molecule you're given cold, or draw a structure from a name, is the level exams actually test. If you can only go one direction (structure to name, but not name to structure), you don't have it fully yet.
Command words matter here too. "State" wants a fact. "Explain" wants the reasoning behind it. "Deduce" wants you to work something out from given information rather than recall it directly. A correct chemistry answer to the wrong command word still loses marks, and it's one of the most common ways students lose marks on questions they actually understood.
A rhythm that fits how chemistry actually builds
- When a topic is new: work through it once with the material open, then immediately try a similar problem closed-book before moving on, rather than waiting until revision season to test yourself on it for the first time.
- A few days later: the same mechanism or problem type, cold, with one thing changed from the version you originally learned.
- Once it's solid: past paper questions specifically on that topic, checked against the actual mark scheme, since chemistry mark schemes are often stricter about showing working than the final answer alone.
- The final week: only the mechanisms and problem types you keep getting wrong. Everything you already have solid doesn't need more time spent on it.
The short version
Chemistry rewards production, not recognition, more directly than most subjects, because so much of it is mechanism and calculation rather than recall. Draw the mechanism blank, change one variable and see what still holds, work problems cold before checking the answer, and pay attention to what the command word is actually asking for. That's the difference between a subject that looks understood while you're reading it and one you can actually do on the day.